In the pre-2026 era, technology innovation defined AI competition: thinking models, larger models, RL environments, agents, and harnesses. In 2026, the piece argues, business model innovation matters more. Anthropic and OpenAI have both re-rated their run rates by segmenting—selling the same models to different buyers at different prices. The article frames the race as two SailGP boats in San Francisco Bay, with each company adding strategy beyond technical one-upmanship. Anthropic moved first with enterprise metered billing in March 2026, doubling revenue in a quarter. OpenAI answered about three months later by cutting the price of its most affordable model, Luna, by 80%, pulling itself to within a boat length.

OpenAI is now approaching a $70b run rate. Margins remain ambiguous, and the article suggests gross profit dollars will be a better way to evaluate the relative strength of the businesses than headline run rates. That such strategic changes can move revenue at this scale within a single quarter indicates the market is still fluid: both companies are expected to near $100b in revenue by the end of the year.

Anthropic‘s leaked S-1 reinforces that even the largest AI buyers have not fully chosen a side. Amazon and Google accounted for nearly a quarter of Anthropic‘s revenue last year, and neither is locked into a long-term contract. With technical innovation, customer segmentation, and price discrimination all in play, the author frames the race as a business-school case unfolding in real time.