Corporate spending data platform Ramp has disclosed that Anthropic’s Claude series saw its share of the enterprise AI subscription market rise to 41% in May 2026, surpassing OpenAI’s 39.5% for the first time — with the latter remaining essentially flat from the previous month. Ramp Chief Economist Ara Kharazian noted that Anthropic’s best-ever month for enterprise procurement happened to be the very month in March when the Department of Defense designated it a „supply chain risk“; „when your model is singled out as ‘too dangerous to use,’ it actually builds a considerable halo effect.“ The above data comes from over 70,000 enterprise customers on the Ramp platform. In roughly a third of transactions where model details are visible, enterprise usage is concentrated across various versions of Claude Opus, especially more recently released ones; Mythos has had a shorter shelf life (available to limited users starting April), and Fable 5 was taken down just days after its launch, and the drag on overall revenue cannot yet be precisely quantified from spending data.
The article notes that this trend provides a fuller context: in late May, Anthropic completed a $65 billion funding round at a $96.5 billion valuation, filed confidential IPO documents that same month, and announced it would soon record its first profitable quarter. Entering June, Commerce Secretary Lutnick sent a letter calling for a full suspension of foreign nationals — including Anthropic’s non-US employees — from accessing Mythos 5 and Fable 5, prompting the company to take down both models entirely. Kharazian believes this friction with the government is likely to further enhance Anthropic’s brand appeal among enterprise users; however, TechCrunch points out that the controversy itself could cast a shadow over the upcoming IPO — as public market investors are typically cautious about companies in disputes with the government. Opus 4.8 remains publicly available and is currently the main model driving actual enterprise spending.