Anthropic Gains on Enterprise AI Spending Share

Anthropic has closed the gap with OpenAI in enterprise paid usage, according to the Ramp AI Index March 2026. Ramp measures the share of U.S. businesses with paid subscriptions to AI models, platforms, and tools and reports overall business AI adoption at 47.6%. Anthropic's paid footprint jumped to nearly one in four Ramp customers, up from one in 25 a year ago, while OpenAI saw a 1.5 percentage point decline in the same window.

The Ramp AI Index uses transactional telemetry from 50,000+ American businesses and billions of dollars in corporate spend, sourced from corporate card and bill-pay flows. The index tracks paid subscriptions rather than API call volume or model parameter counts, so it reflects procurement and subscription decisions rather than raw inference consumption. Key numeric signals from the March update are:

• •Overall enterprise AI subscription penetration: 47.6%
• •Anthropic paid share: nearly one in four Ramp customers (up from one in 25 last year)
• •OpenAI paid share: down by 1.5 percentage points in the latest update

These metrics capture adoption at the customer-account level. They do not directly measure compute spend, prompt traffic, or model selection for latency-sensitive production workloads. Third-party estimates cited alongside the index place Anthropic as high as 40% of some enterprise LLM spend versus 27% for OpenAI, though those figures come from investor and market-analytics commentary and should be treated as supplementary.

This shift matters because subscription share is a leading indicator of enterprise procurement momentum. A rapid rise in paid Anthropic subscriptions signals successful GTM execution, enterprise pricing, or feature fit that resonates with treasury and procurement teams. For practitioners, the implications are practical: more teams will include Claude-family models and Anthropic APIs in RFPs, POCs, and hybrid architectures, increasing the importance of multi-provider orchestration and evaluation.

Anthropic's growth reflects broader trends: enterprise buyers diversifying away from single-vendor dominance, increased emphasis on model safety and controllability, and competitive pricing or contractual terms attractive to finance teams. OpenAI's small decline does not imply technical inferiority but highlights that market share leadership in public attention does not guarantee shackled procurement growth.

Expect more vendor-neutral benchmarking in enterprise procurement, increased investment in model-agnostic routing layers, and new contract terms that emphasize cost-per-inference predictability. Track next Ramp AI Index updates for momentum persistence and corroborate with API telemetry where available.

Practical actions for teams:

• •Re-evaluate vendor shortlists to include Anthropic for POCs and A/B testing.
• •Instrument cost and latency metrics per-provider to understand real-world tradeoffs.
• •Build model-agnostic adapters and feature-flagging to switch providers rapidly as price-performance or compliance needs change.

This market signal is not a technical benchmark result but a procurement and adoption shift with operational consequences. Teams should treat it as evidence that enterprise adoption is diversifying and plan procurement, monitoring, and benchmarking accordingly.