Large artificial intelligence company Anthropic is exploring a potential partnership with British semiconductor firm Fractile to ensure a stable supply of chips for custom inferences and reduce significant overhead related to existing semiconductor solutions.
According to reports, these talks represent a strategic effort to improve the speed and efficiency of existing and next-generation models while reducing the San Francisco-based firm's dependence on Nvidia’.
As global demand for productive AI capacity continues to grow, the financial burden of hardware needed to run these systems has become a fundamental hurdle for developers.
Anthropic, which has received billions of dollars of investment from both Amazon and Google’, currently relies heavily on Nvidia’'s H100 units as well as dedicated processors provided by its cloud partners.
But the high market price and limited availability of these industry-standard chips have narrowed profit margins and pushed firms to look elsewhere.
According to industry analysts, an agreement with a specialist firm like Fractile could allow Anthropic’ to gain more control over its technical infrastructure.
This strategy reflects a broader trend among tech giants including Microsoft and Meta, which are increasingly moving away from general-purpose chips in favor of internal or boutique designs.
A shift in memory architecture and support for British technology
Founded in 2022 by Oxford PhD Walter Goodwin, Fractile has attracted a lot of attention for its unusual approach to processor design.
Unlike standard chips, which need to constantly exchange data between the processor and separate memory modules, Fractile’'s “ memory-computing fusion” architecture keeps data directly on the chip using static random access memory or SRAM, which does not need to be refreshed.
According to the British start-up ’, this method can run large language models up to a hundred times faster than current hardware, while also reducing operational costs. %It can be reduced by 90 percent.
While these performance claims are impressive, the technology is still in development.
Fractile has yet to launch a commercial product, and its proprietary chips are not expected to be ready for full-scale data center deployment by 2027.
Despite the long timeline, the start-up’ is reportedly in talks to secure $200 million (euro 170.5 million) in financing, with a valuation exceeding $1 billion (euro 853 million).
The potential partnership highlights the growing importance of ’'s semiconductor sector on the world stage. If an official agreement is reached, Fractile could join Nvidia, Google and Amazon’ to become the fourth-largest supplier of Anthropic’ chips.
According to market reports, the talks are at an early stage and no binding contract has been signed.
But the interest of a big player like Anthropic shows that the ability to deliver faster and cheaper computing power in the AI race is the determining factor.
