TheChek News Flow

Government-backed venture moves on to commercial footing to tap surging interest in unproved energy source

The UK government plans to attract private investors to the national nuclear fusion developer to which it has pledged £1.3bn in backing, as it seeks to tap surging investor interest in a technology that holds the prospect of eventually delivering abundant clean energy.

Under a new national fusion strategy announced on Monday, UK Industrial Fusion Solutions, the government’s fusion development arm, will be renamed UK Fusion Energy and restructured to operate “more fully as a commercially oriented entity”.

The company “will provide the foundation for the UK to develop first of a kind fusion power plants and directly stimulate the growth of the UK supply chain”, the fusion strategy said. “Working with industrial partners, UKIFS will develop world-leading capabilities as a fusion power plant integrator that can design and deploy fusion power plants globally, with the capability to raise external investment in future.”

The move marks a shift in the UK’s approach to the commercially unproven technology, which has historically relied heavily on public funding.

Private investors have poured billions of dollars into US fusion start-ups in recent years, with China rapidly catching up, betting the technology could one day supply vast amounts of carbon-free electricity.

Tech sector investors have voiced hopes that fusion could help meet the heavy power demand fuelled by the AI boom, which requires huge amounts of electricity for its calculations. Sceptics, however, argue that commercial fusion is decades away.

UKIFS is currently a wholly owned subsidiary of the UK Atomic Energy Authority, the Oxfordshire-based national fusion laboratory, and is responsible for developing the prototype Spherical Tokamak for Energy Production (Step) reactor planned at the site of a former coal-fired power station in Nottinghamshire.

The fusion industry aims to replicate the reaction that powers the Sun by forcing atomic nuclei to combine in a superheated plasma, unlike nuclear fission, which harnesses energy created when atoms are split.

The Step project aims to demonstrate commercial viability by delivering electricity to the grid by 2040, a milestone no fusion reactor has yet achieved anywhere in the world.

According to a government tender notice issued in 2024, the project could cost up to £20bn.

The government did not say when private investors would be brought into the renamed UK Fusion Energy or how much funding the company would seek.

However, progress on the Step programme could help reduce engineering risks and generate technologies that the company could “exploit commercially”, it said.

The UK’s move comes as global competition to develop fusion power intensifies. US fusion start-ups raised $1.6bn in private funding last year, according to PitchBook.

Meanwhile, China has invested between $6.5bn and $10bn in fusion research since 2022, according to estimates by the Special Competitive Studies Project, a US think-tank chaired by former Google chief executive Eric Schmidt.

Nebius said it will provide $12 billion of dedicated capacity across multiple locations

Nebius agreed to a five-year deal worth about $27 billion to supply AI infrastructure capacity to Meta as it continues to see surging demand for its data center computing.

Nebius said it will provide $12 billion of dedicated capacity across multiple locations. Meta also agreed to purchase up to $15 billion in additional available capacity over the five-year period, Nebius said.

Nebius added that it will use large-scale deployments of Nvidia’s next-generation Vera Rubin AI infrastructure platform, which is expected to be available starting in the second half of the year. Nebius will deliver the capacity beginning early next year, it said.

Meta has been investing heavily to expand its AI computing capacity and has signed partnerships with Nvidia and Advanced Micro Devices worth more than $100 billion. It also struck deals earlier this year to make it an anchor customer for new and existing nuclear power in the U.S. needed to fuel its energy-hungry data centers.

Meanwhile, Nebius has seen its position as an AI infrastructure provider surge in recent months as tech companies compete to lock in data center computing capacity to train and run AI models.

In September, it agreed to a $17.4 billion deal with Microsoft to supply AI infrastructure capacity.

Last week, Nvidia agreed to invest $2 billion in Nebius. The two companies will partner to design and deploy large AI data centers and Nebius will gain early access to Nvidia’s latest computing technologies.

Gapping down
In reaction to earnings/guidance:

Other news:

Gapping up
In reaction to earnings/guidance:

Other news:

One of the biggest things he’s expected to announce is a new chip system that combines Nvidia’s technology with that of Groq, the independent chip firm whose tech Nvidia licensed in a roughly $20 billion deal late last year. This is the first time Nvidia has integrated another company’s AI processor directly into its server racks. Nvidia’s current flagship systems rely almost entirely on its own processors and high-speed interconnects to link the components together.

So why has Nvidia broken the mold for Groq? One reason: inference. That’s the technical term for the process of operating AI models that spit out answers to questions asked by ordinary people, as opposed to the process of training AI models. Nvidia’s chips are outstanding for training models, which is why they’ve been so much in demand in recent years. They’re good at the inference stage as well. But Groq’s chips are tailored for inference, which is going to become an increasing part of AI data center workloads.

Nvidia is expected to name OpenAI as a buyer of the new chip, which could power the AI agent that assists with the firm’s coding tasks. It’s worth noting that assuming Nvidia does unveil the Groq-Nvidia system as expected, it will be quite a departure from Huang’s comment in January that what Nvidia would build with Groq would be “quite unique and quite cool, but it won’t affect our core business.”

Groq’s chip—known as a language processing unit (yes, we have GPUs, TPUs and now LPUs!)—is expected to be mass-produced in the second half of this year at Samsung Electronics’ foundry, according to two people with direct knowledge of Nvidia’s plans. It will be the first time Nvidia has manufactured a server chip outside Taiwan Semiconductor Manufacturing Co., the Taiwanese chipmaker that has supplied virtually all of its flagship AI chips.

Diversifying away from a single supplier makes a lot of sense for Nvidia. But the company plans to eventually move production of the LPU back to TSMC, the people said, as the next generation of LPU will integrate better into a coming version of Nvidia’s next AI chip, we hear.

Here are some technical details of the new Nvidia-Groq system for the AI chip nerds out there. The Nvidia-Groq rack will use a different architecture from that of existing Nvidia racks. It will contain 256 Groq chips in a single rack, and Intel processors will help manage communication between them, according to a person involved in the project—a role Nvidia’s own hardware typically performs in its GPU systems. The decision to use Intel components suggests Nvidia’s existing technologies don’t yet integrate cleanly with the LPU.

Still, this is only the beginning of Nvidia’s plans for Groq’s technology. The company is exploring ways to integrate the LPU more deeply into its future road map for chips. One idea under consideration would be to fuse Groq’s processor with a Feynman GPU—Nvidia’s next generation after Rubin—into a single chip. That would improve performance while lowering costs, according to two people involved in the development. Not that Huang will talk about any of that tomorrow!—Anissa Gardizy and Wayne Ma contributed to this item