AI BriefingCompute Wars: OpenAI vs Anthropic
Key point
Anthropic is closing the gap with OpenAI by significantly increasing its compute capacity through AWS infrastructure expansion.
Details
The groundbreaking performance shown by Anthropic's Opus 4.5 appears to be thanks to large-scale compute secured through the AWS Madison and New Carlisle sites. Through this, Anthropic's compute capacity has more than doubled compared to before.
This expansion brings Anthropic close to OpenAI's total capacity, and in terms of practically available capacity for running new models, it could actually be at a higher level.
It takes about 6 months or more from securing capacity to launching a model. Therefore, OpenAI's newly secured capacity is more likely to align with the launch timing of the 'spud' model rather than GPT-5.4.
OpenAI is expected to pull ahead again in terms of compute availability in the second half of 2026, but the gap between the two companies is projected to narrow again in 2027. While OpenAI holds higher planned capacity, Anthropic is also aggressively expanding its infrastructure.
Ultimately, breakthroughs in new AI capabilities are very likely to come from deploying greater compute resources.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Anthropic Pacing to $100B ARR; Rumors of OpenAI Solving Navier-Stokes Spark ControversyOpenAI and Anthropic's RSI Strategy and the Uncertainty of AI Profitability