OpenAI Slams Anthropic in Memo Over Compute Power and AI Rivalry
OpenAI sent an investor memo criticizing its chief rival, Anthropic, by highlighting a significant disparity in projected computational capacity. OpenAI projects reaching 30 gigawatts (GW) of compute by 2030, while it estimates Anthropic will only reach 7 to 8 GW by the end of 2027. The memo underscores OpenAI's belief that its infrastructure growth is materially ahead of its competitor. Both companies are major players in the LLM market, valued at over $1 trillion, and are preparing for potential IPOs while competing with tech giants like Google and Meta. Anthropic, which was founded by former OpenAI employees, recently announced a new model for a cybersecurity project, further intensifying the rivalry.
Ad slot
In a recent memo sent to investors, OpenAI criticized its chief rival, Anthropic, suggesting that the competitor's compute strategy is significantly limited compared to its own aggressive growth plans. The document highlights a widening gap in projected computational capacity as the two AI giants battle for market dominance in the large language model (LLM) sector.
Compute Capacity Showdown
OpenAI's memo, reported by CNBC and Bloomberg, directly contrasts the projected infrastructure growth of the two companies. The core of the criticism centers on computational power (compute):
OpenAI's Projection: The company plans to achieve 30 gigawatts (GW) of compute capacity by 2030.
Anthropic's Projection: OpenAI estimates that Anthropic will reach only 7 to 8 GW by the end of 2027.
OpenAI stated that its growth trajectory is "materially ahead and widening," emphasizing its superior infrastructure ramp.
Market Positioning and Competition
Ad slot
The rivalry between OpenAI and Anthropic is intensifying as both companies vie for leadership in the rapidly evolving AI market. Both firms are collectively valued at over $1 trillion and are reportedly preparing for potential Initial Public Offerings (IPOs).
Industry Pressure: Both companies are working to convince investors of sustainable business models capable of competing with cash-rich tech giants like Google and Meta.
Anthropic's Moves: Anthropic, founded in 2021 by researchers who left OpenAI, has been active in the enterprise sector. Recently, the company announced a powerful new model for a cybersecurity initiative called Project Glasswing.
OpenAI's Strategic Advantage
OpenAI framed its approach as fundamentally different from Anthropic's, pointing to its ability to leverage infrastructure growth and product innovation. The company emphasized several key advantages:
Compounding Advantage: OpenAI argues that each new generation of infrastructure allows for the training of more capable models, making every token more intelligent than the last.
Cost Efficiency: The combination of algorithmic gains and hardware improvements is lowering the cost per unit of intelligence.
Product Leadership: Having launched ChatGPT in late 2022, OpenAI claims its superior products drive higher revenue and allow it to democratize AI by making tools available to millions of people, sometimes for free.
Anthropic did not issue an immediate response to the request for comment regarding the memo.
