Stay Ahead, Stay ONMINE

A New York legislator wants to pick up the pieces of the dead California AI bill

The first Democrat in New York history with a computer science background wants to revive some of the ideas behind the failed California AI safety bill, SB 1047, with a new version in his state that would regulate the most advanced AI models. It’s called the RAISE Act, an acronym for “Responsible AI Safety and Education.” Assembly member Alex Bores hopes his bill, currently an unpublished draft—subject to change—that MIT Technology Review has seen, will address many of the concerns that blocked SB 1047 from passing into law. SB 1047 was, at first, thought to be a fairly modest bill that would pass without much fanfare. In fact, it flew through the California statehouse with huge margins and received significant public support. However, before it even landed on Governor Gavin Newsom’s desk for signature in September, it sparked an intense national fight. Google, Meta, and OpenAI came out against the bill, alongside top congressional Democrats like Nancy Pelosi and Zoe Lofgren. Even Hollywood celebrities got involved, with Jane Fonda and Mark Hamill expressing support for the bill.  Ultimately, Newsom vetoed SB 1047, effectively killing regulation of so-called frontier AI models not just in California but, with the lack of laws on the national level, anywhere in the US, where the most powerful systems are developed. Now Bores hopes to revive the battle. The main provisions in the RAISE Act include requiring AI companies to develop safety plans for the development and deployment of their models.  The bill also provides protections for whistleblowers at AI companies. It forbids retaliation against an employee who shares information about an AI model in the belief that it may cause “critical harm”; such whistleblowers can report the information to the New York attorney general. One way the bill defines critical harm is the use of an AI model to create a chemical, biological, radiological, or nuclear weapon that results in the death or serious injury of 100 or more people.  Alternatively, a critical harm could be a use of the AI model that results in 100 or more deaths or at least $1 billion in damages in an act with limited human oversight that if committed by a human would constitute a crime requiring intent, recklessness, or gross negligence. The safety plans would ensure that a company has cybersecurity protections in place to prevent unauthorized access to a model. The plan would also require testing of models to assess risks before and after training, as well as detailed descriptions of procedures to assess the risks associated with post-training modifications. For example, some current AI systems have safeguards that can be easily and cheaply removed by a malicious actor. A safety plan would have to address how the company plans to mitigate these actions. The safety plans would then be audited by a third party, like a nonprofit with technical expertise that currently tests AI models. And if violations are found, the bill empowers the attorney general of New York to issue fines and, if necessary, go to the courts to determine whether to halt unsafe development.  A different flavour of bill The safety plans and external audits were elements of SB 1047, but Bores aims to differentiate his bill from the California one. “We focused a lot on what the feedback was for 1047,” he says. “Parts of the criticism were in good faith and could make improvements. And so we’ve made a lot of changes.”  The RAISE Act diverges from SB 1047 in a few ways. For one, SB 1047 would have created the Board of Frontier Models, tasked with approving updates to the definitions and regulations around these AI models, but the proposed act would not create a new government body. The New York bill also doesn’t create a public cloud computing cluster, which SB 1047 would have done. The cluster was intended to support projects to develop AI for the public good.  The RAISE Act doesn’t have SB 1047’s requirement that companies be able to halt all operations of their model, a capability sometimes referred to as a “kill switch.” Some critics alleged that the shutdown provision of SB 1047 would harm open-source models, since developers can’t shut down a model someone else may now possess (even though SB 1047 had an exemption for open-source models). The RAISE Act avoids the fight entirely. SB 1047 referred to an “advanced persistent threat” associated with bad actors trying to steal information during model training. The RAISE Act does away with that definition, sticking to addressing critical harms from covered models. Focusing on the wrong issues? Bores’ bill is very specific with its definitions in an effort to clearly delineate what this bill is and isn’t about. The RAISE Act doesn’t address some of the current risks from AI models, like bias, discrimination, and job displacement. Like SB 1047, it is very focused on catastrophic risks from frontier AI models.  Some in the AI community believe this focus is misguided. “We’re broadly supportive of any efforts to hold large models accountable,” says Kate Brennan, associate director of the AI Now Institute, which conducts AI policy research. “But defining critical harms only in terms of the most catastrophic harms from the most advanced models overlooks the material risks that AI poses, whether it’s workers subject to surveillance mechanisms, prone to workplace injuries because of algorithmically managed speed rates, climate impacts of large-scale AI systems, data centers exerting massive pressure on local power grids, or data center construction sidestepping key environmental protections,” she says. Bores has worked on other bills addressing current harms posed by AI systems, like discrimination and lack of transparency. That said, Bores is clear that this new bill is aimed at mitigating catastrophic risks from more advanced models. “We’re not talking about any model that exists right now,” he says. “We are talking about truly frontier models, those on the edge of what we can build and what we understand, and there is risk in that.”  The bill would cover only models that pass a certain threshold for how many computations their training required, typically measured in FLOPs (floating-point operations). In the bill, a covered model is one that requires more than 1026 FLOPs in its training and costs over $100 million. For reference, GPT-4 is estimated to have required 1025 FLOPs.  This approach may draw scrutiny from industry forces. “While we can’t comment specifically on legislation that isn’t public yet, we believe effective regulation should focus on specific applications rather than broad model categories,” says a spokesperson at Hugging Face, a company that opposed SB 1047. Early days The bill is in its nascent stages, so it’s subject to many edits in the future, and no opposition has yet formed. There may already be lessons to be learned from the battle over SB 1047, however. “There’s significant disagreement in the space, but I think debate around future legislation would benefit from more clarity around the severity, the likelihood, and the imminence of harms,” says Scott Kohler, a scholar at the Carnegie Endowment for International Peace, who tracked the development of SB 1047.  When asked about the idea of mandated safety plans for AI companies, assembly member Edward Ra, a Republican who hasn’t yet seen a draft of the new bill yet, said: “I don’t have any general problem with the idea of doing that. We expect businesses to be good corporate citizens, but sometimes you do have to put some of that into writing.”  Ra and Bores co chair the New York Future Caucus, which aims to bring together lawmakers 45 and under to tackle pressing issues that affect future generations. Scott Wiener, a California state senator who sponsored SB 1047, is happy to see that his initial bill, even though it failed, is inspiring further legislation and discourse. “The bill triggered a conversation about whether we should just trust the AI labs to make good decisions, which some will, but we know from past experience, some won’t make good decisions, and that’s why a level of basic regulation for incredibly powerful technology is important,” he says. He has his own plans to reignite the fight: “We’re not done in California. There will be continued work in California, including for next year. I’m optimistic that California is gonna be able to get some good things done.” And some believe the RAISE Act will highlight a notable contradiction: Many of the industry’s players insist that they want regulation, but when any regulation is proposed, they fight against it. “SB 1047 became a referendum on whether AI should be regulated at all,” says Brennan. “There are a lot of things we saw with 1047 that we can expect to see replay in New York if this bill is introduced. We should be prepared to see a massive lobbying reaction that industry is going to bring to even the lightest-touch regulation.” Wiener and Bores both wish to see regulation at a national level, but in the absence of such legislation, they’ve taken the battle upon themselves. At first it may seem odd for states to take up such important reforms, but California houses the headquarters of the top AI companies, and New York, which has the third-largest state economy in the US, is home to offices for OpenAI and other AI companies. The two states may be well positioned to lead the conversation around regulation.  “There is uncertainty at the direction of federal policy with the transition upcoming and around the role of Congress,” says Kohler. “It is likely that states will continue to step up in this area.” Wiener’s advice for New York legislators entering the arena of AI regulation? “Buckle up and get ready.”

The first Democrat in New York history with a computer science background wants to revive some of the ideas behind the failed California AI safety bill, SB 1047, with a new version in his state that would regulate the most advanced AI models. It’s called the RAISE Act, an acronym for “Responsible AI Safety and Education.”

Assembly member Alex Bores hopes his bill, currently an unpublished draft—subject to change—that MIT Technology Review has seen, will address many of the concerns that blocked SB 1047 from passing into law.

SB 1047 was, at first, thought to be a fairly modest bill that would pass without much fanfare. In fact, it flew through the California statehouse with huge margins and received significant public support.

However, before it even landed on Governor Gavin Newsom’s desk for signature in September, it sparked an intense national fight. Google, Meta, and OpenAI came out against the bill, alongside top congressional Democrats like Nancy Pelosi and Zoe Lofgren. Even Hollywood celebrities got involved, with Jane Fonda and Mark Hamill expressing support for the bill. 

Ultimately, Newsom vetoed SB 1047, effectively killing regulation of so-called frontier AI models not just in California but, with the lack of laws on the national level, anywhere in the US, where the most powerful systems are developed.

Now Bores hopes to revive the battle. The main provisions in the RAISE Act include requiring AI companies to develop safety plans for the development and deployment of their models. 

The bill also provides protections for whistleblowers at AI companies. It forbids retaliation against an employee who shares information about an AI model in the belief that it may cause “critical harm”; such whistleblowers can report the information to the New York attorney general. One way the bill defines critical harm is the use of an AI model to create a chemical, biological, radiological, or nuclear weapon that results in the death or serious injury of 100 or more people. 

Alternatively, a critical harm could be a use of the AI model that results in 100 or more deaths or at least $1 billion in damages in an act with limited human oversight that if committed by a human would constitute a crime requiring intent, recklessness, or gross negligence.

The safety plans would ensure that a company has cybersecurity protections in place to prevent unauthorized access to a model. The plan would also require testing of models to assess risks before and after training, as well as detailed descriptions of procedures to assess the risks associated with post-training modifications. For example, some current AI systems have safeguards that can be easily and cheaply removed by a malicious actor. A safety plan would have to address how the company plans to mitigate these actions.

The safety plans would then be audited by a third party, like a nonprofit with technical expertise that currently tests AI models. And if violations are found, the bill empowers the attorney general of New York to issue fines and, if necessary, go to the courts to determine whether to halt unsafe development. 

A different flavour of bill

The safety plans and external audits were elements of SB 1047, but Bores aims to differentiate his bill from the California one. “We focused a lot on what the feedback was for 1047,” he says. “Parts of the criticism were in good faith and could make improvements. And so we’ve made a lot of changes.” 

The RAISE Act diverges from SB 1047 in a few ways. For one, SB 1047 would have created the Board of Frontier Models, tasked with approving updates to the definitions and regulations around these AI models, but the proposed act would not create a new government body. The New York bill also doesn’t create a public cloud computing cluster, which SB 1047 would have done. The cluster was intended to support projects to develop AI for the public good. 

The RAISE Act doesn’t have SB 1047’s requirement that companies be able to halt all operations of their model, a capability sometimes referred to as a “kill switch.” Some critics alleged that the shutdown provision of SB 1047 would harm open-source models, since developers can’t shut down a model someone else may now possess (even though SB 1047 had an exemption for open-source models).

The RAISE Act avoids the fight entirely. SB 1047 referred to an “advanced persistent threat” associated with bad actors trying to steal information during model training. The RAISE Act does away with that definition, sticking to addressing critical harms from covered models.

Focusing on the wrong issues?

Bores’ bill is very specific with its definitions in an effort to clearly delineate what this bill is and isn’t about. The RAISE Act doesn’t address some of the current risks from AI models, like bias, discrimination, and job displacement. Like SB 1047, it is very focused on catastrophic risks from frontier AI models. 

Some in the AI community believe this focus is misguided. “We’re broadly supportive of any efforts to hold large models accountable,” says Kate Brennan, associate director of the AI Now Institute, which conducts AI policy research.

“But defining critical harms only in terms of the most catastrophic harms from the most advanced models overlooks the material risks that AI poses, whether it’s workers subject to surveillance mechanisms, prone to workplace injuries because of algorithmically managed speed rates, climate impacts of large-scale AI systems, data centers exerting massive pressure on local power grids, or data center construction sidestepping key environmental protections,” she says.

Bores has worked on other bills addressing current harms posed by AI systems, like discrimination and lack of transparency. That said, Bores is clear that this new bill is aimed at mitigating catastrophic risks from more advanced models. “We’re not talking about any model that exists right now,” he says. “We are talking about truly frontier models, those on the edge of what we can build and what we understand, and there is risk in that.” 

The bill would cover only models that pass a certain threshold for how many computations their training required, typically measured in FLOPs (floating-point operations). In the bill, a covered model is one that requires more than 1026 FLOPs in its training and costs over $100 million. For reference, GPT-4 is estimated to have required 1025 FLOPs. 

This approach may draw scrutiny from industry forces. “While we can’t comment specifically on legislation that isn’t public yet, we believe effective regulation should focus on specific applications rather than broad model categories,” says a spokesperson at Hugging Face, a company that opposed SB 1047.

Early days

The bill is in its nascent stages, so it’s subject to many edits in the future, and no opposition has yet formed. There may already be lessons to be learned from the battle over SB 1047, however. “There’s significant disagreement in the space, but I think debate around future legislation would benefit from more clarity around the severity, the likelihood, and the imminence of harms,” says Scott Kohler, a scholar at the Carnegie Endowment for International Peace, who tracked the development of SB 1047. 

When asked about the idea of mandated safety plans for AI companies, assembly member Edward Ra, a Republican who hasn’t yet seen a draft of the new bill yet, said: “I don’t have any general problem with the idea of doing that. We expect businesses to be good corporate citizens, but sometimes you do have to put some of that into writing.” 

Ra and Bores co chair the New York Future Caucus, which aims to bring together lawmakers 45 and under to tackle pressing issues that affect future generations.

Scott Wiener, a California state senator who sponsored SB 1047, is happy to see that his initial bill, even though it failed, is inspiring further legislation and discourse. “The bill triggered a conversation about whether we should just trust the AI labs to make good decisions, which some will, but we know from past experience, some won’t make good decisions, and that’s why a level of basic regulation for incredibly powerful technology is important,” he says.

He has his own plans to reignite the fight: “We’re not done in California. There will be continued work in California, including for next year. I’m optimistic that California is gonna be able to get some good things done.”

And some believe the RAISE Act will highlight a notable contradiction: Many of the industry’s players insist that they want regulation, but when any regulation is proposed, they fight against it. “SB 1047 became a referendum on whether AI should be regulated at all,” says Brennan. “There are a lot of things we saw with 1047 that we can expect to see replay in New York if this bill is introduced. We should be prepared to see a massive lobbying reaction that industry is going to bring to even the lightest-touch regulation.”

Wiener and Bores both wish to see regulation at a national level, but in the absence of such legislation, they’ve taken the battle upon themselves. At first it may seem odd for states to take up such important reforms, but California houses the headquarters of the top AI companies, and New York, which has the third-largest state economy in the US, is home to offices for OpenAI and other AI companies. The two states may be well positioned to lead the conversation around regulation. 

“There is uncertainty at the direction of federal policy with the transition upcoming and around the role of Congress,” says Kohler. “It is likely that states will continue to step up in this area.”

Wiener’s advice for New York legislators entering the arena of AI regulation? “Buckle up and get ready.”

Shape
Shape
Stay Ahead

Explore More Insights

Stay ahead with more perspectives on cutting-edge power, infrastructure, energy,  bitcoin and AI solutions. Explore these articles to uncover strategies and insights shaping the future of industries.

Shape

IBM Cloud speeds AI workloads with Intel Gaudi 3 accelerators

For businesses that need more control over their AI development, IBM says they can deploy IBM watsonx.ai software with the Intel Gaudi 3-based virtual server on IBM Cloud VPC in Q2 2025. IBM watsonx.ai includes an end-to-end AI development studio, AI developer toolkit and full AI lifecycle management for developing AI services

Read More »

Gunvor Profits Dropped 42 Pct as Energy Market Chaos Eased

Commodity trader Gunvor Group posted a 42% drop in profit in 2024, as the energy-market volatility that drove mega earnings in previous years declined. Net profit fell to $729 million last year, while revenues grew due to higher volumes of physically traded crude and refined oil products, the trading house said in a statement on Tuesday. Though full-year earnings were the fourth best in Gunvor’s history, second-half profit was the lowest since 2020. After banking billions of dollars in windfall profits as markets whipsawed following Russia’s full scale invasion of Ukraine, commodity trading giants are adapting to a new normal. Prices for gas and coal have declined from multi-year highs, while crude oil futures are trading in a tight range. In an interview with Bloomberg last week, Gunvor’s co-founder and Chief Executive Officer Torbjörn Törnqvist acknowledged that the firm had lost money on some oil trades. The CEO has overhauled Gunvor’s top leadership team in recent months. “Some areas have really performed well; some areas did not perform well. I’m open with that,” he said. On Tuesday, the trading house said it was also impacted by impairments last year, including one for its oil refinery in Rotterdam. Like its trading peers, Gunvor has been investing in assets to help stabilize long-term profits. That includes buying into a gas power station in Spain, a fuel distribution network in Pakistan and upstream natural gas production in the US. Gunvor said its equity was $6.5 billion at the end of the year. Billionaire Törnqvist owns 84.79% of the company, with the remaining shares held by employees. WHAT DO YOU THINK? Generated by readers, the comments included herein do not reflect the views and opinions of Rigzone. All comments are subject to editorial review. Off-topic, inappropriate or insulting comments will be removed. MORE FROM THIS AUTHOR

Read More »

Oil Rises, but Doubts Dominate

Oil crept higher as traders braced for impending US tariff announcements and equity markets rose. West Texas Intermediate rose 0.7% to settle below $72 a barrel, while global benchmark Brent ended the session just below $75. US President Donald Trump is set to announce tariffs later Wednesday and the sweeping duties are set to take immediate effect. Several proposals are said to be under consideration, including a tiered system with a set of flat rates for different countries, as well as a more customized “reciprocal” plan. Equity markets also gained. The tariffs add to a deluge of conflicting drivers since Trump came into office. Sanctions threaten to curb supplies from Russia and Iran, while a production boost by OPEC and its allies starting this month is exacerbating concerns that a glut is looming. The US is threatening to further tighten the screws on Russia. A group of 50 Republican and Democratic senators introduced a sanctions package that would hit the third-largest oil producer and countries that buy its fuel if President Vladimir Putin refuses to engage in good-faith ceasefire negotiations with Ukraine or if he breaches any eventual agreement. “Supply risks are still top of mind,” said Rebecca Babin, a senior energy trader at CIBC Private Wealth Group. “But with equities rallying on the prospect of a less-than-horrible tariff outcome, crude is along for the ride.” Oil Prices: WTI for May delivery rose 0.7% to settle at $71.71 a barrel in New York. Brent for June settlement climbed 0.6% to settle at $74.95 a barrel. What do you think? We’d love to hear from you, join the conversation on the Rigzone Energy Network. The Rigzone Energy Network is a new social experience created for you and all energy professionals to Speak Up about our industry, share knowledge, connect with peers

Read More »

CenterPoint Energy completes $1.2B sale of Louisiana, Mississippi gas systems

CenterPoint Energy has completed the sale of its natural gas distribution systems in Louisiana and Mississippi to Delta Utilities for $1.2 billion, the companies announced Tuesday. Sale proceeds will “support the efficient funding of what we believe is one of the most tangible long-term growth plans in the industry,” CenterPoint President and CEO Jason Wells said in a statement. “We will continue to optimize the funding of our capital investments to support safety, reliability and resiliency for the benefit of our customers and communities.” CenterPoint now has electric transmission and distribution, power generation and natural gas distribution operations that serve approximately 7 million customers in Indiana, Minnesota, Ohio and Texas. Assets sold to Delta include approximately 12,000 miles of main pipeline serving approximately 380,000 metered customers. “Completing this acquisition furthers our vision to establish modern, multi-state natural gas utilities that build stronger, more resilient communities,” Delta CEO Tim Poché said in a statement. Delta is an affiliate of private equity firm Bernhard Capital Partners. Delta is also in the final stages of acquiring Entergy’s two regulated natural gas local distribution companies in New Orleans and Baton Rouge, with the sale expected to close this summer, the company said. Entergy Louisiana’s gas business serves approximately 95,000 homes and businesses in the Baton Rouge area, and Entergy New Orleans’ gas business serves approximately 109,000 homes and businesses in New Orleans, the utility said in a February announcement of the deal. While private equity firms have been known to squeeze profit from companies before reselling, Poché told the Shreveport-Bossier City Advocate that the gas system acquisition “is not a short-term investment at all.” “We’re very attracted to natural gas, in the resiliency that it has as a very significant transition fuel within our country,” he said. For CenterPoint, the utility said in 2024 that the

Read More »

PacifiCorp aims to add 4.7 GW renewables, 1.7 GW storage by 2031

Dive Brief: PacifiCorp expects to procure through requests for proposals about 2,400 MW of solar, 2,270 MW of wind, 1,680 MW of four-hour storage and 510 MW of 100-hour iron air battery storage by the end of 2030, according to the utility company’s integrated resource plan, filed with state regulators on Monday. By the end of this year, PacifiCorp also plans to finalize a 500-MW agreement to buy electricity from TerraPower’s Natrium advanced nuclear project in Wyoming, which is set to be operating by the end of 2031, the utility company said. The cost of the project is confidential. Also, if, as expected, the Biden administration’s greenhouse gas emissions rule for power plants doesn’t take effect, PacifiCorp plans to continue operating some coal-fired generation it had previously expected to retire. The plan calls for converting 562 MW of coal-fired generation to gas, exiting 386 MW of minority-owned coal-fired capacity, and adding carbon capture equipment to two coal-fired units at the Jim Bridger power plant in Wyoming that total 700 MW. Dive Insight: PacifiCorp utilities serve about 2 million customers in California, Idaho, Oregon, Utah, Washington and Wyoming. The IRP, filed every two years with stakeholder input, provides a roadmap for supplying power to those customers over the next 20 years. The utility owns or contracts for about 5,150 MW of wind, 4,530 MW of coal, 3,850 MW of gas, 3,620 MW of solar coupled with 550 MW of storage, 1,200 MW of hydroelectric capacity and 525 MW of stand-alone storage, according to the IRP. Under its resource plan, PacifiCorp expects that coal- and gas-fired generation will account for 16% and 10% of its energy mix in 2031, down from 35% and 19% this year, respectively. Wind and solar will make up 32% and 25%, up from 24% and 10%, respectively, in

Read More »

PG&E announces microgrid awards for $43M as Sunrun joins its 2025 VPP

Dive Brief: Pacific Gas & Electric intends to award up to $43 million in grants for community microgrid projects in Northern California, the gas and electric utility said on March 26. Issued under California’s Microgrid Incentive Program, a $200 million statewide competitive grant program, the funds will support projects serving nearly 9,000 customers in Humboldt, Lake and Marin counties. Four of the projects are in tribal communities, PG&E said. Also this week, distributed energy provider Sunrun announced a “first-of-its-kind program” with PG&E to harness approximately 600 home solar-and-storage systems to provide “targeted load relief to neighborhoods identified with highly constrained electric grids,” potentially avoiding or deferring distribution grid investments. Dive Insight: In 2023, the California Public Utilities Commission approved $200 million in funding for the Microgrid Incentive Program, including $79.2 million for PG&E, $83.3 million for Southern California Edison and $17.5 million for San Diego Gas & Electric. The nine awards announced March 26 represent the first tranche of PG&E’s MIP grants, selected from a pool of 22 applicants. Applications for the second tranche open on April 3 and run through May 30, PG&E said. First-tranche projects will receive a combined $34 million for front-of-the-meter generating resources and other project costs, plus $1 million each to cover grid interconnection costs, the utility said. Eligible microgrid projects must be able to provide at least 24 consecutive hours of energy in “island mode,” interconnect on distribution lines at or below 50 kV, and have “island mode” emissions no greater than the surrounding grid, according to a PG&E fact sheet. Projects must also be located in areas vulnerable to outages due to high wildfire or earthquake risk or lower historical reliability, and serve “disadvantaged or vulnerable communities” in rural, tribal or low-income areas. The awardees will join several active microgrids in Northern and

Read More »

Energy profits hit half a trillion as consumer price cap rises

Energy profits have hit half a trillion pounds since the energy crisis despite rising consumer bills, according to an independent analysis of company reports. Researchers at the End Fuel Poverty Coalition counted the profits declared by UK energy producers such as Shell and Equinor, as well as suppliers including British Gas and grid operators such as National Grid. Nearly half of the £500 billion of profits generated by the industry since 2020, or £207bn, were made by companies involved in the gas industry. According to Scottish Renewables chief executive Claire Mack, bill payers are ‘paying the price’ for exposure to global gas markets, with renewable power likely to provide the best value for consumers. © Supplied by Scottish RenewablesScottish Renewables chief executive Claire Mack. The ceiling on energy prices, designed to protect consumers from price spikes, and enforced by energy regulator Ofgem, known as the energy price cap, rose by 6.4% this week. Trade union Unite protested against high energy prices on Tuesday, arguing that nobody should have to choose between ‘heating or eating’. Union members will campaign across 40 locations in the UK, with further protests planned in the coming two weeks. Prime minister Keir Starmer has meanwhile promised that clean power will lower energy bills, something Kemi Badenoch, leader of the Conservatives, wants to abolish – while right-wing party Reform is also trying to leverage high bills in its political campaign. Politicians have begun to engage energy companies as they seek to invest in clean technologies and lower prices. Energy secretary Ed Miliband broached talks with Centrica this month to extend the Rough gas storage facility in the North Sea beyond 2030. A spokesperson for the Department for Energy Security and Net Zero (DESNZ) said: “We’re open to discussing proposals, just as long as it provides value for

Read More »

Potential Nvidia chip shortage looms as Chinese customers rush to beat US sales ban

Will it lead to shortages? The US first placed export controls on chips sent to China in October 2022 as a means to slow the country’s technological advances. It blocked the sale of Nvidia’s A100 and H100 chips, leading the company to develop the less powerful A800 and H800 chips for the market; they were also subsequently banned. There was a surge in demand for the H20 following the arrival of Chinese startup DeepSeek’s ultra low-cost, open-source AI model in January. And while the H20 is reported to be 15 times slower than Nvidia’s newest Blackwell chips sold elsewhere in the world, it was designed specifically by Nvidia to comply with the further US export controls introduced in October 2023. It is being used by Chinese companies for training, although it’s billed as an inference chip, explained Matt Kimball, VP and principal analyst for datacenter compute and storage at Moor Insights & Strategy. Should Nvidia choose to focus its efforts on manufacturing more of the chips, Kimball said he doesn’t think it will impact supply in the US and Europe, as Blackwell is the main product sold in those markets and H20 is an N-1 Hopper architecture chip. “If you take this a step further and ask whether this large order slows down the production of chips destined for the US and Europe, I’d say the answer is no, as the Hopper family is built on a different process node than the Blackwell family,” he said. Still, Kimball noted, “supply chain management is difficult, especially for smaller organizations that are put to the back of the line as hyperscalers with multibillion dollar orders are first in line for the newest [chips].”

Read More »

European cloud group invests to create what it dubs “Trump-proof cloud services”

But analysts have questioned whether the Microsoft move truly addresses those European business concerns. Phil Brunkard, executive counselor at Info-Tech Research Group UK, said, commenting on last month’s announcement of the EU Data Boundary for the Microsoft Cloud,  “Microsoft says that customer data will remain stored and processed in the EU and EFTA, but doesn’t guarantee true data sovereignty.” And European companies are now rethinking what data sovereignty means to them. They are moving beyond having it refer to where the data sits to focusing on which vendors control it, and who controls them. Responding to the new Euro cloud plan, another analyst, IDC VP Dave McCarthy, saw the effort as “signaling a growing European push for data control and independence.” “US providers could face tougher competition from EU companies that leverage this tech to offer sovereignty-friendly alternatives. Although €1 million isn’t a game-changer on its own, it’s a clear sign Europe wants to build its own cloud ecosystem—potentially at the expense of US market share,” McCarthy said. “For US providers, this could mean investing in more EU-based data centers or reconfiguring systems to ensure European customers’ data stays within the region. This isn’t just a compliance checkbox. It’s a shift that could hike operational costs and complexity, especially for companies used to running centralized setups.” Adding to the potential bad news for US hyperscalers, McCarthy said that there was little reason to believe that this trend would be limited to Europe. “If Europe pulls this off, other regions might take note and push for similar sovereignty rules. US providers could find themselves adapting to a patchwork of regulations worldwide, forcing a rethink of their global strategies,” McCarthy said. “This isn’t just a European headache, it’s a preview of what could become a broader challenge.”

Read More »

Talent gap complicates cost-conscious cloud planning

The top strategy so far is what one enterprise calls the “Cloud Team.” You assemble all your people with cloud skills, and your own best software architect, and have the team examine current and proposed cloud applications, looking for a high-level approach that meets business goals. In this process, the team tries to avoid implementation specifics, focusing instead on the notion that a hybrid application has an agile cloud side and a governance-and-sovereignty data center side, and what has to be done is push functionality into the right place. The Cloud Team supporters say that an experienced application architect can deal with the cloud in abstract, without detailed knowledge of cloud tools and costs. For example, the architect can assess the value of using an event-driven versus transactional model without fixating on how either could be done. The idea is to first come up with approaches. Then, developers could work with cloud providers to map each approach to an implementation, and assess the costs, benefits, and risks. Ok, I lied about this being the top strategy—sort of, at least. It’s the only strategy that’s making much sense. The enterprises all start their cloud-reassessment journey on a different tack, but they agree it doesn’t work. The knee-jerk approach to cloud costs is to attack the implementation, not the design. What cloud features did you pick? Could you find ones that cost less? Could you perhaps shed all the special features and just host containers or VMs with no web services at all? Enterprises who try this, meaning almost all of them, report that they save less than 15% on cloud costs, a rate of savings that means roughly a five-year payback on the costs of making the application changes…if they can make them at all. Enterprises used to build all of

Read More »

Lightmatter launches photonic chips to eliminate GPU idle time in AI data centers

“Silicon photonics can transform HPC, data centers, and networking by providing greater scalability, better energy efficiency, and seamless integration with existing semiconductor manufacturing and packaging technologies,” Jagadeesan added. “Lightmatter’s recent announcement of the Passage L200 co-packaged optics and M1000 reference platform demonstrates an important step toward addressing the interconnect bandwidth and latency between accelerators in AI data centers.” The market timing appears strategic, as enterprises worldwide face increasing computational demands from AI workloads while simultaneously confronting the physical limitations of traditional semiconductor scaling. Silicon photonics offers a potential path forward as conventional approaches reach their limits. Practical applications For enterprise IT leaders, Lightmatter’s technology could impact several key areas of infrastructure planning. AI development teams could see significantly reduced training times for complex models, enabling faster iteration and deployment of AI solutions. Real-time AI applications could benefit from lower latency between processing units, improving responsiveness for time-sensitive operations. Data centers could potentially achieve higher computational density with fewer networking bottlenecks, allowing more efficient use of physical space and resources. Infrastructure costs might be optimized by more efficient utilization of expensive GPU resources, as processors spend less time waiting for data and more time computing. These benefits would be particularly valuable for financial services, healthcare, research institutions, and technology companies working with large-scale AI deployments. Organizations that rely on real-time analysis of large datasets or require rapid training and deployment of complex AI models stand to gain the most from the technology. “Silicon photonics will be a key technology for interconnects across accelerators, racks, and data center fabrics,” Jagadeesan pointed out. “Chiplets and advanced packaging will coexist and dominate intra-package communication. The key aspect is integration, that is companies who have the potential to combine photonics, chiplets, and packaging in a more efficient way will gain competitive advantage.”

Read More »

Silicon Motion rolls SSD kit to bolster AI workload performance

The kit utilizes the PCIe Dual Ported enterprise-grade SM8366 controller with support for PCIe Gen 5 x4 NVMe 2.0 and OCP 2.5 data center specifications. The 128TB SSD RDK also supports NVMe 2.0 Flexible Data Placement (FDP), a feature that allows advanced data management and improved SSD write efficiency and endurance. “Silicon Motion’s MonTitan SSD RDK offers a comprehensive solution for our customers, enabling them to rapidly develop and deploy enterprise-class SSDs tailored for AI data center and edge server applications.” said Alex Chou, senior vice president of the enterprise storage & display interface solution business at Silicon Motion. Silicon Motion doesn’t make drives, rather it makes reference design kits in different form factors that its customers use to build their own product. Its kits come in E1.S, E3.S, and U.2 form factors. The E1.S and U.2 forms mirror the M.2, which looks like a stick of gum and installs on the motherboard. There are PCI Express enclosures that hold four to six of those drives and plug into one card slot and appear to the system as a single drive.

Read More »

Executive Roundtable: Cooling Imperatives for Managing High-Density AI Workloads

Michael Lahoud, Stream Data Centers: For the past two years, Stream Data Centers has been developing a modular, configurable air and liquid cooling system that can handle the highest densities in both mediums. Based on our collaboration with customers, we see a future that still requires both cooling mediums, but with the flexibility to deploy either type as the IT stack destined for that space demands. With this necessity as a backdrop, we saw a need to develop a scalable mix-and-match front-end thermal solution that gives us the ability to late bind the equipment we need to meet our customers’ changing cooling needs. It’s well understood that liquid far outperforms air in its ability to transport heat, but further to this, with the right IT configuration, cooling fluid temperatures can also be raised, and this affords operators the ability to use economization for a greater number of hours a year. These key properties can help reduce the energy needed for the mechanical part of a data center’s operations substantially.  It should also be noted that as servers are redesigned for liquid cooling and the onboard server fans get removed or reduced in quantity, more of the critical power delivered to the server is being used for compute. This means that liquid cooling also drives an improvement in overall compute productivity despite not being noted in facility PUE metrics.  Counter to air cooling, liquid cooling certainly has some added management challenges related to fluid cleanliness, concurrent maintainability and resiliency/redundancy, but once those are accounted for, the clusters become stable, efficient and more sustainable with improved overall productivity.

Read More »

Microsoft will invest $80B in AI data centers in fiscal 2025

And Microsoft isn’t the only one that is ramping up its investments into AI-enabled data centers. Rival cloud service providers are all investing in either upgrading or opening new data centers to capture a larger chunk of business from developers and users of large language models (LLMs).  In a report published in October 2024, Bloomberg Intelligence estimated that demand for generative AI would push Microsoft, AWS, Google, Oracle, Meta, and Apple would between them devote $200 billion to capex in 2025, up from $110 billion in 2023. Microsoft is one of the biggest spenders, followed closely by Google and AWS, Bloomberg Intelligence said. Its estimate of Microsoft’s capital spending on AI, at $62.4 billion for calendar 2025, is lower than Smith’s claim that the company will invest $80 billion in the fiscal year to June 30, 2025. Both figures, though, are way higher than Microsoft’s 2020 capital expenditure of “just” $17.6 billion. The majority of the increased spending is tied to cloud services and the expansion of AI infrastructure needed to provide compute capacity for OpenAI workloads. Separately, last October Amazon CEO Andy Jassy said his company planned total capex spend of $75 billion in 2024 and even more in 2025, with much of it going to AWS, its cloud computing division.

Read More »

John Deere unveils more autonomous farm machines to address skill labor shortage

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Self-driving tractors might be the path to self-driving cars. John Deere has revealed a new line of autonomous machines and tech across agriculture, construction and commercial landscaping. The Moline, Illinois-based John Deere has been in business for 187 years, yet it’s been a regular as a non-tech company showing off technology at the big tech trade show in Las Vegas and is back at CES 2025 with more autonomous tractors and other vehicles. This is not something we usually cover, but John Deere has a lot of data that is interesting in the big picture of tech. The message from the company is that there aren’t enough skilled farm laborers to do the work that its customers need. It’s been a challenge for most of the last two decades, said Jahmy Hindman, CTO at John Deere, in a briefing. Much of the tech will come this fall and after that. He noted that the average farmer in the U.S. is over 58 and works 12 to 18 hours a day to grow food for us. And he said the American Farm Bureau Federation estimates there are roughly 2.4 million farm jobs that need to be filled annually; and the agricultural work force continues to shrink. (This is my hint to the anti-immigration crowd). John Deere’s autonomous 9RX Tractor. Farmers can oversee it using an app. While each of these industries experiences their own set of challenges, a commonality across all is skilled labor availability. In construction, about 80% percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions, he said. “They have to figure out how to do

Read More »

2025 playbook for enterprise AI success, from agents to evals

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to

Read More »

OpenAI’s red teaming innovations define new essentials for security leaders in the AI era

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle

Read More »