Fractile Raises $720 Million Before Its First AI Chips Ship. Investors Put Its Valuation at $7 Billion
UK-based AI chip startup Fractile has raised $720 million in a Series C round to develop high-performance AI inference hardware. The funding comes at a reported $7 billion valuation, ahead of the company's planned first commercial chip launch in 2027.
TL;DR
London- and Bristol-based Fractile has raised $720 million in Series C funding.
The round was led by Thrive Capital, Founders Fund, Lightspeed and Redpoint.
Other investors include CapitalG, BlackRock, DST Global, Balderton, SV Angel, the UK's Sovereign AI Fund and the British Business Bank.
Reuters reported that the funding values Fractile at approximately $7 billion.
The company is developing chips and systems designed to make AI inference faster and more cost-efficient.
Fractile plans to ship its first chips and systems in 2027; it has not yet launched a commercial production chip.
A $720 Million Bet on AI Chips That Have Not Reached the Market Yet
The AI hardware race is attracting enormous investment, and the latest example comes from the United Kingdom.
London- and Bristol-based startup Fractile has raised $720 million in a Series C funding round, making it one of the most heavily funded emerging companies building specialised AI hardware.
The round was led by Thrive Capital, Founders Fund, Lightspeed and Redpoint. Other investors include Google's CapitalG, BlackRock, DST Global, Elad Gil, Balderton, SV Angel, the UK's Sovereign AI Fund and the British Business Bank.
Reuters reported that the funding values Fractile at around $7 billion. The UK government separately confirmed that its Sovereign AI Fund and the British Business Bank were backing the company as part of a funding round worth more than £500 million.
The scale of the investment is striking for a company whose first commercial chips and systems are still planned for 2027.
Fractile is betting that the next major opportunity in AI hardware will not only come from training larger models, but from making those models faster and cheaper to run.
Why AI Inference Is Becoming the Next Hardware Battleground
AI computing involves two broad stages.
The first is training: giving a model vast amounts of data so it can learn patterns and develop its capabilities.
The second is inference: running the trained model to generate responses when someone asks a question, requests an image or gives an AI agent a task.
As AI tools become more widely used, inference demand is growing. Companies need to serve more users, process longer prompts and run increasingly complex reasoning and agent-based workloads.
This creates a different hardware challenge. It is not enough to have powerful processors; they also need to access and move large amounts of information quickly.
Memory bandwidth, the speed at which data can move between memory and computing components can become a bottleneck.
Fractile is designing its hardware around this problem.
Fractile's Bet: Rethinking How Memory and Computing Work Together
Founded in 2022 by Walter Goodwin, Fractile is developing specialised AI chips and systems designed to improve the speed and economics of inference.
The company's approach involves rethinking the relationship between computing logic and memory. Its architecture aims to bring memory closer to the processing components, reducing the delays involved in moving information during demanding AI workloads.
This matters particularly for models working with very long contexts and AI agents that generate large sequences of tokens while carrying out multi-step tasks.
In a recent statement, Goodwin said Fractile's work had achieved breakthroughs including extracting up to 20 times higher speeds from DRAM memory. This is a company-reported technical claim, not an independently verified benchmark of a commercially available production chip.
Fractile describes itself as a full-stack AI hardware company, working across silicon, software and systems rather than focusing on chip design alone.
The company says the new funding will help it accelerate development and ship its first chips and systems next year.
Why the UK Government Is Investing
Fractile's funding round is also important for Britain's ambition to build more of its own AI infrastructure.
The UK government has been working to strengthen domestic capabilities in technologies such as advanced semiconductors, AI hardware and computing infrastructure.
Through its Sovereign AI Fund and the British Business Bank, the government is backing Fractile as part of a funding round worth more than £500 million.
The stated objective is to help promising British AI companies scale within the country, retain skilled jobs and reduce dependence on overseas technology suppliers.
For the UK, the investment is not only about backing one startup. It is also about building a stronger domestic position in the hardware and systems that will support future AI development.
A Huge Valuation, But Commercial Execution Still Lies Ahead
The $7 billion valuation reported by Reuters is a major vote of confidence in Fractile's technology and potential market.
But the company still has important milestones ahead. It must turn its chip architecture into working commercial products, demonstrate performance under real workloads, build reliable supply chains and persuade customers to adopt a new hardware platform.
It will also compete in a market where Nvidia has an established software and hardware ecosystem, alongside other companies developing specialised inference processors.
The first commercial launch in 2027 will therefore be an important test of whether Fractile can translate its technical ambitions and investor backing into products that customers can deploy.
The bottom line: Fractile has secured $720 million to challenge the economics of AI inference, with a reported $7 billion valuation before its first commercial chips ship. The investment shows how strongly investors believe in the next phase of AI hardware. The real proof, however, will come when the chips reach customers.
