Get ONMINE

Stay Ahead, Stay ONMINE

What’s next for AI in 2025

MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of them here. For the last couple of years we’ve had a go at predicting what’s coming next in AI. A fool’s game given how fast this industry moves. But we’re on a roll, and we’re doing it again. How did we score last time round? Our four hot trends to watch out for in 2024 included what we called customized chatbots—interactive helper apps powered by multimodal large language models (check: we didn’t know it yet, but we were talking about what everyone now calls agents, the hottest thing in AI right now); generative video (check: few technologies have improved so fast in the last 12 months, with OpenAI and Google DeepMind releasing their flagship video generation models, Sora and Veo, within a week of each other this December); and more general-purpose robots that can do a wider range of tasks (check: the payoffs from large language models continue to trickle down to other parts of the tech industry, and robotics is top of the list).  We also said that AI-generated election disinformation would be everywhere, but here—happily—we got it wrong. There were many things to wring our hands over this year, but political deepfakes were thin on the ground.  So what’s coming in 2025? We’re going to ignore the obvious here: You can bet that agents and smaller, more efficient, language models will continue to shape the industry. Instead, here are five alternative picks from our AI team. 1. Generative virtual playgrounds  If 2023 was the year of generative images and 2024 was the year of generative video—what comes next? If you guessed generative virtual worlds (a.k.a. video games), high fives all round. We got a tiny glimpse of this technology in February, when Google DeepMind revealed a generative model called Genie that could take a still image and turn it into a side-scrolling 2D platform game that players could interact with. In December, the firm revealed Genie 2, a model that can spin a starter image into an entire virtual world. Other companies are building similar tech. In October, the AI startups Decart and Etched revealed an unofficial Minecraft hack in which every frame of the game gets generated on the fly as you play. And World Labs, a startup cofounded by Fei-Fei Li—creator of ImageNet, the vast data set of photos that kick-started the deep-learning boom—is building what it calls large world models, or LWMs. One obvious application is video games. There’s a playful tone to these early experiments, and generative 3D simulations could be used to explore design concepts for new games, turning a sketch into a playable environment on the fly. This could lead to entirely new types of games.  But they could also be used to train robots. World Labs wants to develop so-called spatial intelligence—the ability for machines to interpret and interact with the everyday world. But robotics researchers lack good data about real-world scenarios with which to train such technology. Spinning up countless virtual worlds and dropping virtual robots into them to learn by trial and error could help make up for that.    —Will Douglas Heaven 2. Large language models that “reason” The buzz was justified. When OpenAI revealed o1 in September, it introduced a new paradigm in how large language models work. Two months later, the firm pushed that paradigm forward in almost every way with o3—a model that just might reshape this technology for good. Most models, including OpenAI’s flagship GPT-4, spit out the first response they come up with. Sometimes it’s correct; sometimes it’s not. But the firm’s new models are trained to work through their answers step by step, breaking down tricky problems into a series of simpler ones. When one approach isn’t working, they try another. This technique, known as “reasoning” (yes—we know exactly how loaded that term is), can make this technology more accurate, especially for math, physics, and logic problems. It’s also crucial for agents. In December, Google DeepMind revealed an experimental new web-browsing agent called Mariner. In the middle of a preview demo that the company gave to MIT Technology Review, Mariner seemed to get stuck. Megha Goel, a product manager at the company, had asked the agent to find her a recipe for Christmas cookies that looked like the ones in a photo she’d given it. Mariner found a recipe on the web and started adding the ingredients to Goel’s online grocery basket. Then it stalled; it couldn’t figure out what type of flour to pick. Goel watched as Mariner explained its steps in a chat window: “It says, ‘I will use the browser’s Back button to return to the recipe.’” It was a remarkable moment. Instead of hitting a wall, the agent had broken the task down into separate actions and picked one that might resolve the problem. Figuring out you need to click the Back button may sound basic, but for a mindless bot it’s akin to rocket science. And it worked: Mariner went back to the recipe, confirmed the type of flour, and carried on filling Goel’s basket. Google DeepMind is also building an experimental version of Gemini 2.0, its latest large language model, that uses this step-by-step approach to problem solving, called Gemini 2.0 Flash Thinking. But OpenAI and Google are just the tip of the iceberg. Many companies are building large language models that use similar techniques, making them better at a whole range of tasks, from cooking to coding. Expect a lot more buzz about reasoning (we know, we know) this year. —Will Douglas Heaven 3. It’s boom time for AI in science  One of the most exciting uses for AI is speeding up discovery in the natural sciences. Perhaps the greatest vindication of AI’s potential on this front came last October, when the Royal Swedish Academy of Sciences awarded the Nobel Prize for chemistry to Demis Hassabis and John M. Jumper from Google DeepMind for building the AlphaFold tool, which can solve protein folding, and to David Baker for building tools to help design new proteins. Expect this trend to continue next year, and to see more data sets and models that are aimed specifically at scientific discovery. Proteins were the perfect target for AI, because the field had excellent existing data sets that AI models could be trained on.  The hunt is on to find the next big thing. One potential area is materials science. Meta has released massive data sets and models that could help scientists use AI to discover new materials much faster, and in December, Hugging Face, together with the startup Entalpic, launched LeMaterial, an open-source project that aims to simplify and accelerate materials research. Their first project is a data set that unifies, cleans, and standardizes the most prominent material data sets.  AI model makers are also keen to pitch their generative products as research tools for scientists. OpenAI let scientists test its latest o1 model and see how it might support them in research. The results were encouraging.  Having an AI tool that can operate in a similar way to a scientist is one of the fantasies of the tech sector. In a manifesto published in October last year, Anthropic founder Dario Amodei highlighted science, especially biology, as one of the key areas where powerful AI could help. Amodei speculates that in the future, AI could be not only a method of data analysis but a “virtual biologist who performs all the tasks biologists do.” We’re still a long way away from this scenario. But next year, we might see important steps toward it.  —Melissa Heikkilä 4. AI companies get cozier with national security There is a lot of money to be made by AI companies willing to lend their tools to border surveillance, intelligence gathering, and other national security tasks.  The US military has launched a number of initiatives that show it’s eager to adopt AI, from the Replicator program—which, inspired by the war in Ukraine, promises to spend $1 billion on small drones—to the Artificial Intelligence Rapid Capabilities Cell, a unit bringing AI into everything from battlefield decision-making to logistics. European militaries are under pressure to up their tech investment, triggered by concerns that Donald Trump’s administration will cut spending to Ukraine. Rising tensions between Taiwan and China weigh heavily on the minds of military planners, too.  In 2025, these trends will continue to be a boon for defense-tech companies like Palantir, Anduril, and others, which are now capitalizing on classified military data to train AI models.  The defense industry’s deep pockets will tempt mainstream AI companies into the fold too. OpenAI in December announced it is partnering with Anduril on a program to take down drones, completing a year-long pivot away from its policy of not working with the military. It joins the ranks of Microsoft, Amazon, and Google, which have worked with the Pentagon for years.  Other AI competitors, which are spending billions to train and develop new models, will face more pressure in 2025 to think seriously about revenue. It’s possible that they’ll find enough non-defense customers who will pay handsomely for AI agents that can handle complex tasks, or creative industries willing to spend on image and video generators.  But they’ll also be increasingly tempted to throw their hats in the ring for lucrative Pentagon contracts. Expect to see companies wrestle with whether working on defense projects will be seen as a contradiction to their values. OpenAI’s rationale for changing its stance was that “democracies should continue to take the lead in AI development,” the company wrote, reasoning that lending its models to the military would advance that goal. In 2025, we’ll be watching others follow its lead.  —James O’Donnell 5. Nvidia sees legitimate competition For much of the current AI boom, if you were a tech startup looking to try your hand at making an AI model, Jensen Huang was your man. As CEO of Nvidia, the world’s most valuable corporation, Huang helped the company become the undisputed leader of chips used both to train AI models and to ping a model when anyone uses it, called “inferencing.” A number of forces could change that in 2025. For one, behemoth competitors like Amazon, Broadcom, AMD, and others have been investing heavily in new chips, and there are early indications that these could compete closely with Nvidia’s—particularly for inference, where Nvidia’s lead is less solid.  A growing number of startups are also attacking Nvidia from a different angle. Rather than trying to marginally improve on Nvidia’s designs, startups like Groq are making riskier bets on entirely new chip architectures that, with enough time, promise to provide more efficient or effective training. In 2025 these experiments will still be in their early stages, but it’s possible that a standout competitor will change the assumption that top AI models rely exclusively on Nvidia chips. Underpinning this competition, the geopolitical chip war will continue. That war thus far has relied on two strategies. On one hand, the West seeks to limit exports to China of top chips and the technologies to make them. On the other, efforts like the US CHIPS Act aim to boost domestic production of semiconductors. Donald Trump may escalate those export controls and has promised massive tariffs on any goods imported from China. In 2025, such tariffs would put Taiwan—on which the US relies heavily because of the chip manufacturer TSMC—at the center of the trade wars. That’s because Taiwan has said it will help Chinese firms relocate to the island to help them avoid the proposed tariffs. That could draw further criticism from Trump, who has expressed frustration with US spending to defend Taiwan from China.  It’s unclear how these forces will play out, but it will only further incentivize chipmakers to reduce reliance on Taiwan, which is the entire purpose of the CHIPS Act. As spending from the bill begins to circulate, next year could bring the first evidence of whether it’s materially boosting domestic chip production.  —James O’Donnell

MIT Technology Review’s What’s Next series looks across industries, trends, and technologies to give you a first look at the future. You can read the rest of them here.

For the last couple of years we’ve had a go at predicting what’s coming next in AI. A fool’s game given how fast this industry moves. But we’re on a roll, and we’re doing it again.

How did we score last time round? Our four hot trends to watch out for in 2024 included what we called customized chatbots—interactive helper apps powered by multimodal large language models (check: we didn’t know it yet, but we were talking about what everyone now calls agents, the hottest thing in AI right now); generative video (check: few technologies have improved so fast in the last 12 months, with OpenAI and Google DeepMind releasing their flagship video generation models, Sora and Veo, within a week of each other this December); and more general-purpose robots that can do a wider range of tasks (check: the payoffs from large language models continue to trickle down to other parts of the tech industry, and robotics is top of the list). 

We also said that AI-generated election disinformation would be everywhere, but here—happily—we got it wrong. There were many things to wring our hands over this year, but political deepfakes were thin on the ground. 

So what’s coming in 2025? We’re going to ignore the obvious here: You can bet that agents and smaller, more efficient, language models will continue to shape the industry. Instead, here are five alternative picks from our AI team.

1. Generative virtual playgrounds 

If 2023 was the year of generative images and 2024 was the year of generative video—what comes next? If you guessed generative virtual worlds (a.k.a. video games), high fives all round.

We got a tiny glimpse of this technology in February, when Google DeepMind revealed a generative model called Genie that could take a still image and turn it into a side-scrolling 2D platform game that players could interact with. In December, the firm revealed Genie 2, a model that can spin a starter image into an entire virtual world.

Other companies are building similar tech. In October, the AI startups Decart and Etched revealed an unofficial Minecraft hack in which every frame of the game gets generated on the fly as you play. And World Labs, a startup cofounded by Fei-Fei Li—creator of ImageNet, the vast data set of photos that kick-started the deep-learning boom—is building what it calls large world models, or LWMs.

One obvious application is video games. There’s a playful tone to these early experiments, and generative 3D simulations could be used to explore design concepts for new games, turning a sketch into a playable environment on the fly. This could lead to entirely new types of games. 

But they could also be used to train robots. World Labs wants to develop so-called spatial intelligence—the ability for machines to interpret and interact with the everyday world. But robotics researchers lack good data about real-world scenarios with which to train such technology. Spinning up countless virtual worlds and dropping virtual robots into them to learn by trial and error could help make up for that.   

—Will Douglas Heaven

2. Large language models that “reason”

The buzz was justified. When OpenAI revealed o1 in September, it introduced a new paradigm in how large language models work. Two months later, the firm pushed that paradigm forward in almost every way with o3—a model that just might reshape this technology for good.

Most models, including OpenAI’s flagship GPT-4, spit out the first response they come up with. Sometimes it’s correct; sometimes it’s not. But the firm’s new models are trained to work through their answers step by step, breaking down tricky problems into a series of simpler ones. When one approach isn’t working, they try another. This technique, known as “reasoning” (yes—we know exactly how loaded that term is), can make this technology more accurate, especially for math, physics, and logic problems.

It’s also crucial for agents.

In December, Google DeepMind revealed an experimental new web-browsing agent called Mariner. In the middle of a preview demo that the company gave to MIT Technology Review, Mariner seemed to get stuck. Megha Goel, a product manager at the company, had asked the agent to find her a recipe for Christmas cookies that looked like the ones in a photo she’d given it. Mariner found a recipe on the web and started adding the ingredients to Goel’s online grocery basket.

Then it stalled; it couldn’t figure out what type of flour to pick. Goel watched as Mariner explained its steps in a chat window: “It says, ‘I will use the browser’s Back button to return to the recipe.’”

It was a remarkable moment. Instead of hitting a wall, the agent had broken the task down into separate actions and picked one that might resolve the problem. Figuring out you need to click the Back button may sound basic, but for a mindless bot it’s akin to rocket science. And it worked: Mariner went back to the recipe, confirmed the type of flour, and carried on filling Goel’s basket.

Google DeepMind is also building an experimental version of Gemini 2.0, its latest large language model, that uses this step-by-step approach to problem solving, called Gemini 2.0 Flash Thinking.

But OpenAI and Google are just the tip of the iceberg. Many companies are building large language models that use similar techniques, making them better at a whole range of tasks, from cooking to coding. Expect a lot more buzz about reasoning (we know, we know) this year.

—Will Douglas Heaven

3. It’s boom time for AI in science 

One of the most exciting uses for AI is speeding up discovery in the natural sciences. Perhaps the greatest vindication of AI’s potential on this front came last October, when the Royal Swedish Academy of Sciences awarded the Nobel Prize for chemistry to Demis Hassabis and John M. Jumper from Google DeepMind for building the AlphaFold tool, which can solve protein folding, and to David Baker for building tools to help design new proteins.

Expect this trend to continue next year, and to see more data sets and models that are aimed specifically at scientific discovery. Proteins were the perfect target for AI, because the field had excellent existing data sets that AI models could be trained on. 

The hunt is on to find the next big thing. One potential area is materials science. Meta has released massive data sets and models that could help scientists use AI to discover new materials much faster, and in December, Hugging Face, together with the startup Entalpic, launched LeMaterial, an open-source project that aims to simplify and accelerate materials research. Their first project is a data set that unifies, cleans, and standardizes the most prominent material data sets. 

AI model makers are also keen to pitch their generative products as research tools for scientists. OpenAI let scientists test its latest o1 model and see how it might support them in research. The results were encouraging. 

Having an AI tool that can operate in a similar way to a scientist is one of the fantasies of the tech sector. In a manifesto published in October last year, Anthropic founder Dario Amodei highlighted science, especially biology, as one of the key areas where powerful AI could help. Amodei speculates that in the future, AI could be not only a method of data analysis but a “virtual biologist who performs all the tasks biologists do.” We’re still a long way away from this scenario. But next year, we might see important steps toward it. 

—Melissa Heikkilä

4. AI companies get cozier with national security

There is a lot of money to be made by AI companies willing to lend their tools to border surveillance, intelligence gathering, and other national security tasks. 

The US military has launched a number of initiatives that show it’s eager to adopt AI, from the Replicator program—which, inspired by the war in Ukraine, promises to spend $1 billion on small drones—to the Artificial Intelligence Rapid Capabilities Cell, a unit bringing AI into everything from battlefield decision-making to logistics. European militaries are under pressure to up their tech investment, triggered by concerns that Donald Trump’s administration will cut spending to Ukraine. Rising tensions between Taiwan and China weigh heavily on the minds of military planners, too. 

In 2025, these trends will continue to be a boon for defense-tech companies like Palantir, Anduril, and others, which are now capitalizing on classified military data to train AI models. 

The defense industry’s deep pockets will tempt mainstream AI companies into the fold too. OpenAI in December announced it is partnering with Anduril on a program to take down drones, completing a year-long pivot away from its policy of not working with the military. It joins the ranks of Microsoft, Amazon, and Google, which have worked with the Pentagon for years. 

Other AI competitors, which are spending billions to train and develop new models, will face more pressure in 2025 to think seriously about revenue. It’s possible that they’ll find enough non-defense customers who will pay handsomely for AI agents that can handle complex tasks, or creative industries willing to spend on image and video generators. 

But they’ll also be increasingly tempted to throw their hats in the ring for lucrative Pentagon contracts. Expect to see companies wrestle with whether working on defense projects will be seen as a contradiction to their values. OpenAI’s rationale for changing its stance was that “democracies should continue to take the lead in AI development,” the company wrote, reasoning that lending its models to the military would advance that goal. In 2025, we’ll be watching others follow its lead. 

—James O’Donnell

5. Nvidia sees legitimate competition

For much of the current AI boom, if you were a tech startup looking to try your hand at making an AI model, Jensen Huang was your man. As CEO of Nvidia, the world’s most valuable corporation, Huang helped the company become the undisputed leader of chips used both to train AI models and to ping a model when anyone uses it, called “inferencing.”

A number of forces could change that in 2025. For one, behemoth competitors like Amazon, Broadcom, AMD, and others have been investing heavily in new chips, and there are early indications that these could compete closely with Nvidia’s—particularly for inference, where Nvidia’s lead is less solid. 

A growing number of startups are also attacking Nvidia from a different angle. Rather than trying to marginally improve on Nvidia’s designs, startups like Groq are making riskier bets on entirely new chip architectures that, with enough time, promise to provide more efficient or effective training. In 2025 these experiments will still be in their early stages, but it’s possible that a standout competitor will change the assumption that top AI models rely exclusively on Nvidia chips.

Underpinning this competition, the geopolitical chip war will continue. That war thus far has relied on two strategies. On one hand, the West seeks to limit exports to China of top chips and the technologies to make them. On the other, efforts like the US CHIPS Act aim to boost domestic production of semiconductors.

Donald Trump may escalate those export controls and has promised massive tariffs on any goods imported from China. In 2025, such tariffs would put Taiwan—on which the US relies heavily because of the chip manufacturer TSMC—at the center of the trade wars. That’s because Taiwan has said it will help Chinese firms relocate to the island to help them avoid the proposed tariffs. That could draw further criticism from Trump, who has expressed frustration with US spending to defend Taiwan from China. 

It’s unclear how these forces will play out, but it will only further incentivize chipmakers to reduce reliance on Taiwan, which is the entire purpose of the CHIPS Act. As spending from the bill begins to circulate, next year could bring the first evidence of whether it’s materially boosting domestic chip production. 

—James O’Donnell

Shape
Shape
Stay Ahead

Explore More Insights

More Articles

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

Chevron Joins TotalEnergies in New Nigerian Exploration Blocks

Chevron Corp has signed a deal to acquire 40 percent in Petroleum Prospecting License (PPL) 2000 and PPL 2001 offshore Nigeria from TotalEnergies SE. TotalEnergies will retain operatorship with a 40 percent interest. Local player South Atlantic Petroleum Ltd owns 20 percent. “This new joint venture aims at derisking and

Read More »

Four things AWS needs to fix at re:Invent this week

When it comes to new AI analytics services from AWS, CIOs can expect more of the same, said David Linthicum, independent consultant and retired chief cloud strategy officer at Deloitte Consulting. “Realistically, they can expect AWS to keep integrating its existing services; the key test will be whether this shows

Read More »

Enterprises run into roadblocks with AI implementations

CompTIA estimates a 37% weighted average adoption rate of AI across respondents, but despite the widespread AI adoption, AI skills training strategies remain reactive rather than proactive. Only one in three companies currently mandates AI training for staff, though that figure will change as 85% of respondents are either already

Read More »

Phillips 66 Closes Partial Sale of German, Austrian Fuel Retailer

Stonepeak Partners LP and newly formed Energy Equation Partners have completed the acquisition of a 65 percent stake in JET Tankstellen Deutschland GmbH, a fuel retailer in Germany and Austria, from Phillips 66. The Houston, Texas-based downstream company retains 35 percent in the new joint venture company, it said in an online statement Monday. The transaction gave the business an enterprise value of around EUR 2.5 billion ($2.9 billion), according to the parties. Phillips 66 said it had received about EUR 1.5 billion in pre-tax proceeds. Phillips 66 said it has entered into a “multi-year agreement” to supply JET from the MiRO Refinery south of Frankfurt, in which Phillips 66 owns 18.75 percent. According to Phillips 66’s announcement of the deal May 15, JET operates 970 sites including 843 that were JET-branded. Energy Equation Partners managing partner Javed Ahmed said in a separate statement by the buyers on Monday, “This investment reflects EEP’s commitment to investing in established players in the energy sector who have the potential to make a meaningful impact on the energy transition, and we are excited to work alongside the entire JET team, including its dedicated service station operators, to realize this vision”. “We are executing with discipline and a clear strategy to create long-term shareholder value”, said Phillips 66 chair and chief executive Mark Lashier. “This transaction strengthens the balance sheet and optimizes our portfolio by monetizing non-core assets at attractive valuations.” “Since 2022 we have sold over $5 billion in assets and enhanced our position in the U.S. Central Corridor and Gulf Coast”, Lashier added. In the first quarter of 2025 Phillips 66 exceeded a divestiture target of over $3 billion, announced last year to support its shareholder return target and other long-term priorities, with the receipt of $2 billion in proceeds from the

Read More »

EIA Sees Gasoline, Diesel Price Dropping in 2025, 2026

The U.S. Energy Information Administration (EIA) projected that the U.S. regular gasoline and U.S. on-highway diesel fuel retail price will drop both in 2025 and 2026 in its latest short term energy outlook (STEO). According to this STEO, which was released on November 12, the EIA sees the U.S. regular gasoline retail price averaging $3.10 per gallon this year and $2.98 per gallon next year. The U.S. regular gasoline retail price averaged $3.31 per gallon in 2024, the EIA’s November STEO showed. The U.S. on-highway diesel fuel retail price is projected to average $3.66 per gallon in 2025 and $3.50 per gallon in 2026 in the EIA’s November STEO. In 2024, the U.S. on-highway diesel fuel retail price averaged $3.76 per gallon, the EIA’s latest STEO highlighted. A quarterly breakdown included in the EIA’s latest STEO forecast that the U.S. regular gasoline retail price will come in at $3.02 per gallon in the fourth quarter of 2025, $2.88 per gallon in the first quarter of next year, $3.04 per gallon in the second quarter, $3.09 per gallon in the third quarter, and $2.91 per gallon in the fourth quarter. The U.S. regular gasoline retail price averaged $3.10 per gallon in the first quarter of 2025, $3.16 per gallon in the second quarter, and $3.14 in the third quarter, the EIA’s November STEO showed. Another quarterly breakdown included in the EIA’s latest STEO projected that the U.S. on-highway diesel fuel retail price will average $3.69 per gallon in the fourth quarter of 2025, $3.59 per gallon in the first quarter of next year, $3.41 per gallon in the second quarter, $3.47 per gallon in the third quarter, and $3.54 per gallon in the fourth quarter. The U.S. on-highway diesel fuel retail price averaged $3.63 per gallon in the first quarter of

Read More »

China Shale Oil Output Outlook Improving

In a BMI report sent to Rigzone by the Fitch Group recently, analysts at BMI, a Fitch Solutions company, said China’s outlook for shale oil production is improving on the back of accelerated exploration and a series of new discoveries. “State-owned companies are making notable progress in shale oil exploration and production, viewing shale as a key source of future supply as conventional fields mature and face steeper decline rates,” BMI analysts stated in the report. The analysts went on to note that China “has significant potential in shale and tight oil” but said “replicating the U.S. shale boom remains unlikely”. “State-owned companies cite abundant tight oil resources in the Ordos, Junggar, Songliao, Sichuan, Qaidam, Santanghu, Jiyang, North Jiangsu, and Bohai Bay basins,” the analysts said in the report. “These firms are incentivized to pursue high-risk, high-cost projects regardless of global oil prices, given national priorities of energy security and self-sufficiency,” they added. The analysts highlighted that the country’s National Energy Administration reported a 30 percent year on year increase in shale oil output in 2024, “to 6.0 million tons (around 120,000 barrels per day)”. “Although shale currently represents a small share of total production, steady growth is expected,” the BMI analysts stated in the report. “Most projects remain pilots with small capacities relative to conventional developments, and production growth will be constrained by challenging geological conditions,” they added. “Unlike the U.S., exploiting China’s shale oil and gas resources is more challenging due to the complex and deep reservoir geology, lower reserves scattered across valleys, lower well productivity, and higher costs of production,” they continued. “Currently projects produce small volumes between 10,000 barrels per day and 50,000 barrels per day, far lower than conventional fields like Daqing. Although China has made some progress in producing shale oil, the pace of

Read More »

Argentina Bags First LNG Sales Contract

Argentina has secured its first agreement for long-term sales of liquefied natural gas, a key step in its bid to become a global supplier of the fuel as drilling ramps up in its Vaca Muerta shale patch. A consortium of natural gas producers led by Pan American Energy Group, which is 50 percent owned by British supermajor BP Plc, agreed to sell up to 2 million tons a year of LNG to Germany’s state-owned SEFE for eight years. The deal, which still needs to be finalized, is “a key milestone for the future development of the Vaca Muerta gas resources,” Pan American’s Rodolfo Freyre, who heads the consortium called Southern Energy, said in a statement. The LNG would start getting shipped to Europe in late 2027, covering most of the capacity of Southern Energy’s first floating liquefaction unit, which is being provided by Golar LNG Ltd. Golar will post a second unit to Southern Energy about a year later, boosting total annual capacity to about 6 million tons. The agreement is further validation of Argentina’s shale ambitions. While oil and gas output in the Vaca Muerta are both booming, the outlook for gas has been more complicated given the larger infrastructure investments and long-term supply deals required to become an exporter. A second project, led by state-run YPF SA, hasn’t yet been officially green-lighted. If it does go ahead, India has expressed interest in being a buyer. The accord comes as SEFE, or Securing Energy for Europe GmbH – a former Gazprom PJSC unit nationalized by Germany after the Kremlin’s invasion of Ukraine – plans to end its legacy contract with Russia by the start of 2027. That’s when the European Union’s ban on dealings with Russian LNG will comes into force, allowing companies to skip contractual obligations. SEFE has been looking for new

Read More »

Uniper Completes Sale of German Coal Power Plant

Uniper SE said Monday it had consummated the divestment of the Datteln 4 coal-run power plant in North Rhine-Westphalia to ResInvest Group. The plant is among the assets it has agreed to sell to satisfy fair-competition guardrails imposed by the European Commission in approving Uniper’s bailout by the German government in late 2022. Commissioned 2020, the Datteln plant has a net output of 1,052 megawatts (MW). It supplies electricity and district heating to households, as well as traction power to rail operator Deutsche Bahn, according to German power and gas utility Uniper. According to its announcement of the sale agreement September 19, the over 100 employees at the site were to transfer to Czechia’s ResInvest. The parties agreed not to disclose the purchase price, Uniper said then. On November 3 Uniper said that as part of the bailout-related divestment package, it had completed the sale of Uniper Waerme GmbH, a district heating network serving over 14,400 customers in Germany’s Ruhr area, to Steag Iqony Group’s Iqony Fernwaerme GmbH. Waerme has a network of over 750 kilometers (466.03 miles), according to Uniper. Waerme “is an expert in the efficient use of heat that is generated during electricity production in combined heat and power plants”, Uniper said in a press release. “In addition, they use a variety of other environmentally friendly alternatives for heat generation. This includes heat from mine gas, waste heat from industrial processes and heat generated in electric boilers and smaller decentralized CHP plants”. On July 9 Uniper said it had sold its 18.26 percent stake in AS Latvijas Gaze, which is involved in natural gas trading and sales in the Baltics, to co-venturer Energy Investments SIA. Latvijas Gaze sells gas in Estonia, Finland, Latvia and Lithuania. In Latvia’s household sector, it is the biggest gas supplier, Latvijas Gaze says

Read More »

Shell, Equinor Launch UK North Sea JV

Equinor ASA and Shell PLC have completed the combination of their oil and gas operations on the United Kingdom’s side of the North Sea. Launched Monday, Adura, the 50-50 joint venture, “will be the UK North Sea’s largest independent producer”, Norway’s majority state-owned Equinor said in an online statement. Adura includes Equinor’s 29.89 percent stake in the CNOOC Ltd-operated Buzzard field, which started production 2007; an operating interest of 65.11 percent in Mariner, online since 2019; and an 80 percent operating stake in Rosebank, expected to come onstream 2026. Shell will contribute its 27.97 percent ownership in BP PLC-operated Clair, which began production 2005; a 50 percent operating stake in Gannet, started up 1992; a 100 percent stake in Jackdaw, for which Shell is seeking new consent following a court nullification; a 21.23 percent operating stake in Nelson, which started production 1994; a 50 percent operating stake in Penguins, which started production 2003; a 92.52 percent operating stake in Pierce, which started production 1999; a 44.9 percent stake in BP-operated Schiehallion, which started production 1998; a 55.5 operating stake in Shearwater, which started production 2000; and a 100 percent stake in Victory, started up earlier this year. Adura expects to produce over 140,000 barrels of oil equivalent a day in 2026, and also has several exploration licenses, Equinor said. “Equinor will retain ownership of its cross-border assets, Utgard, Barnacle and Statfjord and offshore wind portfolio including Sheringham Shoal, Dudgeon, Hywind Scotland and Dogger Bank”, Equinor said. “It will also retain the hydrogen, carbon capture and storage, power generation, battery storage and gas storage assets. “Shell UK Ltd will retain ownership of its interests and projects that are part of the UK SEGAL system, namely Fife NGL Plant, St Fergus Gas Terminal and the Braefoot Bay facility, and in the Bacton

Read More »

Cooling crisis at CME: A wakeup call for modern infrastructure governance

Organizations should reassess redundancy However, he pointed out, “the deeper concern is that CME had a secondary data center ready to take the load, yet the failover threshold was set too high, and the activation sequence remained manually gated. The decision to wait for the cooling issue to self-correct rather than trigger the backup site immediately revealed a governance model that had not evolved to keep pace with the operational tempo of modern markets.” Thermal failures, he said, “do not unfold on the timelines assumed in traditional disaster recovery playbooks. They escalate within minutes and demand automated responses that do not depend on human certainty about whether a facility will recover in time.” Matt Kimball, VP and principal analyst at Moor Insights & Strategy, said that to some degree what happened in Aurora highlights an issue that may arise on occasion: “the communications gap that can exist between IT executives and data center operators. Think of ‘rack in versus rack out’ mindsets.” Often, he said, the operational elements of that data center environment, such as cooling, power, fire hazards, physical security, and so forth, fall outside the realm of an IT executive focused on delivering IT services to the business. “And even if they don’t fall outside the realm, these elements are certainly not a primary focus,” he noted. “This was certainly true when I was living in the IT world.” Additionally, said Kimball, “this highlights the need for organizations to reassess redundancy and resilience in a new light. Again, in IT, we tend to focus on resilience and redundancy at the app, server, and workload layers. Maybe even cluster level. But as we continue to place more and more of a premium on data, and the terms ‘business critical’ or ‘mission critical’ have real relevance, we have to zoom out

Read More »

Microsoft loses two senior AI infrastructure leaders as data center pressures mount

Microsoft did not immediately respond to a request for comment. Microsoft’s constraints Analysts say the twin departures mark a significant setback for Microsoft at a critical moment in the AI data center race, with pressure mounting from both OpenAI’s model demands and Google’s infrastructure scale. “Losing some of the best professionals working on this challenge could set Microsoft back,” said Neil Shah, partner and co-founder at Counterpoint Research. “Solving the energy wall is not trivial, and there may have been friction or strategic differences that contributed to their decision to move on, especially if they saw an opportunity to make a broader impact and do so more lucratively at a company like Nvidia.” Even so, Microsoft has the depth and ecosystem strength to continue doubling down on AI data centers, said Prabhu Ram, VP for industry research at Cybermedia Research. According to Sanchit Gogia, chief analyst at Greyhound Research, the departures come at a sensitive moment because Microsoft is trying to expand its AI infrastructure faster than physical constraints allow. “The executives who have left were central to GPU cluster design, data center engineering, energy procurement, and the experimental power and cooling approaches Microsoft has been pursuing to support dense AI workloads,” Gogia said. “Their exit coincides with pressures the company has already acknowledged publicly. GPUs are arriving faster than the company can energize the facilities that will house them, and power availability has overtaken chip availability as the real bottleneck.”

Read More »

What is Edge AI? When the cloud isn’t close enough

Many edge devices can periodically send summarized or selected inference output data back to a central system for model retraining or refinement. That feedback loop helps the model improve over time while still keeping most decisions local. And to run efficiently on constrained edge hardware, the AI model is often pre-processed by techniques such as quantization (which reduces precision), pruning (which removes redundant parameters), or knowledge distillation (which trains a smaller model to mimic a larger one). These optimizations reduce the model’s memory, compute, and power demands so it can run more easily on an edge device. What technologies make edge AI possible? The concept of the “edge” always assumes that edge devices are less computationally powerful than data centers and cloud platforms. While that remains true, overall improvements in computational hardware have made today’s edge devices much more capable than those designed just a few years ago. In fact, a whole host of technological developments have come together to make edge AI a reality. Specialized hardware acceleration. Edge devices now ship with dedicated AI-accelerators (NPUs, TPUs, GPU cores) and system-on-chip units tailored for on-device inference. For example, companies like Arm have integrated AI-acceleration libraries into standard frameworks so models can run efficiently on Arm-based CPUs. Connectivity and data architecture. Edge AI often depends on durable, low-latency links (e.g., 5G, WiFi 6, LPWAN) and architectures that move compute closer to data. Merging edge nodes, gateways, and local servers means less reliance on distant clouds. And technologies like Kubernetes can provide a consistent management plane from the data center to remote locations. Deployment, orchestration, and model lifecycle tooling. Edge AI deployments must support model-update delivery, device and fleet monitoring, versioning, rollback and secure inference — especially when orchestrated across hundreds or thousands of locations. VMware, for instance, is offering traffic management

Read More »

Networks, AI, and metaversing

Our first, conservative, view says that AI’s network impact is largely confined to the data center, to connect clusters of GPU servers and the data they use as they crunch large language models. It’s all “horizontal” traffic; one TikTok challenge would generate way more traffic in the wide area. WAN costs won’t rise for you as an enterprise, and if you’re a carrier you won’t be carrying much new, so you don’t have much service revenue upside. If you don’t host AI on premises, you can pretty much dismiss its impact on your network. Contrast that with the radical metaverse view, our third view. Metaverses and AR/VR transform AI missions, and network services, from transaction processing to event processing, because the real world is a bunch of events pushing on you. They also let you visualize the way that process control models (digital twins) relate to the real world, which is critical if the processes you’re modeling involve human workers who rely on their visual sense. Could it be that the reason Meta is willing to spend on AI, is that the most credible application of AI, and the most impactful for networks, is the metaverse concept? In any event, this model of AI, by driving the users’ experiences and activities directly, demands significant edge connectivity, so you could expect it to have a major impact on network requirements. In fact, just dipping your toes into a metaverse could require a major up-front network upgrade. Networks carry traffic. Traffic is messages. More messages, more traffic, more infrastructure, more service revenue…you get the picture. Door number one, to the AI giant future, leads to nothing much in terms of messages. Door number three, metaverses and AR/VR, leads to a message, traffic, and network revolution. I’ll bet that most enterprises would doubt

Read More »

Microsoft’s Fairwater Atlanta and the Rise of the Distributed AI Supercomputer

Microsoft’s second Fairwater data center in Atlanta isn’t just “another big GPU shed.” It represents the other half of a deliberate architectural experiment: proving that two massive AI campuses, separated by roughly 700 miles, can operate as one coherent, distributed supercomputer. The Atlanta installation is the latest expression of Microsoft’s AI-first data center design: purpose-built for training and serving frontier models rather than supporting mixed cloud workloads. It links directly to the original Fairwater campus in Wisconsin, as well as to earlier generations of Azure AI supercomputers, through a dedicated AI WAN backbone that Microsoft describes as the foundation of a “planet-scale AI superfactory.” Inside a Fairwater Site: Preparing for Multi-Site Distribution Efficient multi-site training only works if each individual site behaves as a clean, well-structured unit. Microsoft’s intra-site design is deliberately simplified so that cross-site coordination has a predictable abstraction boundary—essential for treating multiple campuses as one distributed AI system. Each Fairwater installation presents itself as a single, flat, high-regularity cluster: Up to 72 NVIDIA Blackwell GPUs per rack, using GB200 NVL72 rack-scale systems. NVLink provides the ultra-low-latency, high-bandwidth scale-up fabric within the rack, while the Spectrum-X Ethernet stack handles scale-out. Each rack delivers roughly 1.8 TB/s of GPU-to-GPU bandwidth and exposes a multi-terabyte pooled memory space addressable via NVLink—critical for large-model sharding, activation checkpointing, and parallelism strategies. Racks feed into a two-tier Ethernet scale-out network offering 800 Gbps GPU-to-GPU connectivity with very low hop counts, engineered to scale to hundreds of thousands of GPUs without encountering the classic port-count and topology constraints of traditional Clos fabrics. Microsoft confirms that the fabric relies heavily on: SONiC-based switching and a broad commodity Ethernet ecosystem to avoid vendor lock-in and accelerate architectural iteration. Custom network optimizations, such as packet trimming, packet spray, high-frequency telemetry, and advanced congestion-control mechanisms, to prevent collective

Read More »

Land & Expand: Hyperscale, AI Factory, Megascale

Land & Expand is Data Center Frontier’s periodic roundup of notable North American data center development activity, tracking the newest sites, land plays, retrofits, and hyperscale campus expansions shaping the industry’s build cycle. October delivered a steady cadence of announcements, with several megascale projects advancing from concept to commitment. The month was defined by continued momentum in OpenAI and Oracle’s Stargate initiative (now spanning multiple U.S. regions) as well as major new investments from Google, Meta, DataBank, and emerging AI cloud players accelerating high-density reuse strategies. The result is a clearer picture of how the next wave of AI-first infrastructure is taking shape across the country. Google Begins $4B West Memphis Hyperscale Buildout Google formally broke ground on its $4 billion hyperscale campus in West Memphis, Arkansas, marking the company’s first data center in the state and the anchor for a new Mid-South operational hub. The project spans just over 1,000 acres, with initial site preparation and utility coordination already underway. Google and Entergy Arkansas confirmed a 600 MW solar generation partnership, structured to add dedicated renewable supply to the regional grid. As part of the launch, Google announced a $25 million Energy Impact Fund for local community affordability programs and energy-resilience improvements—an unusually early community-benefit commitment for a first-phase hyperscale project. Cooling specifics have not yet been made public. Water sourcing—whether reclaimed, potable, or hybrid seasonal mode—remains under review, as the company finalizes environmental permits. Public filings reference a large-scale onsite water treatment facility, similar to Google’s deployments in The Dalles and Council Bluffs. Local governance documents show that prior to the October announcement, West Memphis approved a 30-year PILOT via Groot LLC (Google’s land assembly entity), with early filings referencing a typical placeholder of ~50 direct jobs. At launch, officials emphasized hundreds of full-time operations roles and thousands

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 »
More Articles