TLDRs:
- Anthropic surpasses OpenAI with 32% of enterprise LLM usage, according to a new Menlo Ventures report.
- Claude models power Anthropic’s surge, especially in code-related enterprise tasks with 42% market share.
- OpenAI still dominates consumer AI use, but Anthropic leads in enterprise revenue per user.
- Most enterprises prefer closed AI models, signaling skepticism toward open-source options for mission-critical tasks.
Anthropic, the AI research company behind the Claude model series, has officially overtaken OpenAI in the enterprise market for large language models (LLMs), according to a new report released on Thursday, by Menlo Ventures.
The data reveals that Anthropic now commands 32% of enterprise market share by usage, up from just 12% two years ago, while OpenAI has slipped to 25%, down from its prior 50% lead.
This shift highlights a significant transformation in how businesses are adopting and deploying generative AI, with a clear trend favoring reliability, performance, and specialization over brand familiarity.
Claude 3.5 and 3.7 Drive Enterprise Growth
Anthropic’s surge can be traced back to the June 2024 launch of Claude 3.5 Sonnet, followed by Claude 3.7 Sonnet in February 2025.
These models demonstrated notable improvements in long-context comprehension, reduced hallucination rates, and enhanced multi-step reasoning, features that proved especially valuable to enterprises performing coding and automation tasks.
In fact, Anthropic’s share of enterprise coding usage now stands at 42%, compared to just 21% for OpenAI, according to the same report. The company’s focus on “Constitutional AI”, a framework designed to produce more reliable and safe outputs, has resonated with organizations seeking predictable behavior from AI systems in mission-critical environments.
Enterprise and Consumer Markets Show Diverging Trends
While OpenAI continues to dominate the consumer AI landscape, with over 2.5 billion daily prompts to ChatGPT, the enterprise market tells a different story. Anthropic’s more focused approach has attracted businesses that prioritize stability and security over conversational flair.
Despite a smaller overall user base, Anthropic reportedly generates about 40% of OpenAI’s annual recurring revenue, pointing to higher value per user.
This underlines a key difference, consumers may prefer versatility, but enterprises are willing to pay more for dependable performance tailored to specific workflows.
Closed-Source AI Preferred by Most Businesses
Another takeaway from the Menlo Ventures report is the declining use of open-source AI in enterprise environments.
As of mid-2025, only 13% of daily enterprise AI workloads involved open-source models, down from 19% at the start of the year.
This signals growing enterprise demand for support, security, and accountability, attributes typically associated with closed-source platforms like Claude and GPT models. It also suggests that the battle for enterprise AI dominance may hinge more on technical assurances than philosophical openness.
