
According to people familiar with the matter, Fractile, a UK-based startup developing custom AI chips, is in advanced discussions to raise about $600 million at a pre-money valuation of approximately $6.5 billion. The valuation would be more than six times the company’s valuation in May.
Lightspeed Venture Partners and Redpoint Ventures are expected to jointly lead the funding round, while Thrive Capital and Founders Fund are also expected to participate as co-lead investors. The financing reportedly includes some capital previously committed at a lower valuation.
Just three months ago, Fractile raised $220 million in a round led by Accel, Founders Fund and Factorial Funds, bringing its valuation to around $1 billion. At the time, the company said it was raising capital to bring its first AI chip to market.
Fractile has also reached a preliminary agreement with Anthropic to supply approximately $250 million worth of AI chips, according to people familiar with the matter. The two companies are expected to expand the deal in the future, with the chips currently expected to enter service in 2027.
The latest financing remains ongoing, and the terms could still change, the sources said.
Founded in 2022 by Oxford robotics researcher Walter Goodwin, Fractile is developing computing systems designed to reduce the time required for AI models to generate responses and solve complex problems. The company focuses on AI inference, the process of running trained AI models to produce results.
Beyond chatbots, Fractile believes its technology could support AI applications in areas such as drug discovery and the development of new materials, where faster inference could help accelerate complex computational workloads.
Fractile is one of a growing number of companies developing custom semiconductors specifically for AI inference as the industry seeks alternatives to Nvidia’s dominant position in AI computing. Other emerging competitors include Cerebras Systems and venture-backed startup Etched, which are also developing specialized AI accelerator technologies.