Fractile raises $720M Series C to develop AI chips and compiler stack
British startup Fractile secures a $720 million round led by Thrive Capital and Founders Fund, with government backing, to build purpose-built AI chips and a machine learning compiler aimed at reducing inference costs.
1 source · cross-referenced
- Fractile, a UK-based AI chip startup, raised a $720 million Series C led by Thrive Capital, Founders Fund, Lightspeed Venture Partners, and Redpoint.
- The round was backed by the UK’s Sovereign AI Fund and values Fractile at over £5 billion ($7 billion).
- Fractile’s approach focuses on DRAM-based hardware design to address memory-capacity tradeoffs in agentic AI workloads.
- The company plans to ship its first chips and systems in 2025, with a full-stack strategy including a machine learning compiler.
Fractile, a British startup developing AI chips, raised a $720 million Series C led by Thrive Capital, Founders Fund, Lightspeed Venture Partners, and Redpoint. The round was backed by the UK’s Sovereign AI Fund, which stated the investment aims to ensure the country has a greater stake in the hardware underpinning its future economy.
The company’s valuation was reported as over £5 billion by the British government and corroborated by the Financial Times and Reuters, translating to approximately $7 billion. Fractile describes itself as a frontier AI company, emphasizing a full-stack approach that pairs custom hardware with a machine learning compiler to maximize the speed of LLM execution on its chips.
Fractile’s hardware design centers on DRAM, the most common type of RAM, but with a stacked configuration intended to bypass the traditional tradeoff between memory and capacity. This tradeoff has become more acute with the rise of agentic AI, where longer task chains with large models risk compromising speed without specialized hardware.
According to Fractile CEO Walter Goodwin, the new funding will enable the startup to ship its first chips and systems in 2025. The company frames its work as addressing the compute bottleneck for deploying increasingly demanding AI models in production environments.
- Nov 11, 2026 · Hugging Face
Hugging Face proposes token pooling-aware training to improve compressed multi-vector retrieval
Trust79 - Nov 11, 2026 · The Verge — AI
Google’s Tabletop AI experiment lets users build custom collaborative apps via waitlist
Trust79 - Nov 11, 2026 · TechCrunch — AI
OpenAI launches Business Hub with profiles and agent tools to surface companies in ChatGPT
Trust78