Anthropic explores UK startup's AI inference chips promising 100x speed boost at fraction of cost

Anthropic in early talks to buy DRAM-less AI inference chips from UK startup -- Fractile's SRAM architecture reduces need for pricey memory during extreme pricing and shortage crunch

The Claude developer is exploring a fourth chip supplier alongside Nvidia, Google, and Amazon. Anthropic has reportedly held early discussions with London-based chip startup Fractile about purchasing the company's inference accelerators, The Information reported on Friday, citing people familiar with the matter. The talks would add Fractile as a fourth source of AI server silicon for the Claude developer, which already uses chips from Nvidia, Google, and Amazon. Fractile's chips aren't expected to reach commercial readiness until around 2027, placing any deployment well outside Anthropic's near-term procurement plans and roughly inside the same window as its Google-Broadcom TPU partnership. Founded in 2022 by Oxford PhD Walter Goodwin, Fractile is developing an inference chip that co-locates memory and compute on the same die using SRAM rather than shuttling data to separate DRAM chips. That data movement between the GPU and off-chip DRAM is one of the main bottlenecks in running large AI models at speed. Goodwin told Fortune in July 2024 that Fractile's design stores data needed for computations directly next to the transistors that perform the arithmetic, rather than relying on off-chip DRAM. Based on simulations at the time, Goodwin said Fractile could run a large language model 100 times faster and 10 times cheaper than Nvidia's GPUs, though the company had not yet manufactured test chips. The company raised $15 million in seed funding, co-led by Kindred Capital, the NATO Innovation Fund, and Oxford Science Enterprises. Fractile is now in talks to raise $200 million at a $1 billion-plus valuation, with Founders Fund, 8VC, and Accel among the potential investors. The Fractile team reportedly includes engineers from Graphcore, Nvidia, and Imagination Technologies, and the company is building its own software stack alongside the hardware. Anthropic has deliberately avoided dependence on any single chip vendor, running Claude on Nvidia GPUs, Amazon's Trainium processors through Project Rainier, and Google's TPUs under a deal announced in October that provided over 1GW of compute capacity. In early April, that expanded to 3.5GW of TPU capacity from 2027 through 2031. The interest in Fractile coincides with surging demand on Anthropic's existing infrastructure. The company's annualized revenue run rate passed $30 billion in March, up from around $9 billion at the end of 2025, and its inference costs have been a drag on gross margins. Unlike OpenAI and xAI, which are building or expanding their own massive data center footprints, Anthropic has opted to rent capacity from multiple providers and negotiate leverage through diversified chip supply. Fractile is one of several inference-focused startups pursuing SRAM-based or near-memory architectures, including Groq and Cerebras. Nvidia struck a $20 billion acquisition deal with Groq in December and subsequently launched its own dedicated inference accelerator, Groq 3 LPX, acknowledging the growing commercial pressure to optimize cost-per-token at scale. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Anthropic Eyes UK Startup's Fusion Tech Promising 100x Faster AI Inference at One-Tenth the Cost of NVIDIA's Groq

Anthropic, the creators of Claude AI, are reportedly in early talks with a UK startup whose SRAM tech can boost AI inference by 100x & reduce costs by 10x. Currently, Anthropic sources its chips from various companies, including NVIDIA, Google, and Amazon. This trio allows the company to keep running its AI infrastructure without major concerns that are often associated with relying on a single chipmaker. But as compute demand intensifies in the AI space, many AI firms are now looking to invest in in-house chips that suit their requirements. Based on recent reporting by The Information, Anthropic is said to be in early talks with a UK-based startup called Fractile. Fractile is becoming a highlight in the AI space due to its new technology, which it calls the Memory Compute Fusion Architecture. The architecture works by moving less data to the DRAM, lowering the reliance on off-chip memory, and doing all the data go-through within the chip itself. For this, the company has devised its own SRAM tech, similar to NVIDIA's Groq LPUs or Groq 3 LPX. NVIDIA's acquisition of Groq allowed it to integrate the latest LPU in its upcoming Vera Rubin ecosystem. These chips act as an AI Inference booster & this is done through the incorporation of large amounts of SRAM and super-high bandwidth for scale-in and scale-up. As such, NVIDIA itself terms the Groq 3 LPU as an Inference accelerator, packing 500 MB of SRAM, 150 TB/s of SRAM bandwidth, and 2.5 TB/s of scale-up bandwidth. These are packaged within the Groq 3 LPX Rack that houses 256 LPUs and a massive 128 GB of SRAM for low-latency processing. Fractile's solution is similar, though the company claims that its architecture targets a 100x speed up in AI inferencing while reducing the costs by 10x versus NVIDIA's Groq. The team at Fractile working on the project comes from big firms such as NVIDIA, Graphcore, and Imagination Technologies. These are some big numbers, but the company has yet to design any test chips, so these early talks can initiate the process of Anthropic's in-house chip development. Anthropic still heavily relies on external chip makers as it signed a Multi-Gigawatt deal with Broadcom, and reports suggest that it will soon add a fourth name to its compute portfolio in the form of AMD.