Anthropic says AI cut AMD chip porting from months to three days

Anthropic says it has broken free of its Nvidia dependence with a multi-chip strategy, and that its Mythos model ported Claude to AMD GPUs in three days with almost no human intervention. The company says it now trains and runs Claude across four major chip families as a single compute pool. The figures come from an Anthropic investor presentation and have not been independently verified.

Porting Claude to AMD GPUs

Anthropic says it used AMD GPUs for the first time this year and let Mythos do the work. Claude was running on AMD after three days, and Mythos ported or wrote nearly 100 percent of the kernels itself. Traditional engineering approaches can take six to twelve months to onboard new accelerators. The figures refer to prompt processing of an Anthropic model on AMD MI355X GPUs. Anthropic now claims in an investor presentation that it has solved this problem using AI itself. Nvidia's biggest competitive edge is the combination of hardware and software, specifically CUDA, the software ecosystem that locks developers into Nvidia chips. Porting AI models to other chip architectures takes so much effort that many companies will not even try.

Four chip families as one pool

Tom Brown, co-founder and Chief Compute Officer at Anthropic, says the company trains and runs Claude on all four major chip families. Anthropic allocates compute from a fungible pool across AWS Trainium, Google TPUs, Nvidia GPUs, and AMD GPUs. The company says it gets more out of every dollar of compute than any other frontier AI lab. Anthropic made an unconventional bet on a multi-chip strategy early, when much of the industry considered the effort not worth it. Each chip needs its own custom software and inference stack, which is expensive and time-consuming. Anthropic has been investing in this work for years and has since become incredibly good at it, Brown says. Some chips are better for training, while others are better for inference. In a compute market where supply is tight, a multi-chip approach lets companies focus on the best price-performance ratio, regardless of the vendor. Because of its dominance in the AI chip market, Nvidia currently enjoys a margin of more than 70 percent, which is unusually high for hardware.

Compute multipliers and owned capacity

Anthropic also relies on compute multipliers, research breakthroughs that squeeze more performance from the same physical computing capacity. The company says these multipliers build on one another, so each new breakthrough amplifies the effect of the previous one. Eventually, Claude should be able to discover such compute multipliers on its own, which would amount to recursive self-improvement, a concept that is intensely debated in AI research but has not been proven yet. Anthropic is also building data center capacity through the Anthropic Direct program, leasing chips and operating them in purpose-built data centers to reduce total cost of ownership.

Financial ties to chip makers

Both major GPU makers now have deep financial ties to Anthropic. In November 2025, as part of a $45 billion deal with Microsoft, Nvidia committed to investing up to $10 billion in Anthropic. According to Anthropic's S-1/A registration statement filed with the SEC on November 10, 2026, Nvidia is also participating in a conditional private placement totaling $26 billion, in which seven technology companies are subscribing to shares at the IPO price. AMD announced a strategic investment of up to $5 billion in July 2026; in return, Anthropic plans to deploy up to two gigawatts of AMD Instinct MI455X GPUs in Helios rack-scale systems, with the first gigawatt phase beginning in the first half of 2027.

Sources: The Decoder