With $720M round, Britain’s Fractile gets a leg up in the race for AI chips
First reported by TechCrunch ·
Specialized AI chips that improve inference speed and reduce costs are now available for preorder next year.
British AI chip startup Fractile has secured $720 million in a Series C funding round, co-led by Thrive Capital, Founders Fund, Lightspeed Venture Partners, and Redpoint. The funding also includes investment from Britain's Sovereign AI Fund, with the government aiming to increase the UK's stake in critical AI hardware. The company is developing specialized AI chips designed to reduce the cost of AI model inference and enhance execution speed. Fractile's approach centers on a novel hardware design that stacks DRAM on top of chips, aiming to overcome the traditional trade-off between memory and capacity. This innovation is particularly relevant for agentic AI applications that require longer task chains. Fractile plans to ship its first chips and systems next year.
Fractile's substantial funding, over $720 million and valued at potentially $7 billion, signals significant investor confidence in purpose-built AI hardware as a key sector for future economic growth. The UK government's direct investment through the Sovereign AI Fund underscores a strategic national imperative to secure a domestic position in the AI chip supply chain, moving beyond reliance on internationalFoundries. This round places Fractile among the top-funded startups directly competing to offer alternatives to dominant AI processing solutions.
The company's full-stack approach, integrating a machine learning compiler with its novel DRAM-stacked hardware, is a critical differentiator. This strategy aims to maximize LLM execution speed and efficiency, addressing a growing bottleneck in agentic AI applications that demand faster, more capable inference. By tackling the memory-capacity trade-off, Fractile is positioning itself to enable more complex and lengthy AI task chains without sacrificing performance, a key challenge for current AI deployments.
AI-written summary. May contain errors.