Technology

A Nuclear Startup Just Raised $1 Billion to Power AI Data Centers

Martin HollowayPublished 2d ago4 min readBased on 5 sources
Reading level
A Nuclear Startup Just Raised $1 Billion to Power AI Data Centers

Valar Atomics, a startup that builds small nuclear reactors, has raised $1 billion in a funding round led by Sequoia, one of the most prominent venture capital firms in Silicon Valley. Founder and CEO Isaiah Taylor announced the financing on August 3, 2026. Bloomberg reported that the round valued the company at $6 billion. Shaun Maguire, a partner at Sequoia, joined Valar Atomics' board of directors as part of the deal TechCrunch.

Alongside the funding round, Valar Atomics secured a $200 million line of credit from Erebor and other banks. Additional investors in the round included Apandion Capital, Atreides Management, Conviction, Dream Ventures, HOF Capital, Point72 Ventures, Riot Ventures, Snowpoint Ventures, and Valor Equity Partners TechCrunch.

Valar Atomics builds small modular nuclear reactors. Think of these as smaller versions of traditional nuclear power plants that can be built in a factory and then shipped to where they are needed, rather than being custom-built on site. The company has a reactor in Utah The Information. In June 2026, Valar Atomics showed that its Ward 250 reactor could successfully power an Nvidia Blackwell system, which is a high-performance computer chip used for artificial intelligence. The company also announced a deal with Nvidia to build a 30-megawatt AI factory that does not use water for cooling TechCrunch.

Prior funding rounds included a $130 million early investment and an even earlier round led by Riot Ventures Valar Atomics.

The broader context here is that two big trends are coming together: the massive growth in artificial intelligence computing, and a new wave of smaller nuclear power plants. AI systems require enormous amounts of electricity, and running them also consumes huge volumes of water for cooling in data centers. Valar's waterless AI factory concept directly addresses that water problem. A 30-megawatt capacity is about right for a single large AI computing cluster, which suggests the company wants to place its reactors right next to the computers they power, avoiding the delays and complexity of sending electricity over long distances through the existing power grid.

It is worth flagging that nuclear technology is extremely expensive to develop. A $1 billion funding round is very large for a startup, but in the nuclear industry it is still an early investment. Reactor certification, handling nuclear fuel, and setting up factory production lines each cost billions of dollars over the long run. The $200 million credit line from Erebor and other banks gives the company extra cash to keep operating, but moving from a successful demonstration to actually selling and running reactors commercially involves government approvals that take years, not months.

The group of investors reflects a mix of firms that specialize in deep technology and those focused on growth. Sequoia leading the round, with Maguire taking a board seat, signals that the firm believes nuclear power could become a main energy source for AI computing. Point72 Ventures and Valor Equity Partners bring experience investing in businesses that involve physical assets. Riot Ventures, which backed the company from its earliest days, continued to invest through this round.

The Utah reactor site lets Valar Atomics test its design under real-world conditions. The June 2026 demonstration with Nvidia Blackwell hardware showed that a nuclear reactor can be directly connected to powerful computing equipment, bypassing the grid connection delays that currently slow down new data center construction.

In my view, pairing nuclear reactor development with the surging demand for AI computing power makes sense for both sides. AI companies and cloud providers need steady, clean power that runs around the clock. Nuclear startups need customers willing to commit to buying their power so they can secure financing to build factories. The Nvidia partnership, if it moves forward to actual deployment, could create a model where AI companies build their own power supply alongside their computing facilities, reducing their reliance on the traditional power grid.

The funding from this round gives Valar Atomics the money to keep developing its reactor and start building manufacturing capacity. With Nvidia as a major AI computing partner, and the successful demonstration of the Ward 250 reactor powering Blackwell hardware, the company is now positioned to push toward commercial deployment of its waterless AI factory concept. The $6 billion valuation shows that investors believe demand for nuclear-powered computing infrastructure will last long enough to support the many years it takes to bring nuclear technology to market.