With $720M round, Britain’s Fractile gets a leg up in the race for AI chips

British AI chip startup Fractile has raised a $720 million Series C, with participation from the U.K.'s Sovereign AI Fund. The government framed the backing as a way to ensure the U.K. holds a greater stake in the hardware that will underpin its future economy.

Thrive Capital, Founders Fund, Lightspeed Venture Partners and Redpoint led the round, which was priced as though Fractile were a frontier AI company rather than a component supplier. The British government put the valuation at "over £5 billion" — a level the Financial Times and a Reuters source have reported as roughly $7 billion. That haul makes Fractile one of the best-funded startups chasing purpose-built AI chips, alongside Axelera, Etched and OLIX.

Purpose-built silicon is one of several routes companies are pursuing to cut the cost of inference as models get more demanding in production. Fractile's pitch goes beyond hardware: a full-stack approach pairing its chips with a machine learning compiler designed to maximize the speed at which LLMs execute on them, which is how it sells itself as a frontier AI company.

The core technical claim sits in the memory design. After a pivot, Fractile's approach now revolves around DRAM — the most common type of RAM — stacked on top of each chip in a configuration it hopes will bypass the usual tradeoff between memory and capacity. That tradeoff has grown more problematic with agentic AI, as companies push longer chains of tasks with large models without compromising on speed, while compute supply struggles to keep up. Hopes are high for AI chips generally, despite their limited availability to date.

Chief executive Walter Goodwin wrote on LinkedIn that "there is still a lot of work ahead" and that the new funding will enable the startup to "ship [its] first chips and systems next year." The open questions now are whether the stacked-DRAM design delivers its promised break of the memory-capacity tradeoff in real workloads, how fast Fractile can convert capital into shipped systems in a market where purpose-built AI chips remain scarce, and whether bundling a compiler with silicon is enough to win inference workloads away from entrenched suppliers.