Anthropic, by several accounts, seems to be leading in enterprise AI, but it’s lead isn’t uniform across the board.
New data from the Ramp AI Index — which tracks corporate card and invoice spend from more than 50,000 US businesses — shows a striking pattern in how different industries are splitting their AI subscription dollars between Anthropic and OpenAI. The headline finding: the sectors that have gone deepest on AI adoption are also the ones most decisively in Anthropic’s corner. The sectors that are still cautious about AI tend to lean OpenAI, or are too close to call.
Where Anthropic Dominates
The Information sector stands apart. It records the highest paid AI subscription rates of any industry in the Ramp dataset — and Anthropic’s lead there is the widest in the chart. Finance and Professional Services follow a similar script: heavy AI adoption, clear Anthropic preference. These are industries staffed by knowledge workers whose output maps naturally onto what large language models are good at — drafting, summarizing, analyzing, generating code. The fit is obvious, and the spending data confirms it.
This tracks with what Ramp’s broader data has shown over the past year: Anthropic went from capturing roughly 10% of combined OpenAI-plus-Anthropic business subscription spend in early 2025 to commanding over 65% by February 2026. That shift wasn’t driven by the Fortune 500 alone. Claude Team, designed for small and mid-sized businesses with shared workspaces and administrative controls, emerged as the engine of growth — suggesting Anthropic’s enterprise preference cuts across company size, not just verticals.
The Middle Tier: Close Fights in Transitional Sectors
Education, Manufacturing, Wholesale Trade, Retail, Transportation, and Real Estate cluster in the middle of the Ramp chart. In most of these, OpenAI still holds a marginal lead or is neck-and-neck with Anthropic. Crucially, these sectors are also lower on overall AI adoption rates. They are still in earlier stages of integrating AI into their workflows, and OpenAI’s brand recognition — built on ChatGPT’s consumer dominance — likely plays a role in initial purchasing decisions.
The pattern suggests that sector-level AI maturity is the key variable. Early-stage adopters default to the most familiar name. As usage deepens and teams develop more specific requirements, they appear to migrate toward Anthropic. Claude Opus 4.6, which launched in February 2026 and set new benchmarks across coding and agentic tasks, has likely accelerated this gravitational pull among power users who have outgrown general-purpose chat interfaces.
Where OpenAI Still Leads
At the bottom of the AI adoption curve — Healthcare, Construction, Arts, Agriculture, and Hospitality — OpenAI maintains its edge, or the gap is too small to read clearly. These sectors are late to AI adoption by nature: regulatory complexity in healthcare, fragmented workforces in construction and agriculture, margin pressure in hospitality. For businesses just beginning to experiment with AI tools, ChatGPT’s familiarity and OpenAI’s extensive distribution through Microsoft’s ecosystem provide a natural on-ramp.
Healthcare is the most interesting outlier. Despite being low on the adoption chart, it is an active battleground for both companies. Anthropic has made targeted moves into the space — Claude for Enterprise is now available to HIPAA-compliant organizations, and the company counts Epic, one of the most widely deployed healthcare technology platforms, as a customer. The Ramp data captures spend today; it may not yet reflect where the sector is heading.
The Underlying Dynamic
There is a logic to the pattern that goes beyond brand preference. The sectors where Anthropic leads are sectors where AI is being used for complex, high-stakes, multi-step tasks — exactly the workloads that require reliability, long context handling, and nuanced output. Anthropic’s positioning on safety and consistent model quality has resonated most in environments where the cost of a bad AI output is real.
Anthropic’s revenue growth tells the same story in aggregate: the company crossed a $30 billion annualized run rate in early 2026, up from roughly $9 billion at the end of 2025, with 80% of revenue coming from business customers. That enterprise-heavy mix is the direct result of winning in high-adoption industries first, and letting the pattern spread.
The Ramp data makes one thing clear: AI adoption and Anthropic preference move together. As more industries mature into serious AI users, the competitive map is likely to continue shifting — industry by industry, use case by use case.