Fractile Secures $220M to Make AI Think a Month's Work in a Day

UK-based chip startup Fractile has secured $220 million in a Series B funding round to fast-track the development and commercialisation of its next-generation AI inference hardware, the company announced on Tuesday.

The financing round was led by Accel, Factorial Funds, and Peter Thiel’s Founders Fund, with participation from Conviction, Gigascale, O1A, Felicis, Buckley Ventures, and 8VC, investing alongside existing backers.

Former Arm and Acorn Computers executive Stan Boland has also previously invested in Fractile, and in January 2025, former Intel CEO Pat Gelsinger disclosed that he too had invested in the startup.

The company intends to use the funds to accelerate the production and delivery of its initial silicon chips and compute systems directly into the hands of enterprise customers. In February 2026, Fractile also announced plans to invest £100 million to bolster its UK operations over the next three years, including the growth of its existing sites in London and Bristol and the creation of a new hardware engineering facility in the latter city.

Fractile was founded in 2022 on the bet that, eventually, the world's most capable AI systems would be limited in their impact by the amount of time they take to produce useful outputs and that the only way to truly unlock this latent value was to radically reinvent the hardware used to run frontier AI models.

According to founder and CEO Walter Goodwin, today's large language models are already producing up to 100 million tokens in pursuit of tackling hard problems. At roughly 40 tokens per second, the rate at which these models tend to run on existing chips, a single output of that length takes a month to complete.

The technical and economic limits on inference speed, driven above all by memory bandwidth constraints on current architectures, are what is holding back progress.

To compress that month into a day, output will need to be generated at approximately 1,200 tokens per second, while handling the complexity and capacity challenges of operating large models at very long contexts. This is exactly the problem Fractile has been building from the ground up to tackle with its specialised chips for inference.

The company is building a new generation of processors where memory and compute are physically interleaved to deliver both low latency and high throughput simultaneously. These chips will serve thousands of tokens per second to thousands of concurrent users, at a power budget and scale the company says no other system can match.

A report from The Information indicated that generative AI company Anthropic has held discussions with Fractile regarding the purchase of the startup's inference chips when the hardware becomes available in 2027.

Beyond speeding up today's AI workloads, Fractile's Goodwin sees the hardware enabling entirely new categories of applications. Compressing a month of work into a day will make far more ambitious AI use cases economically viable. From agentic coding to drug discovery, software engineering, and materials science, Fractile says anywhere humanity will benefit from sheer intellectual work applied to complex problems.

The company has established engineering hubs across the UK, the United States, and Taiwan to support its global expansion.