Britain’s Fractile Raises $720M to Take On Nvidia’s Chip Crown

Human Verified 0

Share to: , and

A macro view of stacked silicon chip layers glowing blue

British AI chip startup Fractile has raised $720 million in a Series C round that values the company at about $7 billion, one of the largest bets yet on a challenger to Nvidia’s grip on AI hardware. The round, announced on November 11, was led by Thrive Capital, Founders Fund, Lightspeed and Redpoint, and backed by the UK government’s Sovereign AI Fund. Fractile has not yet shipped a production chip.

What the money buys

Fractile designs chips for AI inference, the stage where trained models answer questions, together with the compiler and software that run on them. Its approach stacks DRAM memory directly on top of the logic, so large models can run fast at very long context lengths — the exact load that long-running AI agents create.

The company says the funding will let it ship its first chips and systems in 2027, after a $220 million Series B that valued it at about $1 billion. The jump — a sevenfold valuation in six months, before a single chip has shipped — is the clearest sign yet of how hot the market for inference hardware has become.

The investor list reads like the AI funding boom’s roll call. Alongside the leads, the round includes CapitalG, Alphabet’s fund, plus BlackRock, DST Global, Elad Gil, Balderton and SV Angel, and the UK’s Sovereign AI Fund and British Business Bank. The government framed its backing as giving Britain a greater stake in the hardware that will underpin its economy.

The valuation itself comes with a wrinkle. The UK government put it at over £5 billion, while the Financial Times and Reuters reported about $7 billion. The government’s own release describes the round as worth over £500 million, while its footnotes say a typical Sovereign AI cheque runs to between £1 million and £10 million.

The public stake, in other words, is small; the headline is the signal. Britain is buying a seat at the table of AI hardware, and the presence of Alphabet’s growth fund among the backers means Google now holds a stake in a company building an alternative to Nvidia’s chips.

Engineers in white suits working in a semiconductor clean room

Our Read

Read plainly, this round is priced on a promise, not a product. Fractile has no shipped chip, no disclosed revenue and no customers on record, yet it is now valued at roughly $7 billion — seven times its May valuation — on the strength of an architecture and a team.

That is either the smartest bet in the inference race or the frothiest, and the honest answer is that nobody outside the company can tell yet. What is clear is the direction of travel: inference, not training, is where the money is going, and Groq’s $8 billion valuation shows Fractile is not the only unproven chipmaker commanding a fortune.

Why It Matters

It matters because inference is becoming the bottleneck of the AI economy. Training a model is a one-off cost; running it for millions of users, all day, is the recurring bill, and the companies that build cheaper, faster inference chips will take a cut of every query.

Fractile’s stacked-memory design is a direct attempt at that prize, and a $720 million war chest buys it the time to prove the idea in silicon. If the first chips land in 2027 as promised, the company joins a crowded race alongside Etched and Axelera; if they slip, the valuation will be the first thing to fall.

Background

Fractile was founded in London in 2022 by Walter Goodwin and Yuhang Song, and now works from London and Bristol. Its Series B, announced in May 2026, raised $220 million at a valuation of about $1 billion. The new round lifts the company into the top tier of funded chip startups, alongside Etched and Axelera.

Source: TechCrunch’s report on the round.

Share to: , and

Faheem R. Khattak
Founder & Senior Editor

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

✓ 10+ Years Covering Google Search & AI

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.