Anthropic’s Claude now accounts for 24.4% of all business AI usage, up from approximately 4% one year ago, with month-over-month growth holding at 4.9%. The data, published by Digit.fyi and corroborated by an Axios analysis of enterprise AI spending, represents a rapid reshaping of the enterprise AI market.
OpenAI’s share is moving in the opposite direction. While exact figures vary by source, both reports confirm a declining trend in OpenAI’s enterprise penetration over the past two quarters. Enterprise buyers surveyed by Axios cited two factors driving the shift: Claude’s reliability in production workloads and Anthropic’s public safety commitments.
The Safety Premium
The timing of the enterprise surge matters. Anthropic’s fastest growth period overlaps directly with the company’s public standoff with the U.S. Department of Defense. After Anthropic refused to allow Claude’s use in autonomous weapons systems and domestic surveillance, the Pentagon labeled the company a “supply chain risk” and ordered contractors to remove it. Today, Reuters reported that military users are pushing back on the removal order because Claude is too deeply embedded to replace.
The enterprise market appears to be reaching its own conclusion about what Anthropic’s safety stance signals. Several companies have announced switches to Claude following the Pentagon controversy, according to Digit.fyi. The logic from enterprise procurement teams, as described by Axios’s sources: a company willing to walk away from government revenue over ethical constraints is less likely to ship reckless model updates that break production systems.
Growth at Scale
A 4.9% monthly growth rate at 24.4% market share is more notable than the same rate at 4%. At its current trajectory, Claude would cross 30% of enterprise AI usage by mid-summer 2026, putting it within striking distance of OpenAI’s leading position. Whether that pace holds depends on several variables: Anthropic’s capacity to serve enterprise demand, the performance of Claude’s next model generation relative to GPT-5’s subagent architecture, and whether the Pentagon dispute escalates in ways that create procurement risk for commercial buyers.
For now, the numbers tell a clear story. Enterprise AI buyers are choosing Claude in growing numbers, and the company’s willingness to draw ethical boundaries on government use has not slowed commercial adoption. By the available data, it has accelerated it.