Stay Ahead, Stay ONMINE

DOE Announces New Team Up ‘to Build Largest DOE AI Supercomputer’

The U.S. Department of Energy (DOE) announced, in a statement posted on its site this week, a new partnership with NVIDIA and Oracle “to build [the] largest DOE AI supercomputer”. “The U.S. Department of Energy, Argonne National Laboratory, NVIDIA and Oracle … announced a landmark public-private partnership to deliver the DOE’s largest AI supercomputer and […]

The U.S. Department of Energy (DOE) announced, in a statement posted on its site this week, a new partnership with NVIDIA and Oracle “to build [the] largest DOE AI supercomputer”.

“The U.S. Department of Energy, Argonne National Laboratory, NVIDIA and Oracle … announced a landmark public-private partnership to deliver the DOE’s largest AI supercomputer and accelerate scientific discovery,” the DOE said in the statement, adding that the partnership will “immediately deliver world-class AI computing resources to DOE researchers while simultaneously building two next-generation AI supercomputing systems at Argonne National Laboratory”.

In its statement, the DOE highlighted that the announcement is “in accordance” with U.S. President Donald Trump’s Executive Orders titled Accelerating Federal Permitting of Data Center Infrastructure and Removing Barriers to American Leadership in Artificial Intelligence.

The DOE noted in its statement that the Solstice system will feature 100,000 NVIDIA Blackwell GPUs and said it will be the largest AI supercomputer in the DOE’s lab complex. Another system, called Equinox, will feature 10,000 NVIDIA Blackwell GPUs, the DOE added.

“Construction at the Argonne Leadership Computing Facility will immediately begin for the Equinox system,” the DOE pointed out in the statement, revealing that it is expected to be delivered in 2026.

“These AI systems will be seamlessly connected with DOE’s vast network of scientific instruments and data assets to address some of the nation’s most pressing challenges in energy, security, and discovery science,” it added.

The DOE went on to state that, as part of the partnership, Oracle will also immediately provide DOE with access to AI computing resources that use a combination of NVIDIA Hopper and Blackwell architectures.

“Scientists from Argonne and across the country will have access to new AI capabilities to drive technological leadership for science and energy applications,” the DOE said.

In the statement, U.S. Energy Secretary Chris Wright said, “winning the AI race requires new and creative partnerships that will bring together the brightest minds and industries American technology and science has to offer”.

“The two Argonne systems and the collaboration between the Department of Energy, NVIDIA, and Oracle represent a new commonsense approach to computing partnerships. These systems will be a powerhouse for scientific and technological innovation,” he added.

NVIDIA Founder and CEO Jensen Huang said in the statement, “AI is the most powerful technology of our time, and science is its greatest frontier”.

“Together with the Department of Energy and Oracle, we’re building an AI factory that will serve as America’s engine for discovery, giving researchers access to the most advanced AI infrastructure to drive progress across fields ranging from healthcare research to materials,” he added.

Clay Magouyrk, CEO of Oracle, said in the statement, “at Oracle, we are proud to partner with the Department of Energy to deliver sovereign, high-performance AI capabilities”.

“Our collaboration at Argonne, tapping into the power of OCI, will provide a critical resource to address the nation’s most complex challenges and accelerate the next wave of scientific breakthroughs,” Magouyrk added.

Paul Kearns, director at the DOE’s Argonne National Laboratory, said in the statement, “the Equinox and Solstice systems are designed to accelerate a broad set of scientific AI workflows, and we are collaborating with Oracle and NVIDIA to prepare thousands of researchers to effectively leverage the systems’ groundbreaking capabilities”.

“This system will seamlessly connect to forefront DOE experimental facilities such as our Advanced Photon Source, allowing scientists to address some of the nation’s most pressing challenges through scientific discovery,” he added.

AMD-Accelerated AI Supercomputers

In a statement posted on its website earlier this week, the DOE announced two new AMD-accelerated AI supercomputers at Oak Ridge National Laboratory (ORNL), “one of which will be built at record speeds thanks to a new public-private partnership model”.

“The supercomputers will help expand America’s leadership in scientific computing, strengthen national security, and drive the next generation of Gold Standard Science and innovation,” the DOE said in that statement.

“With the new public-private partnership model, the Lux AI cluster, powered by AMD Instinct™ MI355X GPUs, AMD EPYC™ CPUs, and AMD Pensando™ advanced networking, will be deployed in early 2026 to expand DOE’s near-term AI capacity and accelerate work on critical national priorities, including fusion, fission, materials discovery, quantum, advanced manufacturing, and grid modernization,” the DOE added.

In that statement, AMD Chair and CEO Lisa Su said, “we are proud and honored to partner with the U.S. Department of Energy and Secretary Wright to accelerate America’s AI compute infrastructure”.

“This partnership exemplifies public-private collaboration at its best. With Discovery and Lux, we are delivering leadership compute systems that combine performance and energy efficiency to advance America’s research priorities and strengthen U.S. leadership in AI, energy, and national security,” Su added.

ORNL Laboratory Director Stephen Streiffer said in that statement, “Oak Ridge’s leadership in supercomputing has transformed how researchers solve problems”.

“With Discovery and Lux, we’re accelerating the pace of Gold Standard Science at a scale that secures America’s leadership in an increasingly competitive world,” he added.

Petrobras, Exxon, Eni Supercomputers

In a statement posted on its site earlier this month, Petrobras announced that it had put the Harpia supercomputer into operation.

“The machine, equivalent to 10 million cell phones or 200,000 laptops, will help the company maintain its leadership in processing capacity in Latin America,” Petrobras said in in the statement.

“Valued at R$435 million [$81.4 million], it is the largest of the five new supercomputers purchased by the company, which together cost R$500 million [$93.6 million] and will increase Petrobras’ computational capacity by more than 60 percent,” the company added.

Petrobras noted in the statement that the acquisition of new supercomputers is strategic in keeping the company at the technological forefront of the oil and gas sector in subsurface seismic imaging.

“Obtaining more detailed seismic images of the subsurface allows us to refine reservoir behavior simulations, enabling more efficient production. In addition, large computational capacities allow Petrobras to compete globally, attract partnerships, and seize business opportunities,” Petrobras Corporate Affairs Director Clarice Coppetti said in the statement.

In a statement posted on its website back in March, ExxonMobil announced the introduction of Discovery 6, “ExxonMobil’s latest HPC system”.

“The Discovery 6 supercomputer combined with our elastic Full Wavefield Inversion (eFWI) technology will allow us to increase resource recovery with less capital,” the company said in that statement.

“The eFWI technology rapidly creates images that reveal information about rock and fluid properties and insights into where hydrocarbons are present. HPC will enable 4D eFWI, adding time-lapse to further optimize development and production,” it added.

In a statement posted on its website in November last year, Eni announced the completion and launch of a new supercomputing system dubbed HPC6.

“HPC6 provides a significant increase in computational power to a peak of 606 PFlop/s, or over 600 quadrillion mathematical operations per second,” Eni noted in that statement.

“The system has achieved a debut ranking of No.5. in the new TOP500’s list (released on November 18, 2024) of super computers,” it added.

“This is an excellent result that ranks HPC6 as first supercomputer in Europe, the world’s first industrial-use supercomputer and the only non-U.S. system among the top 5 in the world,” it continued.

Eni said in that statement that the launch of HPC6 marks a pivotal moment in Eni’s decarbonization strategy, “where technology and innovation differentiate Eni’s approach and deliver value creation through the development of new businesses related to the energy transition”.

To contact the author, email [email protected]

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 won’t sell VMware to new cloud customers

“Things are pretty straightforward for the managed cloud option — there are two obvious choices. Companies can go with IBM and Red Hat, or choose VMware,” he said, noting that while the company has been a big VMware reseller, “IBM had to ask themselves, ‘How are we going to compete

Read More »

USA Energy Sec Says USA Is Ready to Sell More Oil, Gas to China

Energy Secretary Chris Wright said the US is prepared to sell more oil and natural gas to China if Beijing cuts back on purchases from Russia.  “There’s so much space for mutually beneficial deals between the US and China,” Wright said Thursday during a Bloomberg Television interview, noting that the US is the world’s largest oil and gas exporter, while China is the biggest importer.  The energy secretary plans to travel to Asia within weeks, or possibly sooner, following President Donald Trump trip to the continent this week.  During his trip, Trump said he reached deals with Chinese President Xi Jinping and South Korea President Lee Jae Myung to buy more US oil and gas. Trump also cited a “very large scale” transaction involving Alaskan oil and gas in a post on the social media site Truth Social but didn’t provide more details. “There is lots of room from the United States to grow our role in supplying natural gas, oil, and frankly nuclear technology to South Korea,” Wright said in the interview.  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.

Read More »

DTE inks first data center deal to grow electric load 25%

8.4 GW data center pipeline DTE Energy has signed agreements to serve a 1.4 GW hyperscaler and has line of sight to another 7 GW of potential large loads, officials said. $30 billion investment pipeline DTE plans to invest $30 billion in generation, distribution and other infrastructure across the 2026-2030 timeframe. 12 GW New generation DTE expects to add from 2026 to 2032, including batteries, renewables and gas. DTE Energy has signed a 1.4 GW agreement to serve a hyperscale data center and sees “transformational growth” ahead in a project pipeline that could represent up to an additional 7 GW of load. It is the utility’s first hyperscaler agreement at a time when data centers are rapidly expanding around the United States.  Large loads, including AI data centers, could ultimately add 20% to U.S. utilities’ peak demand, most within the next decade, Wood Mackenzie said in a September report. DTE serves about 2.3 million customers in southeast Michigan, including Detroit. “This is an exciting milestone,” DTE President and CEO Joi Harris said in a Thursday call with analysts. “Aside from the 1.4 GW of new load, we are still in late-stage negotiations with an additional 3 GW of data center load providing potential further upside to our capital plan as we advance these negotiations. … And we have a pipeline of an additional 3-4 GW behind that.” The data center contract of 1.4 GW increases DTE’s electric load by 25%, officials said. “We also expect longer-term growth opportunities through the expansion of these initial hyperscaler projects,” Harris said. The generation investment needed to support the additional load “could very well come into the back end of our five-year plan, providing incremental capital investment.” The utility has added about $6 billion to its five-year plan and now expects to invest $30 billion in

Read More »

AEP capital spending plan surges 33% to $72B in utility ‘super cycle’

$72 billion capital expenditure plan Up 33% from AEP’s previous five-year capex plan, partly driven by 765-kV transmission projects in Texas and the PJM Interconnection region. 65 GW peak load in 2030 Up 76% from AEP’s summer peak, driven by 28 GW in data center and other large load agreements. $2.6 billion year-to-date operating earnings Up 13% from a year ago, partly driven by 765-kV transmission projects in Texas and the PJM Interconnection region. 7% to 9% earnings per share annual growth rate Up from 6% to 8% previously. AEP’s stock price jumped 6% Wednesday to $122.11/share. 3.5% Annual residential rate hikes AEP expects its customers will face over the next five years. Surging loads In the last 12 months, AEP’s utilities sold 6% more electricity compared to the previous year, with residential sales up 2.3% and commercial sales up 7.9%, and those sales are expected to continue growing, according to the company. About 2 GW of data centers came online in the third quarter, Trevor Mihalik, AEP vice president and CFO, said Wednesday during an earnings conference call. AEP expects its peak load will hit 65 GW by 2030, up from 37 GW, with demand surging in Indiana, Ohio, Oklahoma and Texas, according to William Fehrman, AEP chairman, president and CEO. The growth estimate includes 28 GW of customers with electric service agreements or letters of agreement, he said. About half of that 28 GW is in the Electric Reliability Council of Texas market, 40% in the PJM Interconnection and 10% in the Southwest Power Pool, according to Fehrman. About 80% of that pending demand is from hyperscalers such as Google, AWS and Meta, Mihalik said. The remaining demand growth is from industrial customers with projects such as a Nucor steel mill in West Virginia and Cheniere Energy’s liquefied

Read More »

A hydrogen ‘do-over’ for California

Melanie Davidson is a hydrogen policy and markets expert. Most recently she led clean fuels strategy at San Diego Gas & Electric. She is former board member of the California Hydrogen Business Council and was a founding staff member of the Green Hydrogen Coalition. Recently, over $2 billion of federal funding for the U.S. Department of Energy California and Pacific Northwest Hydrogen Hubs was terminated. These Hubs were premised on the use of “renewable, electrolytic hydrogen” — meaning hydrogen generated by using renewable electricity to power water-splitting electrolyzers. The resultant hydrogen would have replaced fossil fuels for heavy duty transportation, port operations, and power generation.  The idea of a fully renewable, water-based, hydrogen economy for the West was an exciting one — both in its altruism and the premise, backed by the DOE’s 2021 “Hydrogen Shot.” The idea was for cheap, abundant solar and rapidly declining electrolyzer cost curves to generate hydrogen from water with zero emissions — for $1/kg by 2030, no less.  However, cuts in Hub funding, together with a 2027 sunset date for projects to qualify for the hydrogen production tax credit, are just two more blows to the many pre-existing economic challenges facing a renewable hydrogen future. At least in California, those challenges include rising (not falling) capital costs for electrolyzers and electrical equipment, high interest rates, a scarcity of water rights, and high costs of grid electricity, qualifying renewable energy credits, and land.  The renewable Hubs were anchored on the idea that by leveraging otherwise curtailed solar (terrawatt-hours worth annually), we could generate cheap, abundant, seasonally stored renewable electrolytic hydrogen at distributed locations, then convert the hydrogen back to the grid via fuel cells as needed. It’s an elegant idea, but it doesn’t pencil. The capital costs of those electrolyzers, compressors, liquefiers, hydrogen storage vessels and fuel

Read More »

USA Crude Oil Stocks Drop Almost 7MM Barrels WoW

In its latest weekly petroleum status report, the U.S. Energy Information Administration (EIA) highlighted that U.S. commercial crude oil inventories, excluding those in the Strategic Petroleum Reserve (SPR), decreased by 6.9 million barrels from the week ending October 17 to the week ending October 24. This EIA report, which was released on October 29 and included data for the week ending October 24, showed that crude oil stocks, not including the SPR, stood at 416.0 million barrels on October 24, 422.8 million barrels on October 17, and 425.5 million barrels on October 25, 2024. The report highlighted that data may not add up to totals due to independent rounding. Crude oil in the SPR stood at 409.1 million barrels on October 24, 408.6 million barrels on October 17, and 385.8 million barrels on October 25, 2024, the report highlighted. Total petroleum stocks – including crude oil, total motor gasoline, fuel ethanol, kerosene type jet fuel, distillate fuel oil, residual fuel oil, propane/propylene, and other oils – stood at 1.677 billion barrels on October 24, the report revealed. Total petroleum stocks were down 15.4 million barrels week on week and up 43.6 million barrels year on year, the report showed. “At 416.0 million barrels, U.S. crude oil inventories are about six percent below the five year average for this time of year,” the EIA said in its latest weekly petroleum status report. “Total motor gasoline inventories decreased by 5.9 million barrels from last week and are about three percent below the five year average for this time of year. Both finished gasoline and blending components inventories decreased last week,” it added. “Distillate fuel inventories decreased by 3.4 million barrels last week and are about eight percent below the five year average for this time of year. Propane/propylene inventories increased by 2.5

Read More »

Shell Beats Profit Estimates

Shell Plc beat profit estimates and maintained share buybacks while paying down debt, showing its resilience to weaker oil prices. The third-quarter performance was helped by stronger oil and gas trading, a vital part of the business that struggled earlier in the year amid geopolitical volatility. An expansion in Shell’s liquefied natural gas business after the startup of a new project in Canada also contributed to the positive outlook.  The strength of Shell’s balance sheet — with net-debt falling from the prior quarter — has positioned the company to maintain consistent returns to investors even as oil prices have fallen. Following years of wide fluctuations, the company’s results have reached the point of “boring” consistency for some analysts. Chief Executive Officer Wael Sawan has been on a two-year push to cut costs, improve reliability and offload under-performing assets in an effort to close a valuation gap with the company’s US rivals. Shell’s shares have risen 16% in London since the start of 2025, outperforming its closest peers.  “Shell delivered another strong set of results, with clear progress across our portfolio and excellent performance in our marketing business and deepwater assets in the Gulf of America and Brazil,” Sawan said in a statement on Thursday.  The company maintained its pace of share buybacks at $3.5 billion a quarter. Adjusted third-quarter net income dropped about 10% from a year earlier to $5.43 billion, but was well above the average analyst estimate of $4.74 billion. Net debt declined to $41.2 billion from $43.2 billion at the end of June. Shares of the company were little changed.  After years of outsized profits as demand roared back following the pandemic, the world’s largest energy producers are facing leaner times with crude prices having dropped about 14% this year. Oil market fundamentals point to an oversupply

Read More »

AWS opens giant data center for AI training

Just over a year after construction began, Amazon Web Services (AWS) has opened its giant data center near Lake Michigan in the US state of Indiana. The data center, which is part of AWS Project Rainier, covers 1,200 acres, or 4.86 square kilometers. This makes it one of the largest data centers in the world, CNBC reports. The construction cost amounted to 11 billion dollars, which is currently equivalent to 103 billion Swedish kronor.

Read More »

Samsung’s memory ramp-up may ease AI and cloud upgrade concerns

The company confirmed that its latest-generation HBM3E chips are now being shipped to “all related customers,” a possible sign that supply to major AI chipmakers like Nvidia may be stabilizing. With mass production of HBM4 expected next year, Samsung could eventually help relieve pressure on the broader enterprise infrastructure ecosystem, from cloud providers building new AI clusters to data center operators seeking to expand switching and storage capacity. Samsung’s Foundry division also plans to begin operating its new 2nm fab in Taylor, Texas, in 2026 and supply HBM4 base-dies, a move that could further stabilize component availability for US cloud and networking infrastructure providers. Easing the memory chokehold Easing DRAM and NAND lead times will unlock delayed infrastructure projects, particularly among hyperscalers, according to Manish Rawat, semiconductor analyst at TechInsights. “As component availability improves from months to weeks, deferred server and storage upgrades can transition to active scheduling,” Rawat said. “Hyperscalers are expected to lead these restarts, followed by large enterprises once pricing and delivery stabilize. Improved access to high-density memory will also drive faster refresh cycles and higher-performance rack designs, favoring denser server configurations. Procurement models may shift from long-term, buffer-heavy strategies to more agile, just-in-time or spot-buy approaches.” Samsung’s expanded role as a “meaningful volume supplier” of HBM3E 12-high DRAM will also be crucial for hyperscalers planning their 2026 AI infrastructure rollouts, according to Danish Faruqui, CEO of Fab Economics. “Without Samsung’s contribution, most hyperscaler ASIC programs, including Google’s TPU v7, AWS’s Trainium 3, and Microsoft’s in-house accelerators, were facing one- to two-quarter delays due to the limited HBM3E 12-high supply from SK Hynix,” Faruqui said. “These products form the backbone of next-generation AI data centers, and volume ramp-up depends directly on Samsung’s ability to deliver.”

Read More »

Oracle’s cloud strategy an increasingly risky bet

However, he pointed out, “theatre is not delivery. What Oracle served was less a coronation than a carefully staged performance: a heady cocktail of ambition, backlog, and speculation. At Greyhound Research, we argue that such moments call not for applause but for scrutiny. The right instinct is not to toast, but to check the bill.” Oracle ‘betting the farm’ on AI Rob Tiffany, research director in IDC’s worldwide infrastructure research organization, had a different view, saying, “in an effort to catch up with the other hyperscaler clouds, Oracle has been aggressively building out its Oracle Cloud Infrastructure (OCI) data center regions all over the world prior to their Stargate endeavor with Crusoe, OpenAI, and SoftBank, to capitalize on the AI opportunity.” Speculation about the burst of the AI bubble aside, he said, “the strength and success of the OCI buildout thus far rests with Oracle’s dominant database and Fusion Cloud ERP, and those enterprise customers should be confident  in Oracle’s future.” Scott Bickley, advisory fellow at Info-Tech Research Group, added, “[while it is] extraordinary to see them take on this kind of debt, [Oracle] are really betting the farm on the AI revolution panning out. There are a lot of risks involved if momentum in the AI space loses its current trajectory. There could be a lot of stranded infrastructure and capital.” The ultimate risk, he said “lies in the viability of OpenAI. These guys have said they’re going to spend $1.4 trillion on AI capacity build out, and they’re sitting on a revenue base of $13 billion a year right now. If they go up in smoke, then that could leave a lot of this investment stranded. That would be the worst case kind of Black Swan scenario.” At this point, he said, “CIOs would not want that bubble

Read More »

Google wants to restart closed nuclear power plant in Iowa

The enormous amount of energy required to power a modern data center has prompted major tech companies to sign major partnership agreements with power companies. Most recently, Google signed an agreement with Next Era Energy to restart the Duane Arnold Energy Center in Iowa. The nuclear power plant in question was shut down in 2020 and it is expected to take four years to make it operational again, CNBC reports.

Read More »

Arista fills out AI networking portfolio

The 7280R4-32PE features 25.6 Tbps switching capacity and supports 32x 800 GbE ports with Octal Small Form-Factor Pluggable (OSFP) or Quad Small Form-Factor Pluggable – Double Density (QSFP-DD) optical uplinks. It’s targeted at customers that need to support AI/ML workloads and routing-intensive edge use cases, Arista stated. It supports 25% lower power per Gbps compared to the prior generation, according to Arista.  A second version, the 7280R4-64QC-10PE, is aimed at dense, deep buffer-requiring workloads in data centers with 100G/800G requirements. The box supports 64x 100 GbE and 10x 800 GbE OSFP in addition to 4x 1/10/25 GbE for management or additional low-speed interfaces, Arista started. The box promises 20% lower power requirement per Gbps over the prior generation of the box, Arista stated.  At the high end, the new 7800R4 is the vendor’s latest flagship networking box capable of supporting 36 ports of 800GbE OSFP and QSFP-DD line cards in 4, 8, 12, and 16-slot chassis configurations. The box offers a high radix capacity – meaning it can be fully loaded with line card and support 576 physical 800 Gigabit Ethernet ports or 1,152 400GbE ports, Arista stated.  In addition, the 7800R supports a new 3.2 TbpsEthernet line card called HyperPort that supports 4 800G channels to tie together widely dispersed data centers via a technique Arista calls “scale across.” It’s designed to scale across buildings in the same metropolitan region or across sites in different cities or countries. This routed Data Center Interconnect technology that can extend AI clusters over Metro or long-haul WAN links, according to Arista. “Building on the flexible Extensible Operating System (EOS) software foundation [which runs across all Arista networking gear] and deep buffering, HyperPort delivers up to 44% faster job completion time (JCT) for high-bandwidth AI flows via a single high-speed port, compared to

Read More »

Cisco, Nvidia strengthen AI ties with new data center switch, reference architectures

The new box extends Cisco Nexus 9000 Series portfolio of high-density 800G aggregation switches for the data center fabric, Cisco stated. The Nexus 9000 data center switches are a core component of the vendor’s enterprise AI offerings. They support congestion-management and flow-control algorithms and deliver the right latency and telemetry to meet the design requirements of AI/ML fabrics, Cisco stated. With the Cisco N9100 Series, Cisco now supports Nvidia Cloud Partner (NCP)-compliant reference architecture. “This development is particularly significant for neocloud and sovereign cloud customers building data centers with capacities ranging from thousands to potentially hundreds of thousands of GPUs, as it allows them to diversify their supply chains effectively,” wrote Will Eatherton, senior vice president of Cisco networking engineering, in a blog post about the news. An add-on license lets customers extend the NCP reference architecture to define how customers can mix and mingle Nvidia Spectrum-X adaptive routing capability with Cisco Nexus 9300 Series switches and Nvidia Spectrum-X Ethernet SuperNICs. “The combination of low latency and congestion-aware, per-packet load balancing on Cisco 9300 switches, along with out-of-order packet handling and end-to-end congestion management on Nvidia SuperNICs, significantly enhances network performance. These improvements are essential for AI networks, optimizing critical metrics such as job completion time,” Eatherton wrote. In addition to neoclouds and sovereign buildouts, enterprise customers are a target, according to Futuriom’s Raynovich.

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 »