Fractile, a British startup focused on developing AI chips, has closed a $720 million Series C round backed in part by the UK’s Sovereign AI Fund. The UK government said the investment is intended to help ensure the United Kingdom has “a greater stake in the hardware that will underpin its future economy.”
The financing round was led by Thrive Capital, Founders Fund, Lightspeed Venture Partners and Redpoint. The British government reported the valuation as "over £5 billion," a figure that the Financial Times and a Reuters source have suggested may be roughly $7 billion.
Technical approach and differentiation
Fractile is pursuing a full-stack strategy that goes beyond custom silicon. In addition to chip design, the company is developing a machine-learning compiler intended to maximize the execution speed of large language models (LLMs) on its hardware. This positioning frames Fractile as a frontier AI company rather than a pure hardware vendor.
After a strategic pivot, Fractile’s hardware design centers on using DRAM — the most common type of RAM — stacked directly on top of each chip. The startup aims for this stacked-DRAM approach to mitigate the traditional trade-offs between memory bandwidth and capacity.
That trade-off has become more acute with the rise of agentic AI systems, where organizations run longer chains of tasks with large models while trying to maintain throughput. As compute capacity struggles to keep pace with model demands, purpose-built AI chips are increasingly seen as a hopeful path to lower inference costs, even though such chips have so far been relatively scarce.
Timeline and next steps
In a LinkedIn post, Fractile CEO Walter Goodwin wrote that "there is still a lot of work ahead" and that the new funding will allow the company to "ship [its] first chips and systems next year." The company intends to move toward manufacturing and delivering those systems over the coming year.
Market context
Fractile joins other well-funded startups in the race for specialized AI silicon, alongside companies such as Axelera, Etched and OLIX. The wider goal for these firms is to reduce the cost of inference — making it more affordable to run increasingly demanding AI models in production.
The UK government’s participation signals a strategic interest in strengthening the domestic hardware ecosystem as part of broader economic and technological sovereignty considerations.
