Enterprises prefer Anthropic’s LLMs over OpenAI: report

Anthropic now leads the enterprise large language model (LLM) market by usage, overtaking OpenAI, according to a report from Menlo Ventures released July 31, 2025.

Anthropic is a US-based AI research lab focused on developing large language models.

The report shows Anthropic holds 32% of enterprise market share by usage, while OpenAI is at 25%, a reversal from two years ago when OpenAI led with 50% and Anthropic had 12%.

Google has also seen growth in enterprise adoption of its models during this period.

For enterprise coding tasks, Anthropic’s share stands at 42%, more than double OpenAI’s 21%, according to the report.

The surge in Anthropic’s usage followed the release of its Claude 3.5 Sonnet model in June 2024, and the subsequent Claude 3.7 Sonnet in February 2025.

The report found that most enterprises prefer closed AI models, with only 13% of daily enterprise workloads using open-source models in mid-2025, down from 19% at the start of the year.

🔗 Source: TechCrunch

🧠 Food for thought

1️⃣ Enterprise and consumer AI markets reward different strategies

Anthropic’s rise reveals how the enterprise and consumer AI markets operate with fundamentally different dynamics and preferences.

While OpenAI dominates the consumer space with over 2.5 billion daily prompts to ChatGPT, Anthropic has captured the enterprise market by focusing on reliability and specialized use cases1.

This split reflects deeper market realities: consumers prioritize accessibility and general-purpose capabilities, while enterprises value specialized performance, safety features, and reliability for mission-critical applications.

The revenue implications are striking. Despite Anthropic’s smaller overall user base, it generates approximately 40% of OpenAI’s annual recurring revenue, suggesting significantly higher revenue per enterprise user2.

This demonstrates that enterprise-focused AI companies can achieve substantial revenue growth without matching consumer-focused competitors in raw user numbers, as businesses are willing to pay premium rates for productivity-enhancing AI tools.