1 hour ago
19
Sequoia Capital just wrote a billion-dollar check for a nuclear energy startup. That sentence alone tells you how dramatically the energy conversation has shifted in the last two years.
Valar Atomics, a company founded in 2023 by CEO Isaiah Taylor, closed a $1 billion Series B equity round led by Sequoia, pushing its valuation to $6 billion post-money. The round also included a $200 million credit facility from Erebor and other lenders, giving the company a total war chest of $1.2 billion to move from prototype to production.
From demo to factory floor
The company develops high-temperature gas-cooled small modular reactors, or SMRs, that use TRISO fuel and helium coolant. In English: these are compact nuclear reactors that run on extremely durable fuel particles encased in ceramic layers, cooled by an inert gas instead of water. The design is meant to be safer, simpler, and most importantly, standardizable.
Valar’s approach centers on what it calls “gigasites,” clusters of these reactors deployed together to provide massive amounts of power. The target customers are AI data centers, industrial facilities, and hydrogen production plants that need reliable baseload energy without depending on the traditional grid.
In July 2026, the company hit a milestone that caught investors’ attention: its reactor successfully generated enough power to run an Nvidia AI chip.
The Ward 250 test reactor, backed by the US Department of Energy, is expected to demonstrate criticality, meaning it sustains a controlled nuclear chain reaction.
The funding trajectory tells a story
Valar’s fundraising pace has been aggressive, even by Silicon Valley standards. The company raised a $130 million Series A in November 2025. Earlier in 2026, it pulled in another $450 million at a $2 billion valuation. Now, just months later, its valuation has tripled to $6 billion.
Sequoia partner Shaun Maguire will join Valar’s board as part of the deal.
Why AI is dragging nuclear back from the dead
AI data centers are extraordinarily power-hungry. Training large language models and running inference at scale requires consistent, round-the-clock electricity that solar and wind can’t reliably provide on their own.
Projected electricity demand from AI workloads alone is anticipated to surpass 200 TWh by 2030. To put that in perspective, that’s roughly equivalent to the total electricity consumption of a mid-sized European country.
US government policy is also pushing in this direction. Executive orders have been issued to accelerate nuclear energy deployment, and the Department of Energy’s involvement with Valar’s test reactor suggests regulatory tailwinds that didn’t exist five years ago.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·