Anthropic’s Model Taught Itself to Run on AMD in “Three Days.” Bad News for Nvidia’s Moat
Anthropic claims its Mythos model enabled Claude to run on AMD chips in three days, a process that previously took six to 12 months. This development could reduce customer switching costs from NVIDIA, which has maintained high gross margins. AMD trades at 37x forward earnings, below its three-year average, with analysts expecting revenue to nearly triple by 2027.
How this was made
The 30-second read
Why it matters
The announcement introduces a new competitive dynamic in the AI GPU market, potentially affecting both AMD and NVIDIA valuations.
Market read
The news could shift investor sentiment in the AI hardware space, prompting revaluation of AMD and NVIDIA stocks.
What to watch
Anthropic's claim is limited to a single model and may not generalize across all AI workloads.
Background
Anthropic, an AI startup preparing for an IPO, highlighted a rapid port of its Claude model to AMD GPUs.
Ticker impact
Anthropic disclosed its Mythos model got Claude running on AMD GPUs in three days, a speed far faster than prior efforts.
likely upside as market prices in faster AMD AI capability
The rapid port reduces switching costs for AI workloads, which could translate into higher GPU sales for AMD.
The same Anthropic announcement suggests reduced reliance on NVIDIA's CUDA, potentially eroding part of NVIDIA's AI moat.
likely downside as investors reassess NVIDIA's AI dominance
While NVIDIA still has a large developer base, the news introduces competitive risk.
Market effects
AI hardware competition intensifies, potentially shifting market share toward AMD.
U.S. semiconductor sector may see reallocation of investor capital between AMD and NVIDIA.
The development could influence global AI infrastructure strategies and supply chains.
Counterpoint
NVIDIA's extensive CUDA ecosystem and developer base may limit any material shift to AMD.
Key entities
- companyAnthropic
AI startup preparing for IPO, developer of the Claude model.
- companyAMD
Semiconductor manufacturer whose GPUs were used for the rapid port.
- companyNVIDIA
Current leader in AI GPU market, whose CUDA ecosystem may be challenged.
