Valar Atomics: $1B Series B at $6B Valuation — Sequoia Bets on Nuclear for AI Data Centers

Valar Atomics raised $1 billion at a $6 billion valuation led by Sequoia Capital, plus a separate $200 million credit facility — $1.2 billion in new financing for a company that didn't exist three years ago. That's the short answer. The longer answer is more interesting.

This isn't another AI application raising at an absurd multiple on ARR that doesn't exist yet. It's a nuclear reactor company, and the round closed nine days after TechCrunch first reported Valar was in talks at the same $6 billion number — meaning Sequoia didn't just price the round, it locked in a valuation the market had already leaked. I've made 65+ angel checks and I read this raise the way I read any infrastructure bet: is the underlying physics de-risked, or is the valuation racing ahead of the product? With Valar, both things are true at once, and that tension is the actual story.

Valar Atomics $1B Series B: Round Terms and Lead Investors

Valar Atomics closed a $1 billion Series B on August 3, 2026, led by Sequoia Capital at a $6 billion post-money valuation, with Sequoia partner Shaun Maguire taking a board seat. The company separately closed a $200 million credit facility led by Erebor Bank alongside J.P. Morgan, with Crescent Cove and Hercules Capital participating, bringing total new capital to $1.2 billion in equity and debt combined.

A Valuation That Tripled Before the Ink Dried

The timeline here is unusually tight for a hard-tech company. On July 17, 2026, TechCrunch reported Valar was in talks to raise new funding at a $6 billion valuation. Seventeen days later, on August 3, the round closed at exactly that number, led by Sequoia. That's a valuation that reportedly tripled from where Valar was priced roughly a year earlier — typical for a company that just proved its core technology works rather than one still promising it will.

Founded in 2023 by CEO Isaiah Taylor — a high-school dropout with no traditional nuclear-industry pedigree — Valar had raised roughly $489-600 million across prior rounds before this close, depending on which database you trust. The company is based in Hawthorne, California (some reports say El Segundo), the same stretch of the South Bay that's home to SpaceX and a growing cluster of hard-tech startups betting Los Angeles, not Silicon Valley, is where atoms-not-bits companies get built.

The Nvidia Milestone That Made This Round Possible

Money followed proof. In July 2026, Valar used a nuclear fission reaction to generate enough electricity to power an Nvidia AI chip and host a live website — a small, almost stunt-like demonstration, but a real one: fission-to-silicon with no intermediate grid connection. Valar and Nvidia are now planning a 30-megawatt nuclear-powered AI facility in Utah, which would be the partnership's first commercial deployment rather than a lab demo.

Valar's reactor design is a high-temperature, gas-cooled system using helium as coolant — a choice that trades some of the regulatory familiarity of light-water reactors for higher operating temperatures and, Valar argues, a design simple enough to manufacture in a factory rather than pour in place over a decade. That's the actual bet: nuclear's cost problem has never really been physics, it's been that every reactor gets built once, as a bespoke civil-engineering project, with all the schedule and cost overruns that implies. Valar is trying to turn a construction project into a manufactured product.

Who Else Is In: Palmer Luckey, Palantir's CTO, and a Roster That Looks Like Defense Tech

The investor list reads less like a typical energy-infrastructure syndicate and more like the defense-tech crossover crowd: Anduril founder Palmer Luckey and Palantir CTO Shyam Sankar are both backers, alongside Point72, Valor Equity Partners, Conviction, and several newer crossover funds including Atreides Management and Snowpoint Ventures. That's a signal in itself — the people who bet early on hard-tech, government-adjacent companies like Anduril and Palantir are now applying the same "build the thing physically, sell it to whoever needs sovereign capacity" playbook to nuclear.

I've written before about how big tech is signing power-purchase deals directly with uranium producers to secure baseload capacity years out. Valar's round is the venture-stage mirror image of those deals: instead of hyperscalers buying existing reactor output, VCs are now funding the next generation of reactor manufacturers directly, betting on the supply side of the same shortage.

Why This Matters Beyond One Company

Nuclear has been the perennial "next big thing" in energy VC for a decade, and most of that money went nowhere fast — permitting timelines alone can outlast a fund's life. What's different in 2026 is a customer that doesn't care about permitting the way a utility does: AI data centers need dedicated, behind-the-meter power now, in quantities the interconnection queue can't deliver on a reasonable timeline. I covered how steep that AI data center power demand curve has gotten — and Valar's $6 billion valuation is a direct bet that whoever solves reactor manufacturing speed captures a slice of that demand before the grid catches up.

It's also notable which firm led. Sequoia has been reshaping its own structure and check sizes under its 2026 fund strategy, and a $1 billion check into a pre-commercial nuclear manufacturer is a firm signal that the largest venture funds now think energy infrastructure belongs in the same portfolio bucket as foundation models and chips — not a side bet, but core AI infrastructure exposure.