Menu
SUBSCRIBE
SUBSCRIBE
Menu

AI Data Centers

Rows of lit server racks in a data center hall under a blue-green grid overlay, for the AI Energy Management Alliance

Nvidia, Google, Emerald AI Push Faster Power Access For AI Data Centers

, , , , , , , , , , ,

The AI Energy Management Alliance wants utilities and regulators to reward data centers that can reduce electricity use when the grid is strained.

 

Nvidia, Google, and Emerald AI launched an industry group that wants local utilities to supply electricity to new AI data centers more quickly if they agree to use less power when local demand is highest. The alliance says some new data centers currently wait five to 10 years for utilities to provide enough electricity for them to open.

The AI Energy Management Alliance is proposing industry standards and advocating for changes to government and utility policies. It is not announcing a new rule, and members have not made binding commitments to reduce their electricity use.

Alliance proposes a trade for faster power

AI data centers use large amounts of electricity. When local electricity demand is highest, some utilities may not have enough power to serve new data centers along with existing homes and businesses. Utilities can therefore delay supplying new data centers until more power is available.

The alliance wants utilities to supply electricity to new AI data centers sooner if those data centers agree to use less power during high-demand periods and allow the utilities to confirm the reductions.

A “flexible data center” could reduce its use by postponing some computing work, switching to batteries, or producing some of its own electricity. It would tell the utility in advance how much power it could cut and for how long, and the utility would track whether it followed through.

The alliance has not said how much electricity a data center would have to cut to qualify for faster access.

Study estimates power systems could serve more demand

The alliance cites Duke University research estimating that existing U.S. power systems could serve nearly 100 gigawatts of additional demand if large electricity users agreed to reduce their use during limited periods of high demand.

The estimate does not mean that 100 gigawatts of electricity is currently sitting unused. It represents the additional demand that researchers calculated existing power systems could handle if some large customers temporarily used less electricity when supplies were stretched.

How much additional demand could be served would depend on the location, the available electricity supply, and how much power customers could stop using when asked.

The alliance says those reductions could allow utilities to delay some costly upgrades while providing electricity to new data centers more quickly. The Duke study does not establish that every power system could accommodate more data centers under the same conditions.

Companies have already tested power reductions

The underlying approach is not new, although it is not yet standard across the data center industry.

Google said in March that agreements with several U.S. utilities now allow it to reduce or postpone up to 1 gigawatt of data center electricity use. Google said it can shift selected AI computing tasks when a utility needs to reduce local demand.

A field test published in Nature Energy found that an AI computing system at a commercial data center in Phoenix reduced its electricity use by 25% for three hours without missing its promised performance levels. Emerald AI led the test with Nvidia, Oracle, two Arizona utilities, and the Electric Power Research Institute.

Nvidia and Emerald AI said they tested systems that adjust data center electricity use at five commercial sites. Nvidia said in March that it planned to deploy its DSX Flex software on a larger commercial scale later in 2026 at its AI Factory Research Center in Manassas, Virginia. The software would allow the center to shift computing work when a utility asks it to reduce electricity use.

Alliance includes AI companies, utilities, and power producers

Nvidia, Google, and Emerald AI are the alliance’s founding members. The alliance includes 20 companies and organizations from the AI, data center, utility, and power industries, according to Axios. Participants include Anthropic, AES, National Grid, Constellation Energy, NRG, and RWE.

The organization is a relaunch of the Advanced Energy Management Alliance, formed in 2014 to promote programs that pay large electricity customers to reduce use when needed. The relaunched organization is focused specifically on AI data centers.

The alliance plans to work with its members, including utilities, to decide how data centers’ promised power reductions would be measured and confirmed. It would then ask utilities and regulators to use those rules when deciding which data centers could receive power sooner. The rules would apply only where a utility or regulator chose to adopt them.

Red and blue cutaway illustration of a data center beside power lines, on AI critical infrastructure designation

Advocacy Group Asks The President To Designate AI As Critical Infrastructure

, , , , , , , , ,

A critical infrastructure designation would give AI companies 72 hours to report a cyberattack, put their data centers under federal security standards, and name the firms the government defends first.

 

Key Takeaways

  • Americans for Responsible Innovation (ARI) asked the President to add AI to the critical infrastructure list, which already covers power, water, banking, and hospitals.
  • A 2022 law requires critical infrastructure companies to report cyberattacks within 72 hours, and a designation would extend that duty to AI companies. The law is not in force for anyone yet because CISA has not yet written the regulation that puts it into effect.
  • ARI says the need for critical infrastructure designation is urgent as adversaries are already attacking AI infrastructure.
    • ARI cited recent Iranian drone strikes that destroyed two Amazon data centers in the United Arab Emirates as proof.

 

Americans for Responsible Innovation, a Washington nonprofit that lobbies on technology policy, published a report on August 20 asking the President to add AI to the federal critical infrastructure list. Sixteen sectors sit on that list today.

ARI recommends including in the designation companies that build and train AI models, the data centers powering them, AI chip designers and manufacturers, and the platforms selling access. A critical infrastructure designation puts one federal agency in charge of the sector, requires its companies to report attacks, and ensures the government tells those companies what it knows about threats aimed at them.

Attacks on AI infrastructure have already done damage

ARI cites two attacks as examples of the need for designation. On March 1, Iran flew drones into two Amazon Web Services data centers in the United Arab Emirates, setting fires and cutting power. Banking and food delivery apps across the UAE stopped working, and Amazon’s status page still showed problems eleven days later.

The previous July, someone offered a routine code contribution to the public repository behind Amazon Q Developer, a coding assistant used inside a programmer’s editor. Amazon merged it, and a badly protected credential in its build system handed the contributor administrator access. They planted an instruction in the next official release telling the assistant to wipe the computer it ran on and delete the customer’s cloud account. Amazon published that release to the marketplace, where the assistant was downloaded more than 964,000 times. The instruction never ran, because whoever wrote it made a syntax error.

ARI says one corrupted model could damage many industries at once

AI now runs inside operations across finance, health care, energy, and government. Those companies mostly build their own AI tools, but nearly all of those tools sit on a small number of underlying models, running in data centers owned by a handful of cloud companies.

That is the risk ARI describes. An attacker who corrupts one widely used model does not damage one company. The corruption travels into everything built on that model.

This kind of attack is hard to notice. An AI system produces different output every time it runs, so a tampered one keeps answering and looks like it is working long after its answers stopped being reliable. And currently no rule requires the companies building those models to meet a security standard or disclose a breach, so the businesses running on them cannot spot the problem themselves, and nobody has to tell them. The government cannot see it either.

A designation would assign one federal agency to the sector, set security standards those companies must meet, and require them to report attacks, so a warning reaches the businesses downstream.

“AI systems are becoming increasingly integrated across industries, including health care, finance, and manufacturing, and without formal national security standards, an attack on these systems could lead to the incapacitation of these sectors at once,” said ARI policy analyst Jessica Maksimov, who wrote the report with Terrence Kelly.

AI companies would get 72 hours to report an attack

The designation would put AI companies under the Cyber Incident Reporting for Critical Infrastructure Act of 2022, which gives companies in listed sectors 72 hours to tell the government they were attacked. AI companies currently have no federal reporting duty.

That requirement is not in force for anyone. A law of this kind takes effect only once the responsible agency writes the regulation that carries it out, and CISA, the federal cybersecurity agency, has missed its deadlines to do so. It now expects to publish this fall.

CISA and the Center for AI Standards and Innovation (CAISI), the federal office that writes AI security guidance, would then publish security standards for AI data centers. ARI wants Congress to fund a CISA team to audit them.

A short list of AI companies would get government attention first

The agency would name the systemically important AI companies whose failure would do the country the most damage. The report says government protection, security requirements, and help reducing risk should start with these companies ahead of everyone else in the sector.

The last two sectors that asked were turned down

In 2024, the President declined to label space systems or the bioeconomy, meaning goods made from biological processes, as critical infrastructure. Officials said space systems were already spread across other sectors on the list, and that the bioeconomy’s risks were not sufficiently different from the risks those sectors already cover.

Facebook data center in Oregon under U.S. and state flags, accompanying coverage of a proposed Data Center Bill of Rights

‘Data Center Bill of Rights’ Would Give Communities Power Over AI Projects

, , , , , , , ,

The nonbinding House resolution proposes 10 protections, including limits intended to keep households and small businesses from paying higher electricity costs.

 

Rep. Ro Khanna (D-Calif.) introduced a House resolution calling on Congress to enact a “Data Center Bill of Rights” that would give communities more power to reject, restrict, or pause AI data center projects.

House Resolution 1471 is nonbinding. It would express the position of the House but would not create federal requirements, enforcement powers, or deadlines. Congress would need to pass separate legislation to put the proposed protections into law.

Communities could reject or restrict projects

The resolution says communities should have the right to reject a proposed AI data center through a public process.

Counties and municipalities would retain the power to prohibit or regulate data centers without being overruled by their states. States could also pause new projects until they adopt protections for electricity customers, water supplies, and community involvement.

The proposal would allow data centers to be banned in residential areas and within 2,500 feet of homes, schools, childcare facilities, hospitals, and nursing homes.

Utility customers would receive cost protections

The proposed Data Center Bill of Rights calls for households and small businesses to be protected from higher electricity rates or fees caused by data centers.

Communities would also receive an independent public report before a project is approved. The report would assess the proposed data center’s effects on electricity demand, utility bills, public health, pollution, housing and consumer prices.

It would also disclose the number and types of jobs the facility is expected to create and its projected economic effects over 10 years.

The resolution says data center construction can create hundreds of jobs, but even the largest facilities often employ fewer than 150 permanent workers.

Water use and public subsidies would face scrutiny

Data center operators would disclose their expected water consumption to local water providers before receiving approval. The water provider would determine whether the project is compatible with the local supply, and operators would release water-use reports every six months.

The resolution also calls for limits on noise and air pollution, clean and reliable energy requirements, and agreements requiring data center developers to support local housing, infrastructure, training and permanent technology jobs.

Data center owners, operators and developers would pay state and local taxes. Any tax exemption or public subsidy would include labor rules, public disclosure requirements and provisions allowing the government to recover benefits when those conditions are not met.

The resolution was referred to the House Energy and Commerce and Judiciary committees on Aug. 3.

White dotted circular AI emblem on a blue background, the official graphic of the UN Global Dialogue on AI Governance

United Nations Urges Governments to Accelerate AI Governance as Risks Cross Borders

, , , , , , , , ,

Speaking at the inaugural UN Global Dialogue on AI Governance, Secretary-General António Guterres said governments must move more quickly to establish international AI safeguards before the technology outpaces existing rules.

 

The United Nations is urging governments to move faster on AI governance, warning that the technology is now too important and too fast-moving for countries to regulate independently.

Opening the inaugural UN Global Dialogue on AI Governance in Geneva, UN Secretary-General António Guterres said AI should be governed through international cooperation built on safety, accountability, and fairness. He argued that governments have a limited window to shape how AI develops before its impacts become more difficult to manage.

“We may be the last generation able to set the terms on which humanity and machines coexist,” Guterres said.

The two-day dialogue brings together government officials, AI researchers, technology companies, civil society organizations, and technical experts to discuss how countries should respond to increasingly capable AI systems. A second Global Dialogue is scheduled for May 2027 in New York.

The UN says AI governance cannot wait

Guterres warned that AI is already reshaping economies and societies with increasing influence on issues ranging from national security to public services. He also pointed to the growing use of AI-enabled military technologies, noting that AI chips originally designed for civilian applications are now deployed on the battlefield, where autonomous weapons pose an imminent concern.

The Secretary-General argued that international cooperation is essential because fragmented national approaches could make AI harder to govern.

“When countries align on how to test systems, measure risk and assign responsibility, safety travels with the technology,” he said. “When they do not, a patchwork of incompatible rules raises costs, divides the world — and protects no one.”

Three priorities for global AI governance

Rather than outlining detailed regulations, Guterres grouped his proposals around several broad priorities for future international AI governance.

  • Protecting people: The Secretary-General called for a global AI Child Safety Pledge to strengthen safeguards and accountability for AI systems that are used by or affect children.
  • Human accountability for human rights: Guterres said human rights must remain central to AI governance, arguing that AI should never replace human responsibility for decisions involving justice, healthcare, or policing. “Machines can inform, but humans must decide — and answer,” he said.
  • Expanding global participation: The Secretary-General called for greater public investment to help developing countries build AI expertise and infrastructure, warning that today’s digital divide could become an AI divide. He also announced a new UN-backed initiative, supported by more than 20 countries, to help nations share knowledge and strengthen their AI capabilities.

Transparency and environmental impact

Guterres called on frontier AI developers to publicly disclose the environmental footprint of their AI systems, including carbon emissions, water use, and land consumption. He urged companies to power AI data centers with renewable energy by 2030, warning that their resource demands are growing rapidly.

According to the Secretary-General, AI data centers could consume more electricity than all but five countries by the end of the decade while using enough water to meet the annual needs of the entire population of sub-Saharan Africa.

The Global Dialogue forms part of the UN’s broader effort to build international consensus on AI governance. AI development outpacing governance is a central theme of the summit. Further, officials repeatedly emphasized that AI’s benefits should be shared more broadly while ensuring that increasingly capable systems remain safe, accountable, and subject to meaningful human oversight.