How much did Fractile raise and at what valuation?

London-based chip designer Fractile has raised $720 million in a Series C round, a bet that memory bandwidth, rather than raw computing power, is the biggest obstacle to faster AI. The round values the four-year-old company at $7 billion, according to the Financial Times, which cited people familiar with the deal.

Thrive Capital, Founders Fund, Lightspeed and Redpoint led the round. Other investors include Alphabet’s CapitalG, DST Global, BlackRock, Bain Capital Ventures, T. Rowe Price, Shopify Ventures, Balderton, the UK’s Sovereign AI fund and the British Business Bank.

The valuation marks a steep climb. In May, Fractile raised around $220 million at a valuation of roughly $1 billion. That puts it among the most valuable startups in the UK, alongside companies such as Revolut.

The FT also reports that Anthropic has expressed interest in becoming an early customer for Fractile’s chips, citing people familiar with the discussions.

Why is memory bandwidth a bottleneck for AI chips?

Founder Walter Goodwin’s pitch starts with an imbalance that he says has grown for decades. Over the past 25 years, chip performance has increased roughly a millionfold, while memory bandwidth has grown only about 40 times. That leaves a 25,000-fold gap between how much data a processor can calculate and how fast it can reach the data it needs. Goodwin told the FT that today’s frontier models are “drinking through a straw.”

Fractile’s answer is a three-dimensional stack that combines custom DRAM and logic. The company says the design can deliver substantially higher bandwidth than high-bandwidth memory, or HBM, while still offering enough capacity to run large models with long context windows. Goodwin claims it could make inference, the process of generating responses to user queries, up to 10 times faster.

Has Fractile shipped a commercial AI chip yet?

Those gains have not been shown in a production chip. Fractile has not released a commercial product, and its first chip is planned for next year. For now, the $7 billion valuation rests on the technology and prototypes, not on anything customers are buying.

The design also aims to sidestep a supply problem. Many competing chipmakers depend on HBM to move data quickly between memory and processors. That technology is expensive, and demand from AI companies is squeezing suppliers such as SK Hynix and Micron.

Fractile is entering a market that is getting more crowded. Nvidia acquired Groq last year and launched its own inference chip in March. Startups including Etched, MatX and fellow British company Olix are developing alternatives. Axelera is also among the well-funded players in the race for purpose-built AI chips.

Philip Clark of Thrive Capital sees the inference market becoming “at least as large as the market Nvidia is addressing today.” He points to AI agents as a particular opportunity, since they need to run continuously for hours or days rather than answer a single prompt.

The round also carries weight for the UK’s push to build its own AI infrastructure. Goodwin described the government’s Sovereign AI fund as a driving force behind frontier AI development in London. The UK government said in its own statement that it wants the country to have “a greater stake in the hardware that will underpin its future economy.” It put the valuation at “over £5 billion,” a figure that roughly lines up with the $7 billion reported by the FT.

Fractile is also positioning itself as more than a chip company. It is building a machine learning compiler to help models run faster on its hardware, and it describes itself as a frontier AI company. In a LinkedIn post, Goodwin wrote that “there is still a lot of work ahead” and said the new money will let the startup “ship [its] first chips and systems next year.”

The company has offices in the US and Taiwan and is building for a global market. Its next test is turning prototypes into working chips that deliver the performance it has promised.

Este artigo foi escrito com a assistência de IA.
News Factory APP - notícias agênticas para impulsionar seu SEO e AEO.