Peter Smith
⬤ Fresh numbers from Mercury paint a striking picture of how new startups are picking their first AI partner. By the final quarter of 2025, Anthropic pulled ahead of OpenAI in capturing first-time AI spending from startups banking with Mercury. The trend charts show Anthropic's share climbing steadily quarter after quarter, eventually flipping the script on OpenAI's long-standing lead in the early-stage market.
⬤ Back in late 2022 and through most of 2023, OpenAI dominated the space, snagging the bulk of new startup AI budgets while Anthropic hovered around 20-25% of the market. Things started shifting throughout 2024 as Anthropic chipped away at that gap. Then 2025 hit, and the pace picked up dramatically—Anthropic's portion of first-model picks jumped past the 50% mark by Q4, while OpenAI's slice of the pie shrank in lockstep.
⬤ What's driving this flip? The data commentary points straight to Anthropic's product upgrades. Tools like Claude Code and the Opus 4.5 model represent real leaps forward, giving startups compelling reasons to rethink their AI infrastructure from day one. Early-stage companies move fast when they spot meaningful gains in model performance, developer tools, and overall productivity, and Mercury's numbers show plenty of startups voting with their wallets.
⬤ This shift in first-dollar AI spending matters more than it might seem. The infrastructure choices startups make early on tend to stick as they grow and scale up operations. If Anthropic keeps winning over newly formed companies, that momentum could reshape competitive dynamics across the AI model landscape as these startups mature and expand their footprint.
Peter Smith
