OpenAI CEO Sam Altman has data centers full of memory, but seemingly it's the 32GB of RAM in your desktop that's the problem. In an interview over beers with the YouTube channel Cheeky Pint, Altman said that local inference, and, by extension, people having DRAM in computers at home, is "quite terrible."

When discussing local inference on home computers with open-source models, Altman discussed the assumption that people will be able to run local AI on their own computers, the ones they may already own or want to build. Then he explained why that's a bad idea (at the 29-minute mark):

"From a global allocation perspective, actually, it's quite terrible to have everybody have a bunch of the world's limited supply of DRAM sitting in their own computer and being used 5% of the time," Altman said. "You want maximum utilization in the cloud if you're trying to deal with this very supply-constrained environment. And somehow many people's intuition is the opposite of that."

While OpenAI does have some open-source models, its primary business is in closed-source models that run in data centers, so he's really defending the company's business model.

OpenAI isn't exactly threatened by some laptops and gaming rigs just yet, but local AI could mean people one day paying less for subscriptions to models like GPT-6 Luna if they have the power to run them at home. Altman's talk suggests that the company would prefer that more RAM go to OpenAI's data centers rather than home computers, which also sparks enthusiast fears of not being able to own your own hardware or software, but rather rent an agent and a computer in the cloud.

Some of what Altman is saying is based on an idea that data center buildouts won't slow down. It also assumes that RAM being used in AI data centers 24/7 is more worthwhile than RAM that is used for spreadsheets, web browsing, gaming, and maybe even local AI occasionally.

The rest of the wide-ranging interview includes talk of this week's PHP vulnerabilities, sandboxing models in training, and OpenAI's research in mathematics.

A criminal investigation into the environmental damage caused by Google’s Finnish subsidiary has been launched. Tuike Finland Oy is being investigated after five complaints were received regarding the Muhos data center project, according to an official police statement published today [machine translation].

The police say that the investigation is backed by Finland’s Licensing and Supervision Authority (LVV). Google/Tuike Finland Oy is accused of suspected breaches of environmental law and environmental damage. Other lesser criminal offenses related to ambiguities and shortcomings in the drainage and water management are also being investigated.

Since Google representatives have previously publicly apologized for the unauthorized forest clearings near Muhos, which have turned 816 acres of forest into an expanse of mud, there seems little doubt it is guilty of some wrongdoing. Rather, it is the scale of the punishment and any proposed remedies that will probably be of most concern to execs from the tech giant now.

As well as the awful mess Google has made near Muhos, it seems to have also jumped the gun and plowed ahead with data center development by clearing 500 acres of forest near Otanmäki.

Though the horse has already bolted, Reuters reports that a Google rep said the company would cooperate with the ⁠police investigation and continue to focus on maintaining compliance at its sites. Maintaining?

Last week we reported that wolves had descended on Finnish villages after their habitats had been destroyed to make way for the Google AI data centers. Data center developments were already a thorny issue with locals. Potential construction and operating noise, pollution, utility competition, and environmental damage concerns were overshadowed – for some – by the presence of wild animals that had moved out of the forest.

Google sees Finland as an attractive location for data center development with its cool climate, affordable power, and government endorsement of AI projects. It has also benefitted from being able to develop in low population density areas and from a populace with a low propensity for protest. However, Google is testing the Finnish public with its trampling of the environment. The wolves, pollution, potentially 40% higher electricity bills, and other impacts weren’t helping its case, even before this criminal investigation.

Finnish government representatives have yet to comment on the case.

SpaceXSI, Elon Musk’s data center business powering Grok, is reportedly buying 418 properties in Colonial Hills, Southaven, which sits about half a mile from the powerplant that the company will use to power its data centers in the area. According to The Guardian, real estate agents have been going door to door, offering owners deals on behalf of the company provided that they do not join a class-action lawsuit over the noise pollution coming from the gas turbines that power its GPUs.

MZX, the SpaceXSI subsidiary that operates the Colossus II site, circulated an open letter to nearby residents, saying that it “requires property for its business purposes,” with Southaven mayor Darren Musselwhite saying that they will be used to house SpaceXSI staff that the company will need as it expands operations. Musk himself is reportedly staying in the area in an Airstream trailer four nights per week to personally oversee the data center build-out needed to power Grok AI.

The bigger issue here, though, is the requirement that sellers must sign an agreement that prevents them from joining the ongoing case against the company for noise pollution. The Colossus 1 and 2 sites have been in hot water since they started operating, primarily for their use of unpermitted turbines and the air pollution that they bring. SpaceXSI has since promised that it will remove all its unpermitted turbines in a year, especially as its 1.2-gigawatt powerplant comes online. However, this apparently hasn’t resolved the noise pollution issue that the community has been complaining about.

“It sounds like an airplane engine but it’s continuous, it’s incredibly loud. I’ve been woken up at 2am by it, I’ve been woken up at 3am by it. I sometimes have to just get away from it because it makes me so angry,” piano teacher Elizabeth Rose, who lives a couple of miles from the site, told the publication. One of the plaintiffs in the class-action, Jason Haley, also said to The Guardian, “There’s a high-pitched hum and then another sound like an airplane over your house. It’s been so torturous I can’t sleep; I just leave the house or turn on some music until I fall asleep completely exhausted. It drives you insane. You also lose hope because you’re being ignored or dismissed. There’s no escape.”

The company has denied all the allegations against it, saying that it has adjusted its power output and installed sound-dampening walls to reduce the impact on the neighborhood. Musselwhite also said that city monitors have not detected the site breaching the local ordinance limit of 70 decibels, although some experts say that persistent noise as low as 50 decibels could have adverse health impacts.

Opposition against the noise pollution and Musk’s purchases aren’t unanimous, though. Some residents aren’t as affected by the noise pollution, while local businesses are experiencing increased activity. Other people have also agreed to sell just so they could get away from the noise. The mayor says that the buyout isn’t intended to turn Colonial Hills into a company town. “The buyout has multiple purposes which are to provide homes for some employees and a goodwill gesture from a good corporate citizen for those who perceive the industrial development near them as negative,” Musselwhite said.

Concerns have been mounting everywhere about the environmental impact of data centers, with alarms ringing at various environmental think tanks. The Sierra Club, one of the oldest environmental preservation organizations in the U.S., has now put forth an estimate that by 2030, U.S. data center buildouts will be outputting as much carbon dioxide as the entirety of Mexico, pushing back progress on greenhouse gas reduction by 15 years, Bloomberg reports.

The Club's principal advisor for climate and energy, Jeremy Fusher, puts it plainly: "the numbers continue to sort of boggle our minds." He's referring to an estimate that with the output from pending data center builds, the output from the U.S. power sector is set to put 2.1 billion metric tons of CO2 in the atmosphere by the end of the decade, an uptick of 42% compared to 2025 figures. The researchers' previous estimates were between a quarter and half of the current projections.

An increasing number of countries and states are enacting laws demanding that new data center builds have a large portion of their power generated by green sources. However, those laws rarely apply retroactively, and oftentimes have a phase-in period. Even the Irish law on the matter demanding that new data centers be fed by 80% of renewable power only requires that that happens six years after they're powered up.

In the current AI-driven market, speed of construction is king. Roughly 40% of U.S. data centers are fed by natural gas turbines, and the situation doesn't look like it'll be changing much. It'll likely get worse, as the addition of green sources almost assuredly won't keep up with demand. Much like RAM, GPUs, and storage, turbines were all slurped up by data centers. Their prices have risen by 50%, and production is backlogged for years, with one maker stating it's opening slots for production in 2032.

Having said all that, Brendan Pierpont from Energy Innovation, another climate-concerned organization, points out that although the current path looks dangerous for the environment, it's hard to make a hard prediction of how the situation will play out. He justifies his position with common data center roadblocks: the availability of turbines, delays in grid connections, and other delay factors.

Truth be told, at this point it's hard to predict how many data centers will actually reach completion, given that requests for permits far exceed available grid capacity across the U.S., and a sizable number of buildouts are encountering strong opposition. Plus, although the AI bubble is still alive and well, concerns started to arise about the economics of hyperscalers, indicating some may go the way of the dodo — and their projects along with them.

Fujifilm plans to boost production of magnetic tape by as much as 5x as it works to keep up with demand for the data storage medium that has seen a jump in popularity amid ongoing global AI data center buildouts. The Japanese firm will spend 10 billion yen ($63.3 million) to increase capacity at its Odawara plant throughout its fiscal year 2028. The move will enable it to produce more advanced tapes with capacities of up to 100TB each, Nikkei Asia reports. The expansion of production capacity will mark the company's largest investment in tape-based storage in around 20 years. Fujifilm first began working on magnetic tape technology back in 1954 when it was used for television broadcasting equipment.

Nowadays, magnetic tape is the go-to storage option for data centers everywhere as a form of long-term data backup. While tapes lag behind solid-state drives and traditional spinning hard disks in terms of performance, tapes are hard to beat when it comes to their durability and relatively low cost for storing huge amounts of data. Magnetic tape storage can cost as little as one-quarter that of hard disks, and even less when compared to faster SSDs. That makes them the perfect solution for so-called cold data, information that is rarely accessed but could be considered mission-critical and must be stored safely. Tape storage isn't just cheaper, either, with research showing that it's also less polluting than hard disks and SSDs.

While Fujifilm will produce 100TB tapes, the Linear Tape-Open (LTO) roadmap allows for considerably larger capacities, all the way to 365TB per cartridge. By comparison, SSDs top out at around 256TB and cost considerably more.

Fujifilm is far from the only storage company looking to take advantage of the AI data center boom. Toshiba announced earlier this month its plans to double hard disk production capacity amid shortages, and the company, alongside Seagate and WD, is working towards ever-increasing capacities. Kioxia and SanDisk are also getting in on the act after detailing 332-layer 3D NAND flash in August. The pair hope to bring higher storage capacities to SSDs for data center use.

The demand for AI data center storage is so high that it's at the heart of a global shortage that shows little sign of improving anytime soon. Scarcity has seen prices skyrocket, while manufacturers have chosen to divert resources away from consumer-focused products, instead focusing on the more lucrative enterprise market.

Elon Musk has confirmed that SpaceX closed $40 billion in debt financing for graphics processing units (GPUs), with the chips expected to arrive in the first half of 2027. He disclosed the deal on X, October 25 (embedded below) while responding to investor Chamath Palihapitiya, who had warned that mounting financing pressures and delays in data center construction could leave AI computing capacity sitting idle next year.

Musk agreed with that assessment, but argued that SpaceX is better positioned than many of its competitors because it can borrow against revenue from businesses outside AI. “Pretty much,” Musk replied to Palihapitiya on X explaining that debt markets had largely shut off funding for companies without substantial revenue independent of AI data centers.

He further said SpaceX had closed $40 billion in debt for GPUs because it has investment-grade status and revenue from its Starlink satellite internet and space businesses. “That means the 200+ neoclouds are going to see guppy-level death and consolidation. Some will make it through, but not many,” Musk wrote.

It is still, however, unclear what Musk meant by “closed” or whether investors had agreed to the final debt pricing. The disclosure comes as SpaceX prepares for another enormous spending period. CFO Bret Johnsen had stated during SpaceX’s Q3 2026 earnings call that the company expects capital expenditure of $35 billion to $40 billion in Q4 “because of the additional capacity coming online this quarter.”

Musk also clarified that the borrowing was primarily callable debt for GPU purchases expected to arrive in the first and second quarters of 2027. “This is primarily callable debt for GPU purchases that will arrive in Q1 and Q2. Not a major factor this quarter,” Musk wrote.

Earlier this month, it emerged SpaceX had proposed a $40 billion debt package for Nvidia hardware, estimating that the proposed funding could cover roughly 360,000 Rubin GPUs across 5,000 Nvidia NVL72 racks. The proposed package reportedly included $10 billion in bank loans and $30 billion in investment-grade debt. Musk has now said the financing has closed, but has not confirmed the specific GPU models or quantities covered by the deal.

The spending fits into Musk’s broader push to expand AI computing capacity. In August, xAI outlined plans to increase its data center capacity sevenfold by 2027, targeting 10 gigawatts. In September, the publication also reported that SpaceXAI planned to add another 660,000 Nvidia GPUs during 2026, potentially bringing its total to around 1.44 million if the planned deployments proceed.

And GPUs are only part of the bill. Data centers need power, cooling, networking equipment, and physical infrastructure before those accelerators can do useful work, which underscores SpaceX’s latest move.

SpaceX, meanwhile, has begun selling access to its own infrastructure. In May, Anthropic secured access to SpaceX’s Colossus 1 supercomputer, which includes more than 220,000 Nvidia GPUs and over 300 megawatts of computing capacity. Later in June, SpaceX struck a separate deal with Google, which agreed to pay $920 million per month for access to approximately 110,000 Nvidia GPUs and related hardware, with payments scheduled to run from October 2026 through June 2029.

Russia has lost another data center in its ongoing conflict with Ukraine after drones hit a government data center in the suburbs of Gorodets, some 230 miles east of the Kremlin and over 600 miles away from the frontlines. According to Data Center Dynamics (DCD), the freshly drone damaged data center is operated by the Russian Ministry of Finance and was said to hold several state registries, including the Federal Taxation Service’s tax database.

“The facility is one of two key data processing centers of the Federal Tax Service of Russia operating in a geo-distributed mode,” DCD reported, citing a Russian media outlet. “The damage to this facility could potentially affect the operation of state information systems of the Russian Federation.” Ukraine has targeted Russian data centers multiple times since the start of October, which it started doing after Moscow escalated the conflict by striking Ukrainian data centers. While the Russian finance ministry reportedly operates two other data centers — one at the Federal Data Processing Center in Dubna, Moscow Region, and another in Yerzovka, Volgograd Region — losing a third of its data centers could have a major impact on government services that it provides.

A modern society’s reliance on digital services means that data centers are now critical infrastructure and have become major targets in conflict. Russia’s Yandex bot traffic collapsed after the Ukrainian drone strikes, but its disinformation bot network simply migrated to unaffected data centers to continue operating as normal. Another recent instance of a civilian data center getting caught in a conflict was in the Middle East, after Iranian forces hit AWS sites in the UAE and Bahrain in response to the United States’ bombing campaign that began in February. This has led to disruptions to many digital services in the area, which is why Amazon has advised many of its clients to move to other regions to avoid further disruptions.

Russia’s invasion of Ukraine started in 2014 when it forcibly occupied the Crimean Peninsula, but the current situation started when Moscow escalated the situation and attempted to take Kyiv in 2022. But what makes this 12-year conflict especially deadly is the new technologies and strategies that are being implemented in the front lines, including the widespread and innovative use of drones.

Nvidia is collaborating with German energy firm PreussenElektra GmbH (PEL) to explore converting old nuclear power plants into AI data centers. In an official announcement on October 21st, the U.S. chipmaker and the German firm entered a preliminary partnership to assess whether the latter's decommissioned nuclear power plant sites in Germany can be converted into high-performance AI data centers. The plan does not involve using nuclear power to power any future data centers. Instead, the project is focused on leveraging the existing power, water, and other infrastructure already present at the sites.

“Our former power plant sites have valuable industrial infrastructure and offer opportunities for new, future-oriented uses. Together with NVIDIA, we intend to conduct a sound assessment of whether and how these conditions can be leveraged for AI factories. In doing so, we are combining the responsible reuse of our sites with new economic opportunities for the regions,” says Guido Knott, Chairman of PEL’s Management Board.

The open-ended exploration will assess existing energy grid links, available land, and industrial infrastructure at the former nuclear sites to see if they meet the strict operational requirements of “AI factories.”

The move aims to find alternatives to building massive data centers from scratch amid the unprecedented AI infrastructure buildout. New data centers require well-located land, industrial-zone approvals, proximity to water infrastructure for cooling, and high-voltage grid connections, things nuclear plants already have. By repurposing these facilities, Nvidia hopes to significantly reduce cost and development timelines.

“AI factories require access to power, land and high-capacity infrastructure, and former industrial sites could provide a strong foundation for this next generation of computing,” said Markus Hacker, senior regional director at Nvidia. Together with PreussenElektra, we will assess how its sites could meet the technical and operational requirements of AI factories and contribute to Germany’s high-performance AI infrastructure.”

Germany permanently shut down its last remaining operational nuclear power plants in April 2023, following the Fukushima nuclear disaster, and is dismantling about 30 nuclear power plants across the country. PreussenElektra is currently responsible for decommissioning and dismantling eight of the plants, which are likely the sites to be assessed. The company, alongside Nvidia, has made it clear that the data center assessment initiative is strictly in an experimental feasibility phase.

No investment amounts, approved development plans, or definitive construction agreements have been finalized yet.

Texas has suspended the construction permits of 17 data center projects and disabled a web portal that automated filing for new permits after it was discovered that they were still being issued a month after calls for the process to be paused. Governor Greg Abbott asked for new permits to be halted in September 2026, but watchdog groups FracTracker Alliance and Public Citizen both published reports detailing more than two dozen permits issued to Texas data center projects, reports 404 Media.

Now, the portal that previously automatically authorized permits is unavailable. “TCEQ’s electronic filing application used for automatic authorizations has been temporarily shut down to ensure all applications are reviewed by agency staff to determine if they are related to data centers," a message on the website reads.

The portal, called STEERS, is used to rubber-stamp construction approvals. But despite Governor Abbott's call for no new permits to be approved, it seems that nobody thought to take STEERS offline. Sam Law, a researcher at FracTracker, says that it's been happily approving new projects, including those for OpenAI and Amazon that went through the system earlier this week. All 17 of the suspended permits are for stormwater construction, a process that sees construction companies raise and level the ground. It's also a process that is a vital part of the approval process for all construction projects in Texas.

Governor Abbot put the brakes on 1,800 data center projects while work was done to audit them. Concerns were raised about their power use and generation, how much water they would consume, and the impact that they would have on their communities. “Our top priority is to protect Texans’ safety and quality of life,” Abbott wrote in a statement provided at the time. “Any project that fails to comply with the requirements set forth by the PUCT and ERCOT, and by state law, must be denied connection to the Texas grid. Simply put, Texans must come first.”

The situation in Texas follows New York's year-long ban on data center projects larger than 50 megawatts that came into force earlier this year. The Responsible Data Center Development Act was signed into law as Senate Bill S10642 by New York Governor Kathy Hochul in July. It was another instance of a change in stance on data centers that is gathering momentum in the United States. While President Trump announced an "AI Action Plan" to accelerate adoption of AI tools and infrastructure rollout in July 2025, he announced a mandate forcing AI data centers to pay dividends to communities in October 2026. In Northern Virginia, Amazon bowed to a community backlash to plans to build four new data centers — going so far as to offer its land up for housing instead.

While the continued focus on AI tools helps drive demand for more capacity and more data centers, communities and local officials are increasingly concerned about the impacts they have on their surroundings. Water shortages and deforestation have become key issues affecting the communities in which data centers are being placed, as has the increase in the cost of energy.

Amazon will no longer build four new data centers on the site of the George Washington University's campus in Loudoun County, just outside of D.C., and will instead repurpose the majority of the 122-acre space for mixed-use and residential development following significant pushback from the local community. It will also donate 25 acres of the property to Loudoun County to be used for a community arts center and new recreational facility. On top of all that, it'll even give $5 million to the county's fund for housing, reports The Washington Post.

That's quite the turnaround compared to plans to build in a county that is already dubbed "Data Center Alley" because it has the largest concentration of data centers on the planet. But residents' concerns about the noise generated by such data centers and the incredible amount of power that they require to function have seen the county's approach to new data center buildouts change of late. Amazon first submitted its proposal to build the new facilities in July, and Loudoun County Board of Supervisors Chair Phyllis J. Randall and Amazon met in September, where it was put to the company that it may be better to find an alternative use for the land.

Josh Weissman, Amazon’s vice president in charge of data center development, said the company had jumped at the chance to purchase the former university campus when the opportunity arose, but added that the company considered the county’s “pointed concerns about data center development at this particular site.” Weissman added that “While we see the property’s legal development potential differently than the County does, we concluded that the best outcome for all is to come together on a shared vision for this campus rather than engage in costly litigation.”

The building of new data centers has always been a necessity given our growing use of cloud-based services and storage systems. But the recent boom in AI use has seen demand for additional, and larger, facilities ratcheted up to new levels. But with that demand comes opposition, with local communities wary of the impact these mammoth facilities can have on the areas in which they live, including spikes in the price of electricity. Officials have begun taking steps to try to grease the wheels of data center construction. Just days before this news broke, President Donald Trump called for data center builders and AI hyperscalers to pay local communities a dividend amid energy-related price increases. In a social media post, Trump went on to say that data centers would be required not only to make their own electricity but also to give or sell excess energy back to the local community. Water shortages and deforestation have also become key issues affecting the communities in which data centers are being placed.

Back in Loudoun County, officials pointed to what could be achieved when choosing dialog over confrontation. “By choosing dialogue over reflexive disagreement, Loudoun County and Amazon created an opportunity for the County,” Randall reportedly said via statement. “This agreement provides new places for our Loudoun residents to gather, create, and play. And, most importantly to me, we will provide more attainable homes for working families.”

Lava, an Israeli security firm, says that it has been able to identify 32 Nvidia Blackwell Ultra B300 chips within the network of China Telecom, a state-owned company that provides telecommunications services within the country. These same chips have already been barred from export by U.S. export controls, a report by The Information notes, which begs the question of how they have turned up in a Chinese data center. Such chips can only be sold to whitelisted buyers who agree not to export them to sanctioned parties, including those in China.

However, banning companies from buying Nvidia's high-performance GPUs and AI data center chips hasn't always worked. In 2025, it was reported that Chinese companies smuggled $1 billion worth of Nvidia AI chips into the country in three months. At the time, the B200 chips were the most sought-after, but some distributors were already gearing up to sell B300-series chips when they became available.

While the latest finding shows that the proliferation of export-controlled chips into China continues to be an issue, Nvidia has downplayed the report but didn't deny the findings. "Governments, researchers, and media have all found insignificant diversion of controlled US products, less than one half of one percent in recent court cases — a drop in the bucket compared to the ocean of domestic compute China already has," an Nvidia spokesperson told Tom's Hardware. "With tens of millions of GPUs in service in approved countries, this report again demonstrates that smuggling is a nonfactor, blocked by governments on both sides of the globe."

It's unknown how B300 chips found their way into China Telecom's server racks, nor can Lava confirm exactly where within the company's network they are located. But they are there, with Lava identifying them via Nvidia's own DCGM Exporter tool. The software can be installed by data center operators and is used to monitor server health and performance, and while designed for internal use, it can oftentimes be accessed beyond the data center's network. Oddly, it doesn't always require password authentication, either. Nvidia previously confirmed it was working on a tool that would allow its chips to be located, but it is yet to be released.

It's easy to see why Chinese firms are keen to get their hands on the Blackwell Ultra B300. Announced in March of last year, the data center GPU is one of the most powerful on the planet, offering huge performance gains with 50% more memory and FP4 compute than the previous-generation B200. At the time, Nvidia touted that the chip was "built for the age of reasoning," while name-checking more sophisticated AI LLMs like DeepSeek R1. At the time of its announcement, the B300 was expected to ship by the end of 2025, and it didn't take long for it to become hot property. In August 2026, nine people were indicted in Taiwan over the illegal export of B300 GPUs to China. This came just days after Supermicro confirmed that it had fired several employees following an investigation into smuggling. The complicated smuggling process included the use of a Taiwanese server trading firm called Flying Tiget Tech — a firm with whitelisted status for GPU purchases.

Nvidia's chips have proven so vital to AI data centers that demand for even its older silicon has caused skyrocketing prices. The company's A100 servers have tripled in price, despite being five years old, and can cost up to $82.,000. The ongoing crackdown on Nvidia chip smuggling has even seen some resort to retooling gaming GPUs to run AI workloads where possible.

Google’s Finland country manager has apologized for unauthorized forest clearings near Muhos, which have turned 816 acres of forest into an expanse of mud. The American search and AI giant has also been told to cease work where it has cleared roughly 500 acres of forest, 55 miles southeast of the village of Otanmäki. Now The Guardian reports that wolves, who used to live in the forests, are roaming into villages neighboring these man-made muddy wastelands.

Local anger continues to simmer over the damage done to Finnish forests, as Google seems to have ridden roughshod over regulations in the hurry to build new AI data centers in Finland. However, the report suggests that it isn’t just environmentalists who have a beef with Google. Some families say wolves and other wild animals have entered the countryside villages. This would be a serious human safety concern.

“We had wolves coming into the village because all the wild animals moved out of the forest,” explained Anni Ranta, 53, whose family has lived in Otanmäki for generations. Ranta also fears the U.S.-owned data centers will make Finland an even juicier target for Russian aggression.

Finland is far from the first place targeted by AI data center developers to complain about environmental effects. This latest report also cites locals grumbling about construction/factory noise, streams turning brown, and rare birds being pushed out, among other unwanted side effects. We’ve seen similar concerns voiced across the U.S. and in other countries.

Google’s flouting of environmental policy has been unprecedented and clear, a representative of the Finnish Association for Nature Conservation (SLL) indicated to The Guardian. The Finnish Licensing and Supervision Authority (LVV) seems to agree, as it issued an order for Google to stop development near Muhos and Otanmäki. Nevertheless, a Google spokesperson was upbeat. “We will fully comply with LVV’s requests in this matter,” they said. “We look forward to maintaining an active, open dialogue with all authorities to strengthen our partnership with Finland and local organizations.” Google must be hoping the touted business benefits of the data centers will sway the government and/or regulators.

We last reported on these in-development data centers in Finland earlier this month, after the Finnish government (LVV) gave notice to Google to stop its construction activities in Muhos and Kajaani (near Otanmäki).

President Donald Trump has issued a call for data center builders and AI hyperscalers to pay communities a dividend in order to overcome energy-related price hikes, or even to reduce prices, in order to make data centers more palatable to local communities. In a post on Truth Social post, he warned that data centers would be mandated not only to make their own electricity, as he has previously stated, but also to give or sell excess energy back to the local community.

Local communities all over the United States have protested against AI data centers, often owing to their outsized impact on local communities. Rising electricity costs, noise pollution, water shortages, air pollution, deforestation, and more are the things that small communities have to deal with because of these companies.

To ease the pain of local communities, Trump is demanding that these large companies provide the communities they affect with something that will hopefully make their presence more acceptable. The Trump Administration is keen to stay ahead of China in the AI arms race and is looking for solutions to win back some favor by making the data centers pay something back to the affected communities; how that will practically transpire is yet to be seen. Considering the electricity-focused post, the dividend might be in the form of partially paying rising electricity costs to alleviate a pain point that directly affects many homes within the vicinity of data centers. The Trump Administration had ordered these companies to make their own electricity, as we’ve covered previously. Excess electricity from their generation could be used to further lower the electricity cost for the locals, including a recently passed act that orders data centers to shoulder the cost for grid upgrades to avoid taking it from the taxes of local communities.

“Data Centers that were highly unpopular in a certain area have become extremely popular when a financial deal is struck.” Trump further says in his post, aiming to have more data centers built, “MAKE DEALS, and get them built!!!," the President said in closing.

Whether the dividends will be enough to win over communities is yet to be known, as the economic aspect is only one facet of the problem. There are still plenty of problems to face; enormous water consumption, ecological damage, and pollution are still major challenges that make AI data centers a hard sell anywhere they’re located.

A Norwegian company building fiber networks and data centers across Europe is looking for a majority partner to help fund and build a transatlantic undersea cable that would connect Canada with Europe directly, bypassing the United States. The project follows worsening relations between Washington and Ottawa, with the Canadian government looking to reduce direct reliance on its immediate neighbor, Bloomberg reports.

Project Leif Erikson, if it goes ahead, would stretch more than 2,800 miles, starting at the southern tip of Norway and ending up at Happy Valley-Goose Bay in Canada's eastern province of Newfoundland and Labrador. From there, it would continue underground over Canadian territory, passing through Quebec before reaching Montreal.

According to Bulk's co-founder, Peder Nærbø, the link would be part of a "sovereign cable system" for Canada. The project was touted to investors at a global investors summit, arranged by the Canadian government, in September, where the project claimed to provide "direct, secure, high-capacity data transmission for AI, telecom and hyperscale markets."

Canadian Industry Minister Melanie Joy, meanwhile, met with Nærbø earlier this year, and although the department declined to comment on Leif Erikson itself, it told Bloomberg that security and resilience for Canada's infrastructure are "key priorities" for PM Mark Carney's government. Only a few undersea cables currently connect Canada without passing through the United States first, and none by a company owned by Canada. The biggest, called Topez, is owned by Google owner Alphabet and connects Canada with Japan.

As for Bulk, it has, up until now, laid more than 6,200 miles of cabling locally and at sea. Bulk's goal is to "unlock the Nordics' potential for... renewable energy to be utilized for digital infrastructure." Nærbø believes that "Canada lacks resilience" but "the game-changer was when the current situation with the U.S. escalated."

Leif Erikson has backers, according to the report, but the EU and Canada still need to plan and sign off on the project and its route. Additional financing will be needed to get started, too, with it expected to cost more than $620 million to build. Bulk itself has significant Canadian investment, with 65% of the company owned by a real estate subsidiary of Canada's Sun Life Financial Inc., with several Canadian pension funds also invested.

Leif Erikson is a project that Nærbø has sought to kickstart for the last five years without success. With geopolitical pressures and old alliances fraying, however, new opportunities for its development have given it steam. To that end, Canada and Europe intend to endorse a statement in Montreal on October 29 that the two will look for "the best way to enhance digital links between Europe, Canada and Asia through new subsea cable infrastructure."

It won't be without challenges, however, including a 1,180-mile underground route through Canada that'll require the backing of Indigenous communities. Meanwhile, other European efforts include an undersea cable project at the North Pole to link Europe with Asia to bypass Russia and the Middle East.

More than 300 local data center construction moratoriums haven’t had a massive impact on planned buildouts, with only three projects at about 2.3 out of 21 gigawatts (GW) of data centers in moratorium areas likely to be delayed, with a total of 38 GW projected to come online next year, SemiAnalysis shared on X. Per-project audits of the moratoriums indicate that they tend to miss the site, the permit, or the timeline.

The crux of the report is that, while everyone thinks the moratoriums are killing the U.S. buildout, its forecast has “barely moved over the last 12 months.” This forecast includes 38 GW of U.S. data center IT capacity in 2027, which is more than double that of 2026. Of that, 22 GW is under vertical construction, and the other 16 GW is mostly financed, with site work being done. The enacted local moratoriums cover about 20 GW with only about 1.5 GW of it delayed; New York’s pause adds 0.8 GW more. “We think moratoriums COULD delay the buildout, but they just aren’t yet at scale.”

The firm defines a moratorium as a “temporary legal pause,” and almost always a “freeze on accepting new applications.” More than 400 local instruments are tracked across 17 states, the enacted ones including a few permanent zoning bans, as of September. Four states acted in under two months this summer: New York, Texas, Pennsylvania, and Oregon. To test each active restriction against the records of thousands of projects, the firm’s analysts worked alongside “an army of state-of-the-art research agents.”

Several reasons are given for moratorium ineffectiveness, starting with the fact that they don’t revoke existing approvals. Short ones may expire before later projects need approval, and litigation or equipment lead times may already set the schedule. A moratorium moves the delivery date only if every condition holds, including that it reaches the parcel, is in force, overlaps the timeline, and covers the use in a way the project can’t work around. Capacity already under construction or fully permitted is also out of reach.

For example, Michigan has 45 enacted moratoriums but effectively zero pipeline exposure, and a Google campus there sits on industrial land already zoned for data processing. The firm says SpaceX’s planned data center in Brownsville, Texas, is on land the city removed from its limits and thus out of reach of the city’s moratorium. For a moratorium that did work, a township adopted a pause before NorthPoint, in Pennsylvania, filed for its special exception, with local approval pushed to next year. The company is still pursuing it.

The 2.3 GW figure is also incomplete, as it leaves out the Texas Commission on Environmental Quality (TCEQ) pause. Texas Gov. Greg Abbott told the TCEQ to issue no permits for data center projects. Law firm Bracewell says this applies whether a project is in the Electric Reliability Council of Texas’s (ERCOT) queue or not, and expects the pause to last until ERCOT finishes its review in mid-December. Pauses like the TCEQ’s, and how they play out, could impact buildout forecasts.

The post quotes another by The Economist’s Archie Hall, from a trip that covers Michigan and Ohio, with the reporter stating that data center moratoriums are “much more bark than bite” for the time being. He puts the bigger risk at deterring newer projects, though he guesses that affects where data centers end up more than whether they get built. His impression is that politics also plays a role, and according to the firm’s poll in August, more Americans have a negative rather than positive impression of data centers.

The analysis concludes that moratoriums would need to be “materially more widespread and restrictive” to delay the buildout. On the firm’s weekly podcast last month, Jeremie Eliahou Ontiveros, one of the report’s authors, said the firm would be worried by a strict, state-level moratorium that blocked even off-grid projects, citing Texas as an example. New York Gov. Kathy Hochul, in addition to Abbott, also found ways to bypass their legislatures for pauses they control and can revoke. The next move to watch is the TCEQ’s compliance report to Abbott’s office, due Monday.

Over the past week, Ukraine attacked some of Yandex's biggest data centers, disrupting the operations of various Russian national applications and metropolitan digital services. Russia, which has itself targeted Ukrainian data centers in recent weeks.

Worse yet for Russia, its online disinformation activities appear disrupted by these attacks. Although not as catastrophic as some initial reports claimed, a new report from Cloudflare claims that the bot traffic merely migrated to other hosts.

A new target

Since Russia's full-scale invasion of Ukraine in 2022, the war has moved through several phases. From the pushback of Russian forces in 2022, to the rapid deployment of FPV drones in 2023, to the mid-range drone strike campaigns both sides have focused on since 2025. Over that time, the targets have changed. Ukrainian supporters published "bingo cards," with various refineries and warehouses listed, often superimposing fire graphics to showcase when they'd been damaged or destroyed.

There are tens of major targets across different industries and networks; there are only a handful of Yandex data centers. Although there is a more expansive list of data centers from other providers that support the Yandex cloud network, most are small, only offering a few megawatts of compute capacity.

Considering that targeting just three of Yandex's facilities brought several key digital services such as shopping and taxi apps to their knees in major Russian cities, it may not take much for Ukraine to severely hamper the modern digital elements of Russia's economy and society at large.

While Ukraine isn't immune to similar attempts at disruption, its already-hardened networks and expansive list of allies make it far more resilient.

Friends in the right places

While overreliance on digital services from American companies like Google, Meta, and Amazon has its own drawbacks, in Ukraine's case, it's been a far more resilient strategy than Russia's. Russian attacks on Ukrainian data centers have caused some disruption, but much of Ukraine's digital infrastructure runs on Amazon's AWS cloud platform. That includes distributed hosting, resilient service structure, and AI integration.

That means that when Russia hits Ukrainian data centers, the services can be quickly migrated to another data center, perhaps located in another country entirely. Amazon has the world to fall back on if necessary. Ukraine also has allies in many other technologically advanced nations that can provide digital infrastructure and expertise, including access to the latest AI cybersecurity models.

In contrast, Russian internet and services are far more inward-facing. It has attempted to cut its population off from the outside world by banning Western services like WhatsApp and Telegram, pushing its citizens towards government-run applications like Max.

As the Russian-version of Google, too, Yandex has its fingers in a similar number of digital pies. It runs online search, mapping, online shopping, cloud storage, ridesharing, food ordering, and more. When its systems are disrupted, services that people have come to rely on in a modern digital economy break down, and there are far fewer alternative options.

Russian allies like North Korea, Belarus, and Iran can't offer much in the way of digital infrastructure support. Even China, which has been aiding Russia throughout the conflict, can't help much here. While it is making strides in building out its compute network for AI infrastructure, it runs a relatively walled-garden form of the internet. Integrating Russian applications with that would be complex and time-consuming.

All of this suggests that if Ukraine continues to hit Russian data centers, there's a real potential that Russian citizens and its military could lose access to digital services that have become ubiquitous in 21st-century life.

The effect that would have on Russian society and its war-fighting goals remains unclear, but it would have an impact.

Russia's AI capacity

Losing its most capable data centers alongside sanctions that prevent it from taking advantage of the AI boom that's buoying many other global economies could severely hamper Russian efforts to modernize its economy.

While the U.S., EU, and China are building out hundreds of billions of dollars worth of data centers to power the next generation of artificial intelligence models, Russia is struggling to hold on to what it's got. With major data centers now in Ukraine's sights, including those scheduled for expansion, Russia could see its digital future wiped away in a matter of months.

With even medium-tier data centers often taking years to build, even with ready access to the hardware to power them, Russia could be on the cusp of falling decades behind its contemporaries.

Bots are Down, But Not That Much

One highly publicized impact of these strikes on Yandex data centers has been a reduction in automated online activity from Russia. While that does appear to have reduced bot traffic slightly, particularly around support for far-right parties, it may not be as impactful as some initial reports claim, as Cloudflare has published a report stating that many of the bots have merely migrated to other hosts, reducing, but not eliminating, the bots themselves.

Journalist Ilya Lozovsky posted a lengthy explanation on Twitter/X about why Russian bots haven't been as disrupted as some have claimed, highlighting that bot activity hasn't fallen off as much as legitimate web crawlers and search functions at Yandex.

But while Russian bots haven't gone dark, there are several reports of reduced online activity among far-right groups, particularly those supporting prominent far-right politicians.

Posts supporting the AFD in Germany don't appear as impacted as initial claims suggested — the chart doing the rounds has clear errors in it, and German tech outlet Rohtext claims online discourse in general went down, rather than specific parties being impacted.

What these strikes have done, though, is highlight a clear and present weakness in Russia's modern society, military, and economy. If Ukraine can continue its strikes against Russian data centers, it may well bring key public services to their knees, as well as cripple Russia's ability to compete in modern digital services beyond the end of this war, further inhibiting the Russian economy in the future.

Russian bot traffic has dropped dramatically in the past few days after Ukraine knocked out Yandex’s biggest data centers, with a third attack happening last night. Cloudflare reports that bot traffic from the Yandex network has dropped to just 0.4% of the maximum number of HTTP requests recorded in the past two weeks. It initially dropped to around 57% on Oct. 8, which happened when Ukraine hit a Yandex data center housing two supercomputers in Sasovo, Russia, located around 195 miles to the southeast of the Kremlin. A second, bigger Yandex campus in Kaluga, about 220 miles to the west of Sasovo, was knocked offline yesterday.

These attacks are apparently Ukraine’s response to Moscow’s strikes on its own data centers, some of which resulted in the outage of some of the most popular piracy websites.

Yandex is Russia’s equivalent to Google, and its Yandex Cloud service status reports that at least 84 services are currently offline. Global internet monitoring firm NetBlocks said that Yandex’s network traffic has dropped to just 30% for Yandex Cloud and 63% for Yandex LLC, most likely resulting in slower internet services or complete disruption for the average Russian. Unfortunately, the ordinary Russian wouldn’t be able to use high-speed satellite internet as a backup, as Starlink does not work in Russia, while the competing Rassvet constellation has had a good number of its existing satellites fall back to the surface after launch.

These incidents are a result of the ongoing conflict in Ukraine, which escalated dramatically in 2022 after Russia’s failed attempt to capture Kyiv. However, many observers say that the war actually started in 2014 when Russian forces seized Crimea and occupied parts of eastern Ukraine after the fall of the pro-Russian government that year.

The hostilities, which have dragged on for years, saw the introduction of new technologies and strategies, specifically in drone warfare. It began with the wide deployment of commercially available drones on the front lines, typically for reconnaissance, dropping explosives, and kamikaze-style attacks. However, it has since evolved into long-range warfare, with specially built drones able to reach distances of over 1,500 miles and destroy Tu-95 strategic bombers situated well behind the front lines.

The drone strikes on data centers are relatively new developments. While data centers on bases are often considered valid targets, Russia’s justification for the attacks on civilian data centers was that Ukrainian military intelligence used them. This isn’t the first time that civilian data centers have also been hit by drones and missiles. AWS sites in Bahrain and the UAE have been hit multiple times since Washington started its bombing campaign against Iran, with the company telling its clients to move their data to sites outside the region, as it doesn’t have a recovery timeline.

The U.S. Senate has been conducting a year-long investigation into AI data centers, and its findings suggest that some developers have been misleading the public about the costs and benefits that these projects bring to their communities. According to Time, data center developers have routinely pointed to the number of construction jobs that projects bring to neighborhoods, but many have allegedly refused to share the number of permanent positions these projects bring to the neighborhood once they’ve reached operational status. A few companies that did give numbers point to a ratio of one permanent position for every megawatt of capacity, meaning a 100-megawatt development, which consumes around the same amount of electricity as 100,000 homes, would only hire 100 people.

“Congress must hold Big Tech accountable so these companies pay their fair share,” Senator Elizabeth Warren (D-Mass.) told the publication. Senator Chris Van Hollen (D-Md.) also said, “This report lays bare what we have long known: working Americans and local communities are footing the bill for Big Tech’s massive expansion of data centers, while these companies continue to operate without transparency.”

Aside from the lack of permanent jobs, the authorities also looked at the tax incentives these companies have been getting from local and state governments. The investigation revealed that although property-tax breaks catch critics’ attention during the application process, sales-tax exemptions on computer equipment are reportedly more lucrative, especially as these sites maintain, replace, and upgrade their hardware. The report estimated that 39% of a 1GW data center’s spending is allocated towards expensive GPUs, so if the data centers do not deliver the jobs and economic benefits they’ve promised, then the state is losing out on a lot of revenue.

The investigation also revealed that the surveyed companies — Amazon, Google, Meta, Microsoft, CoreWeave, Digital Realty, and Equinix — don’t want to cover the costs of the additional infrastructure needed to serve their data centers even though they previously promised to “pay their own way.” While the companies said that they were willing to pay for direct costs, they argued that they should not be responsible for the larger investments that could potentially benefit other customers, such as new power plants and transmission lines, even though these were primarily driven by their power demands.

There have already been efforts to legislate the ratepayer protection pledge that data center developers signed earlier this year, but it has failed because senators thought that it was “toothless.” This investigation shows that Congress is not giving up and that it’s conducting an even broader effort to regulate the infrastructure behind AI, especially with regard to its massive power consumption and other issues that residents living around them have been protesting against.

Hot on the heels of a Ukrainian drone strike on a Russian Yandex data center in Sasovo that caused a major outage, a Yandex facility in Kaluga has been struck in a similar fashion. The huge data center was so badly damaged that Russian state media admitted that multiple modules had been removed from service completely.

United24 Media reports that the facility is located in Grabtsevo industrial park and was designed to be the largest Yandex data center in all of Russia. It only entered service in the second half of 2023, lasting three years before this incident.

The Kaluga data center is a sprawling complex that covers around 1.4 million square feet (130,000 square meters). It's so vast that its 63-megawatt power capacity is around 50% higher than some other data centers owned by Yandex. With more than 3,800 server racks, it's perhaps easy to see why it became a target for Ukraine's drone operators. If, of course, the facility was the intended target for the attack — something yet to be confirmed.

As for what all those server racks actually do, it's reported that Yandex's internal services are housed at the facility. It was ultimately built to allow Yandex to expand its cloud offering.

It's as yet unclear how this drone strike is affecting everyday Russians, but the previous Sasovo strike also impacted Yandex Cloud. At the time, users reported problems with Yandex Mail, Disk, Documents, Pay, and Market despite Yandex claiming that key services remained up and running.

Russian state media TASS confirmed the severity of the attack, saying that “Several modules of the Yandex data center in Kaluga have been completely taken out of operation as a result of a UAV attack, [Yandex] reports.”

This incident is just the latest in a run of attacks on digital infrastructure both in Russia and Ukraine. Russian missiles struck a Ukrainian facility earlier this month, taking the world's second-largest torrent tracker offline in the process.

Google subsidiary Tuike Finland Oy has just received a notice from the Finnish Licensing and Supervision Agency, or LVV, ordering it to stop construction on two data centers. According to CNBC, the sites in Muhos and Kajaani cannot proceed until environmental impact assessments have been completed, which have been under investigation for allegedly clearing over 740 acres of Finnish forest for the project.

“We understand the concerns and have fallen short of our own high standards in this instance,” a Google spokesperson told the publication. “We acted in good faith under the Forestry Act and ran nature surveys and put in place measures to protect high-value areas. As we continue our long-term work in the region, we are advancing our landscaping and biodiversity plans, which include tree planting across 130 hectares at Muhos.”

Finland is sometimes called the “Texas of Europe” when it comes to data centers because of its readily available supply of land and electricity, which have been some of the major bottlenecks that data center developers have been facing. Even though the epicenters of the AI data center infrastructure build-up are centered around the U.S. and China, Europe is trying to catch up with its massive $30 billion investment and a target of 100,000 AI GPUs for each site. SoftBank also plans to invest $87 billion specifically in France because of its nuclear-powered electricity grid.

However, other European nations are struggling with the demand for this power-hungry infrastructure. Denmark has paused new data center connection requests as demand hit 60GW — more than eight times the country’s peak power consumption of around 7GW — while Amazon’s data center projects on the continent are facing delays of up to seven years as they wait for new connections.

These stop-work orders aren’t related to electricity supply, though. Instead, the concerns stem from the alleged cutting of trees without the go-signal from the government. The LVV ordered the subsidiary to stop removing trees and topsoil, as well as cease activities such as excavation, quarrying, crushing, soil mass transfers, ditching, and road and storage construction. “In our opinion, the measures in question will change the environment and cause impacts, the identification and assessment of which are among the key objectives of the EIA procedure,” LVV environmental department head Tommi Muilu said in a statement. Still, the subsidiary can proceed with planning, measurement, soil surveys, and other activities that would only have minimal impacts.

U.S. data centers have increasingly faced resistance from residents and communities around them, with 45 projects worth $68 billion blocked in the second quarter of 2026. We haven’t seen the same level of high-profile opposition in Europe, but this is likely because of the stricter permitting and review processes in the region, where concerns are usually raised and addressed even before construction begins.

Data center energy demand is shooting constantly upwards. By 2030, some 945TWh of electricity is expected to be used to meet AI’s demand; the International Energy Agency believes — that's as much electricity as Japan uses today. The energy crunch is inescapable, and the industry’s answer has largely been to focus on the physical: build more efficient GPUs and secure more power supply from an already tight grid. But is that myopic focus on hardware actually worth it? Instead, could most of the work running on those machines be cut back to save power?

Average power usage effectiveness (PUE) has barely changed for six successive years, according to the Uptime Institute’s 2025 survey. PUE can show whether cooling and power systems are wasteful, but doesn’t account for whether the software running on the servers is doing useful work. And given servers account for around 60% of electricity demand in a modern data center, while cooling ranges from about 7% in an efficient hyperscale site to more than 30% in a less-efficient enterprise facility, focusing on eking out more efficiency from the software side of the equation seems sensible.

“The answer to what seems to be the core question – ‘Can software or algorithms meaningfully help with power and energy?’ – is yes, 100%,” said Jae-Won Chung, a PhD candidate in computer science and engineering at the University of Michigan and researcher with the ML.Energy initiative, in an interview with Tom’s Hardware Premium.

Chung suggests computing systems should be seen as a stack with hardware at the bottom, backed up by systems software, algorithms and applications. The hardware end of that equation is the hardest bit to tackle: chips are slow and costly to replace. But the other three layers are easier to change, and any gains in efficiency made there can compound.

ML.Energy’s tests of the Alibaba Qwen 3 235B A22B Thinking model suggest that running inference in FP8 – a slightly lower-precision format – consumed a third less energy than bfloat16 versions on problem-solving tasks. Chung’s own Perseus training optimiser identifies parts of a large-model training job that have less work to do, then slows them so they finish alongside busier parts. The system cut training energy by up to 30% without reducing throughput or changing the hardware.

This isn’t a tweak around the edges being made by tinkerers: some of the biggest GPU firms are doing something similar. Nvidia’s Blackwell power profiles fine-tune everything from GPU compute and memory frequencies, power limits, NVLink states and cache settings to fit the workload. Nvidia believes that can save up to 15% of energy while retaining 97% or more of overall performance, allowing power-constrained facilities to run more GPUs, raising throughput by as much as 13%.

‘The loudest rack in the room’

Chetan Visrolia, business development manager for data center infrastructure at SHI, said in an interview with Tom’s Hardware Premium that the most obvious waste he encounters is often legacy equipment supporting old code. “It is always the lowest-hanging fruit for quick savings,” he said. “It can easily be found in any data center, as it is the loudest rack on the floor.”

Cloud instances can also be made a smarter size, and AI prompts that are repeatedly used can be stored in cache rather than recomputed on the fly. There are also efficiencies to be made by getting small models to handle routine requests while holding back the larger ones for harder tasks. At the same time, better batching and compilers can raise accelerator utilisation, and crunching down prompts while throwing limits on outputs can also tamp down the number of tokens processed.

Software can also change when and where electricity is drawn. Sophie Hall, a doctoral student at ETH Zurich’s Automatic Control Laboratory who has studied workload shifting across Google’s global data-center fleet, argues in an interview with Tom’s Hardware Premium that electricity consumption isn’t necessarily the main problem.

“It’s more like: when do they use it, where do they use it, and how is it interacting with the grid?” she said. If batch jobs aren’t time-sensitive, they could be delayed until local demand falls or routed to a region with spare capacity and lower-carbon electricity. Hall’s research uses day-ahead planning and real-time scheduling to respond to grid signals while preserving performance guarantees.

Taking that kind of action was less necessary when hyperscalers built way more capacity than they actually needed and the grid wasn’t a bottleneck, Hall said. But the massive demand for AI has changed the equation.

There are limits, though: individual countries’ data sovereignty rules can prevent a workload from crossing borders, while moving large datasets across networks is something that inference providers may be wary of in case things go wrong. “The data needs to literally, physically, geographically move to a different location,” Hall said. “That’s not that easy.”

The scale of the data center demand powered by AI means that software isn’t a silver bullet or a substitute for better chips and new electricity infrastructure. But it can act as a control layer that squeezes out every last drop from those increasingly expensive assets, while also helping shift demand to places where there is more capacity.

Jevons ’ paradox is also at play, and needs to be considered when thinking about making systems run better. Efficiency doesn’t reduce a facility’s electricity use if each watt saved gets used up generating even more tokens. “Power is the core bottleneck in AI data centers,” said Chung. “We really want to make the best use of every watt we consume.”

Ukrainian drones hit a major Yandex data center in Sasovo, Russia, causing a massive fire and forcing the company to shut down the facility, according to Reuters and Meduza. Many services using Yandex Cloud are down following the attack. The site reportedly houses two of Yandex's three supercomputers used for artificial intelligence development, though it remains unclear whether the systems sustained any damage, as for now their fate is unclear. Russian missiles struck Ukrainian digital infrastructure earlier this month.

The Sasovo facility, one of Yandex's five major data centers, accommodates tens of thousands of servers, including two Nvidia A100-based supercomputers named Chervonenkis and Lyapunov. Chervonenkis, the company's most powerful system, was once ranked as the world's 19^th most powerful supercomputer in the Top 500 list in 2021, but by today's standards its 21.53 FP64 PFLOPS is enough only for the 101^st place in the list. Due to Western sanctions, Yandex can no longer upgrade the system. Both machines have been used to train AI models, including YandexGPT, Russia's leading large language model. Yandex's third supercomputer, Galushkin, is located in Vladimir.

So far, the company has not disclosed whether its supercomputers were affected by the strike, though judging by the massive fire caused by the drone incident, the systems' chances of survival are low. The potential damage to Nvidia-powered computing infrastructure could be extremely painful for Russia's AI industry, which is already constrained by restrictions on exports of advanced AI accelerators and servers to Russia. If the machines are indeed destroyed or damaged, many of its elements may be lost forever, and Yandex will be forced to use Chinese infrastructure for future model training.

The attack also disrupted Yandex Cloud infrastructure, so the company is recommending customers use alternative zones. Although Yandex claims that its key consumer services remained operational, users reported problems with Yandex Mail, Disk, Documents, Pay, and Market. Other affected services reportedly included Cian, Litmarket, Russian Railways, Spartak Moscow, cryptocurrency exchange Rapira, Astral, and the 1C business solutions. Corporate systems operated by T-Bank, Ozon, and Wildberries also experienced disruptions, according to Meduza. A Moscow-region bus operator suspended online ticket sales, while Detector404 registered 58,268 complaints over 24 hours. Many Russian media outlets, including state propaganda RIA Novosti, Vesti, and Interfax, were also down following the attack on 'infrastructure partner,' reports RBC.

It is noteworthy that the Yandex data center occupies former premises of the Sasta machine-tool factory, where Yandex established operations in 2014. Sasta manufactures industrial equipment supplied to various Russian manufacturers, including defense-related enterprises, and is therefore under Western sanctions. Regional authorities reported a fire at an industrial facility following the attack, but it remains unknown whether the intended target was the factory or Yandex's computing infrastructure. Still, no matter the target, Ukraine's sanctions against Russia's economy and military complex seem to be generally more efficient than those imposed by the West.

The incident follows Russian strikes against Ukrainian telecommunications and computing facilities. Since September 23, attacks have damaged or destroyed data centers that belong to at least eight Ukrainian companies, including Virtual Systems, MiroHost, CosmoNova, DataGroup, and CityHost. Among other websites, popular torrent trackers 1337x and NordicQuality temporarily went offline after servers that hosted them were damaged.

Ukraine is consequently moving some digital infrastructure underground, while Russian businesses are discussing deployments in countries considered friendly by Moscow, which is probably limited to just a few states, including China.

EV charging firm Xeal plans to use its existing U.S. network to deploy the “world’s first edge inference compute network using idle EV charging capacity.” The big idea is that it already has 1,600+ locations and 200MW of permitted, installed electrical infrastructure, so it aims to “deploy over 100,000 NVIDIA GPUs” at these sites. Thus it has developed Laitent Pods (pictured), with each containing up to 48 Nvidia Hopper or Blackwell Ultra GPUs. We hope charging location security will be adequate to keep thieves away from these cabinets, as they are probably worth $2M+ each.

Nikhil Bharadwaj, co-founder and CEO of Xeal, explained that Laitent Pods can deliver inference compute today while new data centers are still being built. “Laitent is the fastest way to bring new compute online, tapping already permitted and grid-connected, yet idle energy capacity, located right where it's needed for ultra-low latency, with no water hookup requirements,” highlighted the exec. Grid interconnect is said to be one of the trickiest things for those looking to bring new compute online, but Xeal’s EV charging sites, which “typically operate at less than 10% of permitted capacity,” can deliver processing power via a distributed network across a metropolitan area.

An Nvidia rep also praised the upcoming Laitent Pod network for being an innovative way to bring low-latency inference closer to users while making better use of infrastructure that has already been built.

In the rendered image from Xeal, a Laitent Pod looks quite a lot like a broadband fiber cabinet that you might have on your street. A pod will boast the following specs/features:

• Powered by idle EV charging capacity
• Up to 48 Nvidia Hopper or Blackwell Ultra GPUs
• Self-contained independent cooling, no water hookup required
• Quiet operation at less than 65 dBs (comparable to a washing machine)
• Fixed rental or ancillary revenue for site owner
• Adds up to $1 million in property value for little-to-no investment
• All power bills covered as a tenant pass-through
• Potential savings for EV charging customers
• Built to withstand the outdoors, NEMA 4 enclosure
• Installed and running in hours

It claims that the location of its pods around metro areas, a feature dubbed Metro Edge, offers sub-20ms latency, a significant speedup over inference located in a distant data center.

We’ve previously reported on EVs and distributed computing, but in a slightly different context. Last year we discussed Elon Musk’s proposal to make use of the untapped asset of the millions of Tesla cars out there. He called it “100 gigawatts of inference” that could be used when the vehicles are “bored.”

Meanwhile, we must remember that EV charging stations are already being targeted by thieves. We reported about a ‘plague’ of electric car charger cable thefts driven simply by the copper scrap value of the hefty cables used at the sites. Imagine the security that would be required if/when Laitent Pods are deployed, stuffed with extremely expensive GPUs. 48 Blackwell GPUs would be worth around $2,000,000 according to 2024 figures – so surely much more by now.

The first Laitent Pod will be online by the end of the year, says Xeal.

Private data center construction spending is up over 73%, from $49.05 billion to $84.95 billion, year over year (YoY), according to the Census Bureau’s detailed construction spending tables for August. Spending is at a record annual pace, and growth is the fastest since April 2025’s +88.2%. Data center buildouts are now about 63% of all private office construction. The report, released last week, uses Census definitions that exclude any value of “racks or servers in data centers” from its construction spending numbers. It also does not speculate on any cause of the spending increase.

The Census data covers the Value of Construction Put in Place, measured monthly, to calculate what a full year would total at that month’s pace. Data center construction is defined as “[including] buildings that contain the hardware needed for storing, processing, and transmitting digital information,” a subcategory of private office. Power plants and transmission lines built to meet data center power demand fall under Census’s separate power category.

The $39.1 billion gap between general office and data center construction spending is the widest in the series since Census data begins in January 2014. The former is at $45.8 billion, down 9.7% YoY and 34.8% since Jan. 2023. Data centers have, in fact, beaten the general office every month since Sept. 2025.

Data center construction now sits at over nine times its Jan. 2021 level of $9.26 billion. Data center construction has become the majority of private office construction, about 63% versus 47% a year earlier, and with forecasts for data center power use still climbing, that is likely to remain the case moving forward.

Private nonresidential construction was at $773.0 billion, down 1.0% YoY, or $688.1 billion and down 5.9% YoY if data centers are excluded. Of the $8.0 billion monthly rise in private nonresidential spending, $5.9 billion or 74% was for data centers. Manufacturing, down 19.8% YoY, fits the same pattern, and data center construction is now about half its size.

Power construction, up 9.7% YoY, is the other big gainer. Taken together, the picture painted is that the construction industry’s growth depends heavily on data center projects. This follows Associated Builders and Contractors chief economist Anirban Basu’s statement in a Sept. 1 release that July’s nonresidential gain was “entirely due to data centers,” and that power’s growth was boosted by their electricity needs.

Growth has gotten faster for four months now. In addition to August’s reported gain, May was +36.3% YoY, June was +55.2%, and July saw +65.3% by our math. As the numbers do not include server spend, which is also growing due to rising hardware costs from AI demand, the Census only shows part of the picture. The acceleration in construction is nevertheless a strong indicator of where things are headed. September’s results are expected on Nov. 2 and may continue the trend.

Oracle has reportedly made a deal with Tencent for a five-year lease at several Oracle data centers in Southeast Asia, covering about 100,000 advanced AI chips unavailable in China. The deal is estimated to be worth about $7 billion, which would come out to about $1.60 per chip per hour over five years, with about 30% upfront, according to a report by the Financial Times (FT). The deal lands as compute rental prices continue to climb, per Chief Strategy Officer James Mitchell on Tencent’s August 12 earnings call. Neither company has commented on the report, but such leases are legal under current U.S. rules, per the FT.

The estimated chip-hour price is around 43% less than the roughly $2.80 per GPU-hour for Nvidia’s H100 with a one-year contract, per research firm SemiAnalysis in August. It’s also about 48% lower than the $3.09 per GPU-hour for a three-year contract with Nvidia’s B200, per a 2025 filing by Japanese data center operator Datasection for an unnamed customer. Market trends have pricing running the other way, with the FT reporting longer terms and larger upfront payments for such leases. Tencent president Martin Lau said on the same call the company could sell its existing orders at “more than 30% profit” compared to what it “paid just a few months ago.”

No chips are named. The FT says that such leases could reach even Nvidia’s top processors, but that does not mean this deal involves Nvidia hardware. One-year H100 contracts on SemiAnalysis’s index were above the estimated rate at about $1.70 per GPU-hour even as of October 2025, with the B200 rate being even more expensive despite its longer, three-year term. Going by raw B200 server costs, which exclude buildings, power, networking, and financing, 100,000 B200s would cost $5.44 billion, or about 78% of the reported value, based on Datasection’s 2025 purchase of servers with 5,000 B200s for $272 million.

Without specifics on the hardware being rented, it is not possible to determine whether this is a bulk deal or if the leasing costs are closer to the actual hardware cost. The pricing on paper, however, better fits Hopper-class at volume. Hopper would include the H100 and H200 and, with H200 imports still limited, could meet the criterion of the chips in the deal not being available in China.

The Oracle data centers are located outside China, with Tencent merely renting time on them. Details including the countries involved, sites, and start date remain undisclosed. Oracle has two Singapore cloud regions, with a Malaysia one announced. The FT says the biggest clients of Southeast Asian data centers are ByteDance and Alibaba, with Chinese firms using the capacity especially for AI training, which domestic chips cannot yet do. Tencent previously inked a $1.2 billion-plus deal for Datasection Blackwell rentals in Japan and Australia, the FT reported in December 2025.

Some Oracle clients have prepaid for GPUs or brought their own hardware since its March call, and fiscal Q4, from March to May, added $67 billion of AI contracts, mostly of those two kinds, per its June call. June through August, per its quarterly filing, the company “received $11.4 billion of prepayments from customers that included a significant financing component.” Deferred revenue rose from $15.4 billion to $30.8 billion over the quarter. No customers were named, so this can’t be tied to Tencent. The company reported 97.9% GPU utilization during its September 10 earnings call, with renewals at a 20% average premium, which was mostly on GPUs four years or older.

Tencent prepaying about 30% makes sense given Oracle’s statements earlier this year. It’s a good deal for both parties. Oracle would improve its cash flow following a quarter of negative free cash flow, and Tencent Cloud’s May price raises, per Mitchell, could help recoup the cost of the rented chips.

The legal picture, given the friction between the United States and China over AI export controls, is thorny. The Remote Access Security Act, a bipartisan bill that passed the House 369–22 in January, would give the U.S. government the right to regulate remote cloud access to sensitive technology, which could cover the chips under the reported deal. The bill specifically targets cloud loopholes and applies to any foreign person rather than naming countries, to prevent adversaries from leveraging cloud networks to bypass U.S. export bans. The bill currently sits in the Senate Banking Committee.

Separately, Export Administration Regulations license requirements under the Department of Commerce’s Bureau of Industry and Security (BIS) have covered shipments of advanced AI chips, including to China-headquartered companies abroad, since November 2023. Cloud rentals were left out, with the BIS only asking for comment on them. Renting remains one loophole around BIS guidance on restricted chips. Bloomberg reported that the BIS was reviewing offshore access with legal rentals included. The deal is also not barred by Tencent being on the Pentagon’s 1260H list of “Chinese military companies,” as this is not a sanction. No statement about the deal by any regulatory government official accompanied the report.

According to The Information, the Department of Commerce is drafting a rule to block Chinese firms from renting compute in third countries, such as Thailand and Singapore, which would align with the reported details of this deal. U.S. President Donald Trump and Chinese President Xi Jinping met late last month in Washington, covering AI and extending a trade truce for a further two months.

However, the deal is of a type that might face trouble from both the House-passed bill and Commerce’s reported rule. But the latter cannot be enforced under current law, an attorney at Baker McKenzie told The Information. At the summit, the two leaders agreed to set up a channel to handle AI-related incidents, but nothing was announced on chip controls. While the U.S. has cleared some firms, including Tencent, to buy Nvidia’s H200, China has let in only about 10,000 for Tencent so far, the FT reported in August. The next meeting between the two, at APEC in Shenzhen in November, may clarify the situation.

The reported deal comes weeks before Oracle’s Investor Day on Oct. 28 in Las Vegas, during its Oct. 25–28 Oracle AI World event at the Venetian. That would be a chance for the company to disclose details about the deal, such as announcing what hardware is to be leased. Given the hardware math, our assumption is that it involves older or non-Nvidia hardware rather than an action born of financial stress at Oracle. The month following Oracle’s event, Tencent is expected to report its Q3 results. These events could also reveal more concrete details about the arrangement.

Google’s latest data center construction project in Finland is being investigated after the company representing the search giant allegedly cleared over 300 hectares of forest. The reported issue is that Google’s local operator, Tuike Finland, cleared the 420 football fields' worth of trees without obtaining a mandatory environmental impact assessment beforehand, says AFP.

Google sees Finland as an attractive European data center host country due to a number of favorable factors. The source notes that it is a country with a cool climate, plus stable and cheap energy supplies (mostly nuclear, wind and hydro). Moreover, Finland sees data centers as a solution to address some current woes such as slow economic growth and record-high unemployment levels.

This new development near Muhos, a small town with fewer than 9,000 residents, isn’t the first for Google in Finland. It already has a facility up and running in Hamina, in the south, but hopes to expand in Muhos, Kajaani, and Vaala, in the north. Overall, it is a massive expansion project, and part of a 13-billion-euro ($15-billion) investment in Finland over the next two years. Finland’s potential data center appeal has also made it a location of choice for Microsoft and TikTok.

While the local mayor looks forward to the massive investment, growth of business, and general economic boom for the region, residents are waking up to the experience of others where data centers have sprung up. Concerns about impacts on the environment, energy supply stability, and energy prices are simmering under the surface.

Google seems to have trampled on some of these concerns already, with these reports of trees already felled and excavators clearing the ground before government agency approval. Not a good start.

"In our view, operations at this construction site should be suspended until it is clarified whether the project is in compliance with the law," Hanna Halmeenpaa, chair of the board of the Finnish Association for Nature Conservation, told AFP at the edge of the freshly deforested area near Muhos. Even some nature preservation areas have been logged, reckons the chair. Halmeenpaa was responsible for filing the suspension request with the conservation agency.

Google denies prematurely devastating the forest. It told AFP that tree felling had been "in compliance with Forestry Act requirements." Moreover, its Finland comms manager said, in a contrasting statement to that of Halmeenpaa, that Google has been ensuring "high-value nature zones within the sites have been protected during forestry works."

We haven’t seen any other reports indicating that preparation and construction work has ceased at the new data center development site near Muhos. So it seems like Google will plow ahead, hoping or knowing that the political and/or business will is sufficiently strong to ensure it won’t have to undo any of the actions it has already completed.

Amazon has issued grave warnings to communities opposing data centers, stating that their actions could cause irreversible damage to the U.S. economy and national security. In a lengthy 3000-word post earlier today, October 2, CEO of Amazon Web Services (AWS), Matt Garman detailed the many benefits of Artificial Intelligence, comparing the technology’s impact on the economy to those of modern transportation systems and the Internet, before attempting to debunk “myths” around data centers and calling on local communities to be more receptive to build-outs as a matter of national urgency. The post also details Amazon's plans to do more for data center host communities through its “Built Together” community investment frameworks.

After outlining AI’s impact and the need for data center buildouts, Garman explained why it was imperative for the U.S. to lead the AI race. “The countries that lead in AI will shape it and get the most from it in the short and long run. The choices we make today will determine that outcome, and a lapse in our nation’s focus or resolve could put us behind, and potentially irreparably so,” he wrote, presenting urgent buildouts as a key factor in leading that race.

Next, Garman addressed growing anti-data center movements that have reportedly blocked data center projects worth $68 billion in the second quarter of 2026 alone, warning that if the 100 data center moratoriums being considered across the country should go through, “the US could be writing its own losing ticket to this race, and the consequences would last generations.” He also pointed out that unlike transportation infrastructure, the comparably critical data center buildout is not funded by taxpayer money, but by private firms.

The post dismissed many of the concerns about the buildouts as myths, then proceeded to attempt to debunk them. Garman said that, contrary to popular belief, “Data centers actually use very little water relative to other industries,” a position many hyperscalers have repeated. Amazon earlier claimed that its data centers consume only 0.075% of the water Americans use to water their lawns and gardens.

Garman also claims the narrative that data centers are driving up energy bills is “misleading and convenient scapegoating.” He argues that AI is simply precipitating an expansion of power infrastructure that was always bound to happen. “Once we collectively build out this new capacity, again funded by the private sector, it should lead to flat or lower rates in the future for residents living in communities with these expanded grids,” he said. On the other hand, contradictory reports claim that AI data centers trigger a "massive irreversible" 76% electricity price spike. The U.S. House also recently passed a bill that directed regulators to consider making AI data centers pay for grid upgrades to minimize the impact on residents. The bill was eventually rejected in the Senate, with opponents saying it was "toothless" and didn't do enough to protect residents.

Speaking on pollution, Garman says that data centers emitting huge amounts of pollution with their large backup diesel generators is a myth. He claims that while data centers have rarely used backup power in case of a power outage, this is only necessary because many industries depend on the facilities. This claim doesn’t exactly address growing concerns of pollution from gas turbines powering off-grid data centers. Elon Musk recently came under fire for using unlicensed gas turbines at the Colossus data center.

Lastly, Garman said claims that communities do not benefit from data centers are false, citing “considerable tax revenue, positive economic impact, large numbers of good jobs, and education programs” as some of the benefits to the communities where data centers are built. Microsoft pledged to invest $1 billion in local infrastructure improvements in its Hyperion data center community and had reportedly awarded $1.6 billion in contracts to local businesses since construction began, in addition to paying teacher bonuses. As benefits specifically on Amazon’s part, Garman listed an estimated $3 billion in taxes to St. Joseph County, Indiana, over time; more than $1 billion already contributed to communities across the U.S.; thousands of jobs created in building and running the data centers; and education programs with over 10,000 educators and learners enrolled.

The company says it plans to do more. Through its “Built Together” initiative, Amazon says it will “add more than $1 billion over the next five years to what we’re already doing — with communities in the driver’s seat to determine what’s most useful for them across education, job training, energy affordability, water and energy preservation, and local priorities.”

Garman insists that data centers can’t be separated from the revolutionary impact of AI. U.S. President Donald Trump appears to agree, at least on the “revolutionary impact” aspect, and has repeatedly expressed a desire to maintain a lead over China. On the other hand, data center critics claim that responsible implementation of the build-outs, and not AI itself, is the bone of contention.

Like several industry leaders and hyperscalers, Garman insists that many of the issues were being blown out of proportion, and implied that much of the anti-data center sentiment could be traced to nefarious foreign influence. “With any change, there will be important questions raised, but there will also be misinformation and outright lies, and in the age of social media and 24/7 news, myths take hold faster than ever before. In fact, this buildout is so important geopolitically that there are widespread reports of various countries intentionally seeding misinformation in the U.S. about data centers to trick us into slowing down,” he said.

The United States Senate has decided against advancing the Ratepayer Protection Act, a bipartisan bill designed to shield everyday Americans from rapidly rising electricity costs driven by data center buildouts nationwide. On September 30, the bill fell short of the 60 votes required to pass the bill by just three votes, ending 57-43, according to a Reuters report. The Ratepayer Protection Act bill had cleared the U.S. House of Representatives just two weeks earlier, with the Senate as the last procedural step before President Donald Trump could sign it into law. Following the vote, the bill is now officially dead.

Introduced in the Senate by Senator Jon Husted, the bill was the legal culmination of President Trump’s “Ratepayer Protection Pledge.” It sought to institute a federal standard for states and Public Utility Commissions, directing regulators to consider whether entities that consume large amounts of electricity, such as data centers, should cover the incremental costs of energy infrastructure needed to serve them.

Proponents say the act would prevent data center operators from passing infrastructure costs to residents and small businesses, a growing concern in communities seeing a rise in data center construction. “If America wants to lead the world in AI and strengthen our national security, we have to build the energy infrastructure to support it. But we must do that without passing the costs on to working families and small businesses. The Ratepayer Protection Act would keep America globally competitive while protecting Ohioans and Americans from higher electricity bills,” said Husted.

On the other hand, opponents, mostly Democrats, labeled the bill “toothless,” arguing that its non-binding nature provided no real protection because it did not mandate that states adopt the standard; it simply directs regulators to consider it as part of their rate-setting processes. Only four Democrats voted for the bill; the general consensus was that it does not go far enough to protect consumers from soaring electricity bills.

With the bill’s collapse, Congress has yet to impose any new federal protection on data center-related grid costs. Some lawmakers have suggested they may reevaluate the issue after the mid-term elections in November, possibly with revisions that would require utility regulators to ensure data center operators cover additional costs caused by their operations. The debate over data center energy costs and who should cover them will likely remain a key issue during campaigns, especially in states where new facilities are planned or where buildouts are ongoing.

Meanwhile, local communities continue to push back against data centers over several issues, including electricity bill hikes. Some states are already taking steps to address the hikes. An Oregon power company hiked data center bills by 30% and cut residential electricity costs by 1.3%. Certain hyperscalers are also moving their data centers off-grid, with others pledging to supply electricity to the grid or offset increases in residents’ bills.

Microsoft’s “biomimicry” program is reshaping the land around its data centers with restored wetlands and native gardens to help the buildings blend into local ecosystems at 20-plus sites, reports The Verge. Critics question the company’s motives after recent energy deals, such as a 20-year deal inked in June to buy power from a planned Chevron gas plant in West Texas. Public Citizen’s Texas director, Adrian Shelley, called the plan “basically lipstick on a pig.”

Microsoft says the approach is now part of all new projects in the U.S. The goal is to blend “data centers into the natural landscape” at sites across the United States and Germany, gauging community input and using ecological research to map the local environment. Microsoft’s team uses a tool called Ecosystem Intelligence, which runs on Azure and scores various features of the site, including water quality, biodiversity, and noise abatement.

Solutions, according to Microsoft’s director of global land strategy, Kaitlin Chuzi, are customized to the local habitat and should restore what was lost, while plant buffers quiet data center humming and help “minimize the visual impact on communities near our data centers.” Named sites include Mecklenburg County, Virginia; Mount Pleasant, Wisconsin; East Point, Georgia; Bedburg, Germany; and other projects in Ohio, North Carolina, and Wyoming.

The Mount Pleasant site, home to Fairwater, which Microsoft calls the “world’s most powerful AI data center,” has 100 acres of prairie and forest in a Lamparek Creek restoration with the Root-Pike Watershed Initiative Network. Its executive director, Dave Giordano, told the outlet that Microsoft developed its stormwater ponds with forebays to retain sediment, one of a number of accommodations for the environment. “Not all data center companies are the same,” he said. It is a living example of the company’s answer to AI data center backlash.

San Antonio District 6 Councilmember Ric Galvan told the outlet that there are more than a dozen data centers in his district, with more proposed. He cited concerns about the impact on traffic and the environment, including tree removal, following a 2022 approval of a Microsoft plan to remove 2,642 trees from a 33-acre tract for a data center.

Microsoft spokesperson Jennifer Lyons, however, said in an email to the outlet that the company is investing in more than 400 acres of woodland and pollinator prairie restoration in San Antonio and intends to leave the “land more productive than we received it.” Galvan said Microsoft may be “covering their bases,” which he said is fine if residents’ concerns are addressed. He proposed a moratorium on new data centers in August, backed by Public Citizen, but the council opted against a blanket pause and is weighing zoning rules instead, with a vote scheduled for Oct. 15.

Shelley also pointed to adjacent effects, specifically pollution from new gas plants built to supply energy for data centers. “I cannot really see any level of biomimicry investment that makes a gas plant a desirable neighbor,” he told the outlet. Chevron says its plant would provide “approximately 2.67 gigawatts” of power if the plan goes through, with first power anticipated in 2028.

Meanwhile, Texas Gov. Greg Abbott has paused approvals for data centers seeking to connect to the state’s main power grid until regulators audit them. In September, he barred the state’s environmental agency from issuing new data center permits until the audit is done. Water use, a common criticism, was only about 0.4% of the state’s total from data centers in 2025, according to the Houston Advanced Research Center (HARC). It projects water usage of 0.5% to 2.7% by 2030.

Emissions have risen by Microsoft’s own accounting, too. Its 2026 Environmental Sustainability Report, released in July, puts the total up 25% in fiscal 2025, “driven primarily by the expansion of our datacenter infrastructure” and by Microsoft pausing its use of unbundled renewable energy certificates, a type of accounting credit. Microsoft has 2030 goals of becoming carbon negative and water positive, meeting the latter in fiscal 2025 by its own measure.

Holly Alpine, who led ecosystem restoration and climate resilience work near Microsoft’s data centers until leaving in 2024, still sees a gap between data center buildouts and sustainability goals. Microsoft said in an email to the outlet that AI at a global scale must have “a range of energy solutions in the near term.”

Microsoft had no clear answer when questioned if it would ever change its mind on a site. Chuzi called it “an ongoing learning process to understand what that means to be a good neighbor.” HARC’s Erin Kinney told the outlet the real test is how Microsoft calculates, tracks, and shares the benefits with communities. Accurate measurement would make it easier to have genuine cooperation moving forward so that both data center companies and local communities can come together to mitigate environmental impacts.

Amazon announced a new community program called “Built Together,” where the company promised to invest $1 billion over the next five years on data center communities. According to the company blog post, it will provide the funding to the residents while they determine the best use of the funds to be allocated, including projects that reduce electricity bills.

“With Built Together, Amazon will add more than $1 billion over the next five years to what we’re already doing — with communities in the driver’s seat to determine what’s most useful for them across education, job training, energy affordability, water and energy preservation, and local priorities,” AWS CEO Matt Garman wrote. He said that Amazon is focusing on three key issues that matter most to residents living near its data centers: education and workforce pathways, community energy affordability and water solutions, and flexible funding for local priorities with communities in the lead.

The company says that education and workforce pathways will allow the people living near data centers to pursue associate’s degrees related to in-demand fields at zero out-of-pocket costs, with Amazon covering the shortfall after financial aid. It even claimed that it has already established agreements with local community colleges and expects to launch them in the next few months, with a goal of connecting more than 300,000 students from data center communities to its free degree access. Aside from this, Garman also said that the company is expanding its Modular Training Centers, which provide free certification programs for skilled trade workers so they can eventually get employed at Amazon or its partner construction firms, with a target of funding training for up to 100,000 workers per year.

Amazon is also looking to provide grants to schools, community buildings, and even homes in data center communities to help deploy energy-efficient solutions. These include heat pumps and HVAC systems, insulation, water heaters, batteries, solar installations, and more, with an aim to reduce electricity costs by around 20% to 40% annually and reduce household electricity expenses by up to $700 per year. Last, but not least, Garman said that it’s partnering with local nonprofit organizations and community foundations to ensure that its earmarked funds are best spent according to the needs of the residents. “Every community is different,” said the AWS chief. “What matters most in rural Oregon is not the same as what matters in suburban Virginia or a mid-size city in Ohio.”

The first part of Garman’s post actually addressed many of the concerns raised by community members near these developments, calling them “false” and “misleading,” despite various reports that have raised concerns about utility hikes and noise pollution. Still, it seems that these initiatives are designed to alleviate some of these issues and make data centers more acceptable to communities where they’re being proposed. This is especially true as many projects are being delayed due to bipartisan AI data center bans, especially as polls show that the majority of Americans resist having one put up near their residences.

Other AI hyperscalers and data center developers have offered incentives to communities even before Amazon made this announcement. SpaceXAI offered a 50% discount on Starlink plans to placate its residential neighbors in Memphis and Southaven, while another developer said that it will give $10,000 checks to 4,500 households if its 1,300-acre facility gets approved. One data center site in Maryland even proposed an astounding $110-million package to the community, which is said to be the biggest-ever package offered in U.S. history. Many critics decry moves like these as “bribes,” though, just so developers can get their projects going, but New York State recommended a similar pathway, suggesting a $1 million investment per data center megawatt built in the community.

These values might seem like a lot of money, especially as Amazon’s promised $1 billion investment amounts to $200 million per year. But even though this might seem like a good amount, it’s a tiny fraction of the trillions of dollars projected to be spent on data centers through 2050. Amazon itself is expected to have a capital expenditure of $220 billion on AI data centers this year, making its community efforts cost just 0.1% of its overall AI infrastructure budget.

NorthPoint Development wants to build a data center on 1,300 acres in the Pocono foothills, near Hazle Township, Pennsylvania. As well as seeking to charm the local government with various funding packages, it has proposed paying checks of $10,000 per household. Despite the Pennsylvania town’s median income of $60,000, the offer isn’t being warmly welcomed by locals, reports the Wall Street Journal. Those interviewed were widely against the data center development plans and cited several reasons for being averse to accepting the payout. The major concerns regarded potential impacts on property values, the specter of noise pollution, and a distrust of both the AI moguls and local government representatives.

Offer letters began to be received by Hazle Township back in June. The WSJ recently talked to several residents about NorthPoint’s $10,000 offer. While some say the $10,000 would be handy, it couldn’t find many residents who were “eager to welcome the data center for the payout,” in the journal’s words. We’d point out that opinion polling can often be inaccurate, as what people say and what they do, voting-wise, doesn’t always line up.

Interviewees cited a number of key reasons for their data center development aversion. We distilled the following points from the source:

• Noise pollution concerns
• Water and energy cost worries
• Dislike of AI
• Distrust of AI moguls and local authorities
• Potential negative impact on property prices

Traditionally, developers have offered towns and cities packages worth tens of millions, including funding for local services. But direct cash payments to residents represent a new phase of their charm offensive, the WSJ indicates. Perhaps the idea for the direct payments to residents came after “a raucous public meeting last year, where angry locals asked what they had to gain by supporting the data center,” ponders the source.

The locals won’t just get the direct payout, though. Other funds up to $120 million over the next 15 years have been dangled like a carrot for community and local services investments — improved schools, emergency services, and so on.

Meanwhile, Hazle Township’s government already rejected the data center project on zoning grounds last year, and it has enacted a temporary moratorium on data center construction. Finally, the $10,000 offer to households is simply not enough in many interviewed folks’ opinions. Residents could feel insulted by it, regarding the offer as a bribe, concludes the WSJ report.

The Russian military has started striking data centers within Ukraine in its latest escalation against the country. According to the BBC, a Russian drone hit the data center, which was situated in a central business district in Kyiv, damaging it and killing four people in the area. Multiple reports also suggest that other data centers and network infrastructure have been hit across the country as well, with President Volodymyr Zelensky saying that Moscow wants to maximize disruption for the common people.

“All this affects people’s ability to stay connected, study, work,” Zelensky said. He also wrote in a social media post, “The Russians are constantly expanding their escalation operation: American and other businesses, data centers, internet providers — for Russia, all ordinary life is simply a target. It is important that the world responds to all of this and that Russia feels the response to its terror.” Russia, on the other hand, confirmed that it attacked the Datagroup site, claiming that Ukrainian military intelligence used it, and that it’s also targeting other data and telecommunications centers owned by New-Telco, United DC, Kyivstar, and Parkovyi.

These attacks have left about 100,000 households in Ukraine without internet connectivity, and some companies have begun migrating their data across the border to avoid disruption. Nevertheless, one government official said that Ukraine’s internet network is highly decentralized, but people should still expect some local disruptions. However, Ukraine Foreign Minister Andrii Sybiha said that Russia’s targeting of data centers would affect the flow of “life-saving information.” “Rapid alerts about missile and drone threats are essential — they save lives. This is crucial civilian infrastructure… It is not a military target,” Sybiha said. “It is essential to keeping everyday life functioning.”

Data centers are considered critical infrastructure as modern society increasingly relies on online services. Because of this, they’re now considered strategic targets in a conflict. While government-owned data processing sites have historically been one of the first targets in military operations, Iran’s response to the U.S.’s bombing campaign in March marked one of the first instances where commercial civilian data center infrastructure was deliberately targeted. This included drone and missile attacks on AWS Middle East regional data centers in Dubai and Bahrain. It also directly threatened the offices of Nvidia, Microsoft, Apple, Google, and 14 other American tech companies in the region, as well as OpenAI’s Stargate AI data center in Abu Dhabi.+

The end goal is finally in sight for Elon Musk, nearly two years after he announced plans to expand the Colossus supercomputer to over a million GPUs. The billionaire said on X that 220,000 Nvidia GB300 GPUs will be operational by next week, with another 220,000 coming online in November. He also added that another 220,000 units will come online by late December “if we get lucky.”

These numbers would add to the 110,000 GB200 and 440,000 GB300 GPUs already operating at Colossus 2, plus the 150,000 H100, 50,000 H200, and 30,000 GB200 GPUs at Colossus 1. This would bring SpaceXAI’s GB300 GPUs to 1.1 million units, with its total GPUs in operation to 1,440,000 units. This is quite an achievement, especially given that the company is behind its rivals by several years — SpaceXAI is only three years old, while Anthropic and OpenAI are six and ten years old, respectively. Interestingly, the Colossus 1 site, which features a combination of Hopper and Blackwell GPUs, is inefficient for training Grok, so Musk rented it out to Anthropic for inference instead. On the other hand, Colossus 2 solely uses Blackwell GPUs, ensuring that there won’t be any bottlenecks.

While Elon Musk essentially begged for Jensen Huang to give him these GPUs, getting his hands on them isn’t currently the hardest part. One of the biggest issues facing data center build-outs right now is power, and SpaceXAI solved this by bringing its own. This has resulted in some controversies with the surrounding neighborhood, ending in a lawsuit against the company for running unpermitted gas turbines. It has since promised to remove them, but only over a span of one year as its own 1.2-GW power plant comes online.

Running a million GPUs or more on a single site is an impressive achievement, but SpaceXAI isn’t the only one trying to reach this goal. Broadcom said in 2024 that it has three hyperscale customers gunning for this target by 2027 but didn’t mention who these customers were. As for Elon Musk, the 1-million-GPU target is just the beginning. He said that SpaceXAI will grow its data center capacity sevenfold by 2027, and he’s even aiming for 50 million H100-equivalent GPUs by 2030. These data centers won’t be limited to the ground as well, with SpaceX planning to launch an Orbital Data Center System with a million satellites, despite Jensen Huang saying that it’s a “dream” for now.

The DataOne data center in Vineland, New Jersey, has just been hit with a record $1.07 million fine for operating unpermitted portable generators on the construction site. According to the New York Times, state inspectors spotted 62 of these units powering the facility back in late July.

Residents had already complained about it at a February town hall meeting. Inspectors said that there were no generators on site when they first visited in December last year. However, some residents said that “some type of unit” was spotted by satellite imagery back in February, leading to their complaint at the community assembly.

Residents of Vineland have already made multiple complaints about the development. Some of the issues t include the unpermitted construction of a 1.5-million-gallon LNG tank and unacceptable noise levels, especially at night, that affected homes within a half-mile radius of the construction site. They also said that there was little to no public consultation on the project before it went into full swing, with the people only finding out that a data center was being constructed in their neighborhood after Microsoft announced a $33 billion deal with Nebius, which in turn contracted DataOne to supply the compute they need.

Despite a record fine, some residents feel that the punishment isn’t enough. “It’s a ‘don’t ask for permission, just ask for forgiveness later’ kind of deal,” longtime Vineland resident Steve Brown told the Times. He also added, “It’s still kind of a drop in the bucket for a company that has several billion dollars of capital.” Long-time community member Tiffany Leone-Vespa also added, “What’s $1.07 million? They should have done it for each generator they were running illegally. They’re billionaires. It’s like pocket change.”

On the other hand, DataOne disagreed with the fine. A company spokesperson said that “We will apply for the air permits for the temporary generators,” and also added that it’s transitioning from the polluting and noisy portable turbines to more modern fuel cells. New Jersey ordered it to comply with the permitting requirements and to stop using the generators if they’re not granted within 45 days.

This is a similar situation that the citizens of Memphis and Southaven faced when SpaceXAI deployed 69 unpermitted generators to provide power to its Colossus 2 data center. It has become the subject of a lawsuit, especially as the community raised concerns that they were spewing thousands of tons of pollutants in the primarily black neighborhood. However, the Department of Justice stepped in and asked the court to dismiss the case because the site “supports mission-critical operations.” Elon Musk’s company eventually agreed to remove all unpermitted turbines, but said that the process will take over a year as replacement 1.2-gigawatt plants gradually come online.

Google is set to launch an experimental satellite, named MVP, into orbit next week to test running an AI data center in space. As part of a wider initiative called Project Suncatcher, the satellite will be aboard a SpaceX Falcon 9 rocket when it launches from the Vandenberg Space Force Base in California. Once in space, and assuming everything goes well, the tech on board the satellite will be capable of handling simple AI requests from space.

The New York Times points out that while the move is an important one, the satellite will be limited in terms of its processing power. It will feature just four Tensor Processing Units (TPUs) and be similar to a single Earth-bound server in terms of capabilities. It will be powered by solar panels capable of providing a single kilowatt of power to the tensor chips.

While clearly limited, the orbiting server is set to handle AI requests for up to a year, although it will remain in orbit for up to six years. After that, gravity will take hold, and the satellite will burn up during re-entry into Earth's atmosphere.

Before that happens, James Manyika, Google's senior vice president of research, expects to learn more about how space may become a home for future data centers, though probably not for a long time.

“We don’t expect, to be perfectly frank, that we’ll have anything usefully operational in the next few years,” the SVP told the Times. He went on to compare the venture to Google's driverless cars, saying that Google was testing those "like 15 years before anything showed up." The idea of users accessing a version of Gemini running in space is some way from coming to fruition.

But it's also something that Google and other companies are set to investigate further. Alongside Google, Elon Musk (xAI, Tesla, SpaceX), as well as Jeff Bezos (Amazon) and Sam Altman (OpenAI), have been making noises about the prospect of orbital data centers for some time now. China has already put its own AI compute into orbit. Even Intel has a patent for orbital data centers of its own.

Putting servers into space does have its issues, not least the problem of getting them there. Radiation and computers don't get along, and the radiation found in space can cause "bit flips" that change the way chips process data. Google's answer, for now, is to restart the chips every so often to reset them. The company has also devised a cooling system that will vent heat directly into space — since the usual fan-based solution simply won't work without any air.

Despite these accommodations, the four chips in Google's test satellite will still get too hot, too quickly. They'll only be able to run for around 15 minutes before they need to be shut down to allow them to cool off.

Regardless, it's still early days for the concept of running servers in space. Google plans to launch a pair of satellites next, with long-term plans for networked fleets of more than 80 satellites also already floated.

By Manyika's estimation, Google has 15 years to go from a four-chip server to a full-on data center. The clock is ticking.

UK-based “neocloud” Nscale built its business on a major deal with a ByteDance subsidiary named Spring, which accounted for $24 million of its $33 million annual revenue in 2025. This contract allowed Spring to access 2,304 Nvidia B200 chips at Nscale’s Glomfjord, Norway, data center. Even though the Financial Times reports that this was legal and exploited loopholes in the U.S.’s ongoing export controls against China, it still exposes the company to regulatory and reputational risk, especially as Washington continues to restrict Chinese access to its most advanced AI accelerators.

This information only came to light after Nscale released its S-1 documents as it filed for an initial public offering in the U.S. There was also no direct mention of the TikTok owner or its subsidiary anywhere in its S-1 filing; instead, it was noted in a supporting document, a $105 million loan agreement with Macquarie, that Spring (SG) Pte Ltd, a Singaporean ByteDance subsidiary, was a significant customer. This credit, alongside a $35 million in equity, helped fund the purchase of advanced Nvidia AI GPUs, which also caught the attention of the chipmaker.

A month after Nscale drew the first tranche of the loan, it closed a deal with Microsoft and Nvidia, with Jensen Huang committing to invest more than $660 million into the company. This funding is part of an agreement between Nscale and Nvidia, under which the latter promised to buy large volumes of AI chips in the future. Nvidia has since increased its commitments to over $2 billion, plus an additional $860 million guarantee on Nscale’s lease at a Texas facility. Nscale has since been able to secure larger deals with other AI hyperscalers, such as a $44 billion deal with Microsoft and another $45 billion contract with Anthropic. The company said that this would make its Spring contract fall to below 20% of its overall revenue, and it is expected to shrink further as Nscale secures other Western clients.

Because of Spring’s link to ByteDance, Macquarie compelled Nscale to monitor how the Chinese subsidiary used the AI chips and report “compute anomalies or suspicious configurations” that might be in violation of the U.S.’s export controls. Nscale also had a third party conduct due diligence on both the TikTok parent and the Singapore-based subsidiary to ensure that the deal complied with the law.

Chinese companies have been taking advantage of these loopholes in the U.S.’s export controls to gain access to advanced AI chips. While they cannot import the most advanced Nvidia AI GPUs, like Blackwell and Vera Rubin, into China, they could still rent these AI chips from companies based abroad or by using subsidiaries founded outside of China to purchase them directly. However, the U.S. has recently closed the latter loophole, with the Trump administration working to close the former, too. It’s unclear if these moves would affect the Spring-Nscale deal, but the recent delays on gigawatt-scale data center projects in the U.S. mean that Nscale could potentially get more clients as AI hyperscalers race to secure compute from smaller facilities.

Anthropic and OpenAI are reportedly scrambling to secure smaller AI data center deployments as they race to bring more compute capacity online while gigawatt-scale projects undergo construction. According to a September 18 CNBC report, the two AI giants are exploring agreements for existing facilities with roughly 20-30 MW of capacity, despite already committing tens of billions of dollars to much larger AI infrastructure projects.

Anthropic has reportedly sounded out potential deals of that size across the UK and Nordic countries, according to CNBC, citing four people familiar with the discussions. OpenAI has also explored smaller deployments in the Nordics, while sources said both companies have discussed similar opportunities in the United States.

“We’re building a diversified compute portfolio to meet growing demand for AI around the world,” an OpenAI spokesperson told CNBC. The company added that different workloads require different infrastructure and that it evaluates potential deployments based on performance, reliability, timing, and cost. Anthropic declined to comment.

The smaller facilities would complement the enormous AI campuses both companies are already pursuing. Anthropic, for example, reportedly signed a roughly $45 billion agreement with Nscale for around 460 MW of compute capacity at a West Virginia development. OpenAI, meanwhile, said earlier this year that its Stargate infrastructure commitments had already surpassed the project's original 10 GW target, before announcing another 3 GW in Georgia and 8 GW in Ohio.

These projects provide the large clusters of accelerators needed to train increasingly large AI models, in which thousands of GPUs must work together via high-bandwidth interconnects. However, they also take years to develop, requiring land, grid connections, substations, cooling infrastructure, and huge amounts of power before the accelerators can begin operations.

Meanwhile, both OpenAI and Anthropic need immediate capacity to meet the inference demand for their models. As it turns out, inference can be distributed across multiple smaller clusters, making smaller data centers appropriate for these tasks and increasing the demand for such facilities.

Smaller deployments also circumvent some of the issues plaguing large data center buildouts. In the U.S. alone, local opposition blocked 45 data center projects worth $68 billion in the second quarter of 2026 over land usage, noise pollution, and electricity and water consumption. Existing, small facilities do not face the same pressure.

“Securing a few megawatts at an existing powered site can be more practical than waiting for a much larger block in one location,” Jabez Tan, head of research at Structure Research, told CNBC. “For workloads that can operate across separate sites, a collection of smaller deployments can add up to substantial capacity.”

Renting capacity is already standard practice for AI companies, as AI labs lease capacity from hyperscalers, dedicated data center operators, and rapidly expanding GPU-focused neoclouds. This allows them to spread workloads across multiple providers.

Edit: 8/24/2026 6:25am PT: Energy provider Entergy reached out to clarify the savings to customers outlined below, stating, "The correct number is $2.65 billion over 20 years in customer savings. This information is found in our Meta fact sheet that is publicly available at www.entergy.com/datacenters and in our press release first announcing the savings on March 27, 2026." We have corrected the story below accordingly.

Amended article follows:

Local opposition blocked or delayed $68 billion in data center projects between April and June this year, according to a September 21 Bloomberg report citing data from research group Data Center Watch. The report says 30 state houses have implemented rules on data center location and resource consumption, with communities pursuing moratoriums on planned construction even before developers apply for permits. Conversely, hyperscalers have spent over $1 trillion on data infrastructure since 2023, with an additional $745 billion in Capex expected in 2026 alone.

The blocked-project figures reflect widespread opposition to data centers across the United States over concerns about land usage, noise pollution, and electricity and water consumption. In July this year, 142 AI data center protests were simultaneously staged across 42 states in one weekend. Data Center Watch’s research estimates that there are 843 opposition groups across all U.S. states except Hawaii.

One leading concern is the environmental impact of AI data centers. Each consume huge amounts of water to cool hundreds of thousands of accelerators powering artificial intelligence, with AI reportedly set to consume up to 600 billion gallons of water by 2030. The hyperscalers, on the other hand, insist the issue is being blown out of proportion.

OpenAI CEO Sam Altman says 38,000 ChatGPT queries use as much water as producing one almond, while Microsoft says its new data center architecture consumes as little water as a restaurant. Meanwhile, in May this year, an AI data center project secretly consumed 29 million gallons of water within a year and three months before low water pressure alerted nearby residents to its existence.

Protesters also commonly cite electricity consumption as a major concern. Analysts project that data centers will consume 20% of U.S. power by 2035. Local residents are already feeling the effects. An AI data center in Virginia led to an irreversible 76% hike in residents’ electricity bills, prompting the state to eventually mandate hyperscalers to cover the cost of the transmission infrastructure required exclusively for the project.

Many regions are now implementing similar mandates, with hyperscalers offering to offset the bills in many cases. These concerns, along with the extreme power demands of AI data centers, have also led hyperscalers to build on-site power, with some offering to supply power back to the grid. Meta, for example, recently signed an agreement with utility company Entergy Louisiana for 7 GW of power for its Hyperion facility. Entergy claims Meta's payment provide other customers with $2.65 billion in savings over 20 years.

Meanwhile, the shift to off-grid, on-site power generation is creating yet another issue driving the protests. Due to their scalability and speed of availability, many hyperscalers are relying on gas plants, which mostly burn fossil fuels to generate electricity. Elon Musk's Colossus 2 data center recently came under fire after 59 unpermitted natural gas turbines powering the data center released thousands of tons of pollutants into predominantly black communities in Mississippi. Elsewhere, Amazon’s custom 35-turbine gas plant has been authorized to emit 33 million tons of greenhouse gases annually, a move that will make it the largest single source of CO₂ pollution in the U.S.

The growing sentiments present a sort of dilemma. AI is proving to be quite the revolutionary technology and has seen massive adoption and skyrocketing usage. To meet demand, hyperscalers are erecting data centers at breakneck speed, which residents now oppose. Proponents say data centers are an absolute necessity and any opposition stifles growth.

Opponents, on the other hand, say they are only demanding responsible execution and have nothing against AI. President Trump has condemned communities that have opposed data centers, saying that they “want to end up being backwards and poor.” He has also implied that China could be behind the sentiment, orchestrating movements behind the scenes. Meanwhile, officials across the country are reportedly facing death threats and gunfire over AI data center projects, even as AI data center investments are projected to hit $32 trillion by 2050.

The U.S. House of Representatives just passed a bill that creates a federal standard requiring data centers to pay for grid upgrades made in their favor. H.R. 9340, also known as the Ratepayer Protection Act, amends the Public Utility Regulatory Policies Act of 1978, which would require each State regulatory authority and each non-regulated electric utility to consider the adoption of the bill within two years of its passing, if it is signed into law.

This bill would ensure that data centers with a capacity of 100 megawatts or more would have to pay “the full, incremental cost of any generation, transmission, or distribution upgrade necessary to serve the load of such large-load customer, including in the event of such large-load customer terminating a contract or other agreement with the electric utility pertaining to the sale of electric energy, or otherwise ceasing the purchase of electric energy from the electric utility.” This bill closely follows President Donald Trump’s “Ratepayer Protection Pledge,” where he made AI hyperscalers, utility providers, and state governors promise that they will pay their own way when it comes to their electricity demands. All this stemmed from the surprise price hikes that many residential users and small businesses suffered from because of the massive demand by AI data centers and has become one of the primary reasons why the majority of Americans now oppose data center developments in their communities.

Oregon is actually one of the first states to have taken concrete steps in controlling the utility price increases when it passed the POWER Act in 2025. This law is even more stringent, with any development using more than 20 megawatts required to pay its fair share, and has already resulted in a 30% hike for data center electricity bills and a 1.3% reduction for residential power costs. Virginia has also followed suit soon after its governor signed the Ratepayer Protection Pledge in July 2026, with Virginia’s State Corporation Commission requiring data centers to pay for all required transmission infrastructure.

Even though the House of Representatives has already passed H.R. 9340, it still needs to go through the Senate before finally heading towards the White House for signing by the President. But even if it passes through the remaining hurdles, and the U.S. adopts a federal standard where large data centers pay for grid upgrades done in their name, it’s still up to each state regulator if they will adopt the standard. Furthermore, states have up to two years to make a final decision, meaning there’s a chance that the various temporary data center bans and moratoriums would have expired even before state regulators would have enacted this bill.

New York State Governor Kathy Hochul has just announced the Community Investment Framework (CIF), which serves as a negotiation guide for towns and municipalities in the state when they’re dealing with data center developers, recommending that communities demand $1 million per megawatt for the privilege of housing data centers. According to Empire State Development [PDF], the state’s economic development agency, the CIF offers “a structured approach to identifying, negotiating, and documenting investments from data center developers/operators/owners that address local priorities and create lasting community benefits.”

The document recommends six principles during data center negotiations: an investment benchmark of $1 million per megawatt of utility demand; long-term planning, including plans for when the data center leaves the facility; a roadmap to manage ongoing costs for projects funded through the CIF; a list of local priority investment areas to address the needs of the community; local ownership confirmation of projects receiving funding from data centers; and clear timelines for any program, especially those that require phased, multi-year investments. Aside from these, it also gave several recommendations, such as focusing on community engagement, engaging decision-makers and counsel, incentivizing good faith and timely negotiations, and maintaining flexibility, among others.

While data centers have been around for decades now, AI data centers are relatively new and are putting more pressure on electricity and water supplies. They also offer limited job creation and community value compared to other industries, which is why Loudoun County, Virginia, which hosts 250 data centers, has started reining in these facilities as well. More than 500 states, counties, and localities across the nation have already passed a moratorium or temporary ban against data centers, including states like New York and Texas. However, these have an expiration date, meaning government units only have a limited time before they would have to start processing data center applications once more.

This framework is a useful guide for when municipalities have to start negotiating with data center developers again, helping ensure that these projects would bring more benefit to the community while limiting their negative effects. One Pennsylvania town actually created a similar comprehensive document, outlining 43 specific demands before approving a proposed data center in Plymouth Township. Unfortunately, the said developer decided to cancel the application and file a new one instead, with the town accusing it of demanding approval “by tantrum.”

Reports suggest that various data center developers have requested up to 9,000 MW from the state grid. However, this does not include operators that are opting to bring their own power sources or build their own power plants on site. So, if New York towns and municipalities adopt the CIF in their negotiations with data center developers, this could potentially add $9 billion or more towards their coffers. This may seem like a lot of money, but it’s a drop in the bucket compared to the expected $32 trillion that some experts suggest will be invested in data centers by 2050.

More than six months after Amazon Web Services (AWS) data centers in Abu Dhabi and Bahrain were attacked by drones, the company has now had to tell customers to move their data to facilities in other regions. The Wall Street Journal reports that AWS has admitted that it's time for customers to jump ship to data centers located elsewhere in the world. In its AWS dashboard update, AWS confirmed "most customers have been able to re-establish their operations in other Regions by restoring backups or copying data that remained accessible. AWS Support remains available to help customers who need assistance moving their applications to alternate Regions."

The move comes as AWS continues repairs on the damaged data centers, but it stopped short of offering a timeline as to when customers can expect either location to be back online. AWS says that it remains committed to supporting its customers in Bahrain, but only said it would provide a further update early next year. It did not offer any specific recovery time frame for the restoration of services. Regarding the UAE, AWS says that it is working to replace affected infrastructure, but again only said it would provide an update on the restoration of services in the coming months.

Both affected data centers became targets for the Islamic Republic of Iran after the United States and Israel launched strikes on that country in February of this year. Months later, the Bahraini data center was struck by cruise missiles, and Iranian state media claimed it had been "destroyed."

The move is bad news for the United Arab Emirates. The U.A.E. has spent years trying to become a world leader in data processing, a pitch that led it to lobby U.S. officials to allow it to buy AI chips for its data centers.

Having made plenty of noise about OpenAI's decision to take a spot in a 5-gigawatt data center complex in May 2025, U.A.E. officials have since seen those plans stall. The WSJ points out that no lease has been signed well over a year later. OpenAI was expected to take 20% of the data center's total capacity, the kind of loss that is sure to sting officials if plans don't get back on track.

For its part, the U.A.E. is clearly well aware of the threat that the ongoing regional instability poses to its ambitions, and it's looking for ways to address the fears of potential partners. Officials have already hinted at the potential for underground data centers. While the move would make an already costly endeavor even more expensive, it could help defend new data centers from drone attacks. Building drone defense systems near data centers is another option.

Drone attacks can be catastrophic for any building, and data centers are no different. The data centers operated by AWS suffered fires and damaged infrastructure as a result of the strikes. In their efforts to fight the fires caused by drone strikes, emergency workers used water, causing further damage.

The Abu Dhabi and Bahrain data centers are relied upon by a variety of businesses in the banking and financial sectors. Those businesses cannot tolerate downtime, especially of the protracted variety. There's little sign of data centers slipping from the crosshairs of drone operators any time soon, either, and AWS is far from unique in finding its data centers under attack.

Iran's Islamic Revolutionary Guard Corps (IRGC) claimed in April that it had hit a data center linked to Oracle in Dubai. That unit further threatened American tech companies, including Microsoft, Oracle, Alphabet, and Nvidia, this spring. It accused the companies of contributing to attacks on Iran. With that in mind, and with tensions in the Middle East looking unlikely to ease any time soon, it's perhaps unsurprising that AWS has asked customers to move data to safer data centers around the world.

The estimated natural gas consumption of data centers in the U.S. is expected to massively increase as these facilities increasingly rely on gas turbine generators for their power. According to Bloomberg, data centers are projected to use up to 15 billion cubic feet per day by 2035, a 117% increase from the previous forecast of 6.9 billion cubic feet. This number tracks with other data center forecasts, which suggest that data centers will use 20% of U.S. power by 2035, amounting to about 194 gigawatts.

Many data center projects have already been delayed by the lack of available power infrastructure, with power plants expected to take so much longer before they come online. It’s for this reason that many developments have turned towards onsite generators, so much so that AI demand is now compounding the jet engine shortage already plaguing the aviation industry.

Elon Musk was among the first to use gas turbines to power a data center when he deployed them at the Memphis Supercluster in 2024, even though he didn’t have permits for some of them. Now, it seems that the world’s richest man has seen this trend and has invested a billion dollars to buy a portable gas and diesel turbine leasing company. He even announced that SpaceX will start in-house turbine blade manufacturing to help cut down on the manufacturing bottleneck plaguing the jet engine industry.

New technologies like small modular reactors are currently being developed as an answer to AI data centers’ insatiable demand for power, like Ampera’s 3D-printed modular thorium nuclear reactor or Valar Atomics’ Ward 250 nuclear microreactor. Many AI hyperscalers, including Amazon, Google, Microsoft, Nvidia, and Oracle, have even invested in projects like these in a bid to generate massive amounts of clean energy for AI. However, they’re expected to take a few more years before they could become commercially viable — time that tech giants do not have. Because of this, Musk said that “natural gas will still be needed to supplement and bootstrap solar for several years.”

The deployment of natural gas turbines in data centers isn’t good news for the communities living around them, though. The NAACP said in its lawsuit against SpaceXAI that the use of these turbines at Colossus 2 increased nitrogen oxide exhaust by 111%, PM2.5 particles by 83%, and formaldehyde emissions by 88%. While the company has already pledged to remove all its unpermitted generators, the process will take at least a year as the portable turbines are slowly being replaced by a 1.2-gigawatt on-site power plant.

Aside from this, the massive demand for natural gas could potentially put a strain on the supply, causing prices to rise and hit the average consumer. Domestic natural gas producers are projected to raise their output by 35 billion cubic feet per day in the next decade, but this still falls short of the forecasted demand by around 11 billion cubic feet per day. So, unless output manages to catch up with the demand, prices are expected to shoot up and cause a scramble for available supply. Still, some experts suggest that there are still more than enough undeveloped gas fields within the U.S. to allow the industry to increase natural gas supplies and reduce costs.

Residents of Frederick County, Maryland, could be the beneficiaries of what is purported to be the biggest residential benefits package yet to be offered by data center developers, in a move a new report claims is a sign of a growing trend that Big Tech is trying to get ahead of fears and community pushback surrounding AI infrastructure. The $110 million deal includes new schools, water reclamation, and more, The Information reports.

According to the report, the Frederick Digital Campus offering could be a sign that data center developers like Amazon, Microsoft, and Oracle are wising up to growing residential pushback and concerns around the building of large AI data centers in their communities, with developers "sweetening financial offers to municipalities and regulators to gain approval for new facilities" while "getting smarter" about ensuring they shoulder the cost of utilities like electricity. The report says AI builders are turning towards tangible benefits, rather than rhetoric, to get their projects approved.

The Maryland site, if approved, would see residents of Frederick County benefit from a $110 million investment in total, including a $30 million elementary school, $40 million of recreational facilities, a $14.5 million workforce training center, and a further $10.5 million for "agricultural preservation." That comes on top of a purported $215 million in annual property taxes the campus would pay upon completion, a 40% uptick in tax revenue.

It appears the developers have also offered concessions regarding construction, reducing the square footage by almost 20%, and reducing potable water (water safe for human consumption and use) usage by 80%, with up to $100 million also proposed for a water reclamation system.

The proposal is yet to be approved, but if passed, the campus would boast Amazon and Aligned Data centers amongst its tenants. The report reiterates the deal "reflects a rapidly emerging consensus by both tech companies and host governments to eliminate giveaways to developers and to accelerate benefits to towns in the vicinity of the facilities."

A further cited example from Pennsylvania claims AWS announced it would not seek any economic incentives to reduce the tax burden on its 4.5GW, 36-building data center campus in Homer City.

The report further cites occasions where big tech companies are taking the side of consumers and residents over power rate debates, with Microsoft recently said to have challenged an American Transmission Co. and We Energies’ proposal for its Wisconsin data center, claiming the plan wasn't robust enough to protect retail customers from footing the bill if demand was lower than expected. In another case, Google and Amazon are said to have lobbied Virginia regulators to ensure they would fund transmission upgrades required for their infrastructure, rather than let an energy company cover the cost by marking up customer bills.

With concerns around data center buildouts impacting local water supplies, energy rates, and even contributing to noise pollution, it's clear that Big Tech companies appear to be trying to grease the wheels on a local level by investing more directly in some local communities. Big Tech has reportedly now spent more than $1 trillion on AI infrastructure, so even local investments to the tune of hundreds of millions of dollars are a drop in the ocean for companies.

Regulators are trying to pump the brakes on data center buildouts, with some 500 data centers on hold in the US because of various moratoriums and legal pauses.

Patagonia, a region of Argentina with 17,300 glaciers, has the potential to turn into a data center hotspot. The cool climate and abundant energy resources there are drawing interest from the likes of Amazon, Google, and OpenAI, according to Reuters. However, there needs to be more work on tapping into the resources and equipping high-speed networking, like fiber, to make the place irresistible. There’s also the question of political stability with a presidential election next year.

Argentina, and Patagonia in particular, has some juicy lures for potential hyperscale data center operators, and the cold climate is a bonus. That environment helps keep data centers cool and use less power while maintaining agreeable thermals. Patagonia has tens of thousands of glaciers across the Northern Patagonian Ice Field and the Southern Patagonian Ice Field combined.

Another bonus to would-be Patagonian data center operators are the range of energy sources there. Reuters mentions hydroelectric power, wind, and shale gas, but some operators who are already established are also adding solar.

Then, the social and political climate looks good for data centers in Argentina. According to the source, the populace doesn't hate data centers yet. At least there seems to be no campaign groups or public backlash. It is also reported that President Javier Milei has been making changes to court Big Tech. That’s a positive, and may be a negative. Positive in that the Argentines are now being accommodating to the needs of huge AI data centers. Negative, as there’s a presidential election next year, where things could take a dramatic turn, or not. There are also some regions in Patagonia where data center expansions may be complicated by the wishes and needs of indigenous communities. Businesses offering sizable, long-term investments aren’t keen on uncertainty.

With the pretty good environment for hyperscale buildouts established, it isn’t that surprising to see Reuters say OpenAI has confirmed a $25B and 500MW clean energy data center in the country. Lesser-known (to us) companies like Poland's Green Capital also plan a $3B and 300MW data center in Patagonia. It is also reported that FlexDomes, a U.S.-based green data center company, is planning the first stage of a 120MW facility.

Meanwhile, in neighboring South American countries, there are large developments from the likes of Amazon, Google, and others. Argentina must have these data center whales in its sights, as they are known to be scouting for expansion into new locations.

As well as political / business risk, Argentina has some structural weaknesses that need to be addressed by those developing in the country. It is said to have connectivity and power grid infrastructure deficiencies, for example. In other locations, we have seen these addressed on almost a per-project basis, with some state intervention, but it adds some high costs to setting up.

Data center builds are one of the hotly contested items worldwide. Many states and cities have upheld moratoriums on new buildouts due to power usage, water consumption, and noise concerns. Thailand is the latest nation to pump the proverbial brakes, with the government requesting that existing buildouts hit the pause button while coming up with a more precise regulatory framework.

Government heads requested that agencies compile information about current and future data centers during this week, in a bid to create unified legislation. The country currently has few laws specific to data centers, leading to legal voids like zoning a data center as a "warehouse" right next to a hospital. Much like everywhere else, the country has seen growing complaints about data centers' water and power usage.

After the week is out on September 11, the government expects to take about a month to come up with a regulatory draft, making the pause technically an indeterminate timeframe — though further delays wouldn't benefit either party, as Thailand considers the industry critical to the country’s competitiveness. The catch is that pausing construction isn't enforceable, so companies can elect to plow ahead regardless while the legislation is discussed.

NESDC (National Economic and Social Development Council) secretary-general Danucha Pichayanan is very specific: "we don't have the power to suspend the construction of the 49 data centers" currently under construction. However, when the legislation arrives, it will apply retroactively to ongoing projects, though with an adjustment period for in-progress builds. New builds will naturally need to comply with the legislation from the get-go.

Some companies may elect to soldier on with the belief their buildouts will be in compliance with the reasonably predictable content of the new laws, or betting that they could win future legal challenges — a dynamic that's already in play elsewhere with Project Jupiter. The incoming legislation is expected to cover power consumption (and possibly generation), closed-loop cooling, and water-surplus guarantees, meaning builders already have a fairly good idea of what they'll need to do.

Not complying with the request for a pause might be a game of political chicken, though, as the government may retaliate with regulatory delays and additional costs to power connections. Just last Thursday, Thailand's energy ministry raised concerns over a Bangkok data center that may be planning to hold diesel stockpiles far in excess of the 200,000 liters it has permission to.

On that topic, Thailand revised laws on industrial power delivery last July, and among other requirements, demands a bank guarantee of ฿4.5 million ($134,000) per megawatt. Half the money will be refunded when the actual usage reaches 50% of proposed utilization, and the rest when usage hits 70%. This mechanism is meant to stop preemptive allocation of power delivery capacity that might remain unused for months, years, or not at all — once again, mirroring datacenter playbooks elsewhere around the world.

At face value, the amount of $134k per megawatt sounds like a small price to pay for a builder to get ahead of the pack, but its per-megawatt nature means that a hyperscale facility pulling 200 MW or more needs to place around $27 million in escrow.

With the power laws recently revised, water consumption is seemingly the largest concern, as the current legal framework allows companies to make deals directly with utilities without additional safeguards. For example, in Chonburi, an operator signed a decade-long deal with Eastwater Stecon Utilities for 3.3 million cubic meters annually, or about 36,900 residents. It's expected that the new laws will add much stricter requirements to avoid localized deserts.

A Bitcoin mining data center in El Reno, Oklahoma, has been condemned after it leaked 3 million gallons of water, reports Koco TV News. Schools, businesses, and city and county offices were all closed last week due to the leak.

According to the source report, the Athlon Blockchain LLC-owned data center shouldn’t even have been operating, as the city issued it a Stop Work Order in 2023 due to “multiple fire and life safety code violations as well as the expiration of the building and electrical permits that had previously been issued.” The city asserts that none of the costs related to the water leak at the property will be passed on to citizens, according to the source.

We don’t know whether the Bitcoin mining data center development went ahead despite the Stop Work Order in 2023 due to ignorance or a devil-may-care attitude toward regulation. Koco reports that the owner of the property was given until the end of 2023 to comply with safety codes and obtain new permits, while construction should have been put on hold. Now it is believed that the firm building the facility set up the data center and began operations without any further consultation with the city. Perhaps the owners thought they could sidestep the Stop Work Order by housing the data center in an array of shipping containers.

Clearly, regulators should check that codes, directives, and ordinances are obeyed. This incident shows a disastrous leak from an insufficiently policed operator with serious repercussions for local citizens. Thus, the city has now committed to routinely reviewing compliance. As for the waste and costs involved in this case, a statement on behalf of the city says it is “preparing to take the necessary legal actions to recoup all costs and expenses related to the water leak from the property owner, including city resources expended in locating and addressing the leak.”

There has been a lot of controversy recently regarding data center usage of water. Concerns have been heightened due to widespread drought conditions in the Northern Hemisphere this summer.

Denver Water, which supplies the city and its surrounding areas, has imposed Stage 1 drought restrictions in its service area starting August 2026 as water supplies continue to drop. According to Gadget Review, commercial properties like data centers can only water grass twice a week — on Tuesdays and Fridays — until September 30. Starting October 1, lawn sprinklers will be banned for use until the Board of Water Commissioners declares otherwise. However, one resident who was out walking their dog saw that the new AI data center built close to their home was watering its lawn and using so much water that some areas have started to puddle.

“We can’t water our lawns until next summer, at the earliest, but you better believe that they are watering their entire green with so many sprinklers that it’s already making a giant pond, but I can’t water my f****ng tomatoes. I hate it here,” one resident claimed in the video.

This has understandably raised a lot of outrage, especially with all the widespread, bipartisan resistance that these data centers have been facing. However, there are still a lot of unknowns when it comes to the video. Firstly, we do not know when the video was shot; there’s a possibility that it was recorded on either a Tuesday or a Friday, meaning the site was well within its rights to water its lawns, at least until September 30.

The publication also noted several other possibilities: it could be using recycled, non-potable water, which Denver Water explicitly states, “are approved to irrigate any day of the week”; it could hold a variance from the utility provider, granting them permission to water their grass under “special circumstances,” or it was just plainly violating the law. Without further information, there’s just no way we can determine what is going on based on a single video. Nevertheless, this compounds the negative perception that data centers are already facing. Although closed-loop cooling technologies like those from Microsoft and Nvidia have essentially cut data center water consumption for cooling to zero, we still cannot get around the fact that several sites have already caused water issues for their neighbors.

Many people are raising issues about data centers’ massive resource consumption and the other negative effects like noise and air pollution that they could potentially bring to their surroundings. It’s for these reasons that Denver, and hundreds of other jurisdictions, have passed temporary bans and moratoriums on data center development.

Independence City Council member John Perkins lost his position after 68.05% of voters opted to oust him in a special election because he supported data centers. According to NBC News, Perkins was one of the five councilors who said yes to the tax breaks amounting to more than $6 billion for the Nebius data center being built in the area.

Many members of the community have been campaigning against this development ahead of the March City Council meeting when these tax breaks were approved, but the council ended up voting 5-2 anyway in favor of the data center. Perkins is the third to have lost their seat after this vote, with Jared Fears, who was running for re-election in the city council, and Bridget McCandless, who was running for mayor, losing their respective bids.

“Politicians aren’t entitled to elected positions. They’re elected by the people to serve those people. When they stop serving the people they were elected to represent, they’re voted out of office,” a member of the Stop the AI Data Center in Independence Facebook group said. “This is democracy in action.” While Perkins hasn’t released a statement yet after the release of the unofficial results of the Special Recall Election, he previously said in an Op-Ed on The Kansas City Star that he stands by his decision to support the Nebius tax breaks and that recalling him would be a setback for the city.

“You can disagree with that call, and I respect that. However, disagreement isn’t a crime, and I don’t believe it is grounds to attempt to overturn an election that happened two years ago,” Perkins wrote. “One vote is not a reason to erase the previous hard work I’ve done for my district.”

Despite the massive support for ousting Perkins, we should also note that the voter turnout for this issue has been abysmally small. Out of the 14,250 registered voters in Jackson County, only 1,643, or 11.5%, went to the polls to vote for or against the councilor.

Another Missouri town has voted out half of its city council, which approved a data center project within its jurisdiction. The people of Festus, Missouri, ousted four of the eight members of its council in a regular election in April 2026 after they approved a 360-acre data center project in the area. Another petition was raised to have the remaining four members recalled, but the measure was shot down by the council.

The popularity of data centers has fallen across the country, with the number of bans and moratoriums over these developments exceeding 500, spread across towns, cities, counties, and even states. Interestingly, the pushback against these developments straddles beliefs and political affiliation, making it harder for hyperscalers to get new projects off the ground.

OpenAI CEO Sam Altman has downplayed the water consumption issue that AI data centers are facing, saying that it has been blown out of proportion over time and does not keep up with the latest developments in data center cooling technology. Altman said this in an episode of the Sources Podcast with Alex Heath, where the two talk about the backlash against AI development. He claimed that producing a single almond in California is equivalent to 38,000 ChatGPT queries.

“You know, I saw this thing going around about the water usage of ChatGPT. And it was like every time you run a single ChatGPT query, it’s like, you know, you run your shower for like six hours. And the water never comes back, and it’s just done. I don’t have the exact calculation in front of me, but I think the real number is something like — doing this from memory, it might be wrong, but it’s close — for every 38,000 ChatGPT queries, that is the same amount of water that is used in the production of a single almond in California… This is like the full-on, you know, total true water accounting, not just what’s running in one data center,” Altman told the host.

“There’s a question of like where this came from because the people that are scarfing down 12 almonds at a time don’t feel like they’re doing something horrible from a water perspective, for the most part. It is true that data centers at one point used evaporative cooling, but they have not done that in a long time. Like if you look at a modern very large data center, it uses the equivalent amount of water as an office building in terms of, you know, people like running the sinks and the toilets and whatever. So, that has been a robust meme and difficult to disprove, but I don’t think it holds up to any scrutiny.”

It’s true that almonds are some of the most water-intensive crops being grown in the U.S. Researchers say that it takes about 1.1 gallons of water to grow a single almond. By comparison, Altman's own statements in the past indicate that one ChatGPT response uses somewhere between 1 to 50mL of water, ostensibly using older evaporative cooling techniques, or about 0.0002 to 0.013 gallons per query. While this is nowhere near Altman’s 38,000 estimate, it’s still about 85 to 5,500 ChatGPT queries per almond using an unspecified cooling method. Nevertheless, training AI models in older facilities can use far more water, with estimates suggesting that GPT-3 used up almost 185,000 gallons of water with older systems, while the increasing popularity of AI, especially AI agents which use up a lot more tokens than just chatbots, means that data centers could still use more water compared to almond farming.

In fact, multiple data centers across the United States have faced issues with their water consumption, with one site in Fayette County, Georgia, using up 29 million gallons over a span of 15 months and another development in Morgan County, Georgia, allegedly causing the water of its neighbors to turn muddy. Altman admitted that data centers that used older technologies like evaporative cooling used a lot of water, but he also mentioned that new techniques and technologies have been developed that cut down data center water use. Some examples of this include Nvidia’s liquid cooling system that runs “hotter than a hot tub,” which cuts water use by 100%, and Microsoft’s closed-loop cooling systems, which use the same amount of water as a restaurant. Amazon even claimed that its data centers consume only 0.075% of the water that Americans use for their lawns and gardens.

You can watch the complete podcast episode below.

GoPro just announced a merger with Starman Optical, Inc., an optical-photonics company, saying that the deal would allow the combined entity to produce U.S.-made optical transceivers for use in AI data centers. According to the press release, Starman will pay shareholders $285 million in cash for the action camera company, amounting to $1.14 per share, and will also pay off its outstanding debt of $92 million. Despite that, current GoPro investors will retain 10% ownership of the post-merger company.

Starman said that the move will give it access to GoPro’s intellectual property, which amounts to more than 2,500 patents, as well as its imaging solutions, advanced optics, and other technologies that it has developed. This deal is crucial in keeping the camera company afloat, especially as it has recently faced declining sales and increasing costs, leading the camera maker to warn investors that there is “substantial doubt about the company’s ability to continue.”

On the other hand, the merger would expand GoPro’s portfolio to include optical transceivers, which are getting increased demand as AI data centers transition to fiber optics for their networking needs. Aside from data centers, the deal would also allow GoPro to expand into defense, government, robotics, and the aerospace industries, where it “intends to leverage its IP, optics, and imaging capabilities” to supply the increasing demand for solutions made within the U.S.

While GoPro has successfully moved its manufacturing out of China to protect its supply chain from U.S. tariffs, it still has no local manufacturing. So, this merger would allow it to have production within American borders and help isolate it from further economic instability brought about by surprise tariff announcements.

This could help it secure its future, especially as the FCC has recently been making moves to keep foreign-made tech like routers, LiDAR-equipped drones, and robots. It also makes sense for the two companies to partner — even though consumer action cameras and optical transceivers for data centers sound like they have nothing in common, they share a lot of technologies between them, like imaging sensors and advanced lenses or optics.

This isn’t the first company with seemingly no relation to data centers to expand or pivot towards it. Struggling shoemaker and apparel brand Allbirds pivoted towards it earlier this year, resulting in a 580% jump in its stock price after the announcement, although it has since settled back down to its pre-announcement prices. AI data centers are indeed a tempting investment, with consultancy firm PricewaterhouseCoopers saying that investment on the technology could hit $32 trillion by 2050.