Anthropic's Dominance Over OpenAI in the Business LLM Arena

Highlights from ZDNET

A casual survey might easily nominate OpenAI's ChatGPT as the favored Large Language Model (LLM) in the business sector. However, by the middle of 2025, it's actually Anthropic that's leading in enterprise utilization, according to a report from Menlo Ventures, a notable venture capital entity.

It's important to recognize that Menlo Ventures has a vested interest in Anthropic, having supported the company through numerous funding rounds, which includes orchestrating their Series D and being part of a significant $3.5 billion Series E investment that positioned Anthropic’s valuation at $61.5 billion.

Despite potential biases, other industry voices also acknowledge Anthropic's prominent position. As reported in AI Magazine, their Claude LLM family has witnessed tremendous growth, catapulting to $3 billion in annual recurrent revenue—a notable achievement even by aggressive AI benchmarks.

Trailing Anthropic in this arena is OpenAI at 25%, with Google capturing 20%, and another competitor holding 9%. At the far end, a name with only 1% stands alongside various other minor players. Menlo Ventures attributes Anthropic's swift rise to the effectiveness of its models.

It should be noted that Ziff Davis, parent company of ZDNET, initiated legal action against OpenAI in April 2025 concerning alleged copyright violations, which may influence perceptions.

The Catalysts Behind Anthropic's Ascendancy

Three primary elements are fueling Anthropic's prominence. First is the notion dubbed "AI's first killer app" by Menlo Ventures: code generation. Although initially tentative, now more developers prefer AI-driven coding solutions, with Claude securing 42% market prominence, twice that of OpenAI’s share.

Concrete examples illustrate the burgeoning popularity of Anthropic's offerings, with Claude aiding in the creation of a $1.9 billion environment within a year. This includes new categories and tools such as AI-driven Integrated Development Environments (IDEs), and diverse app development platforms.

Another factor is Anthropic's innovative training methodology for its LLMs, employing user-friendly corrective feedback systems that effectively adapt for coding functionalities.

Furthermore, Anthropic pioneers step-oriented problem-solving LLMs that integrate external resources for enhanced resolutions. These advancements offer significant utility in improving AI responses through iterative processes and tool integrations using a novel open-source framework, enhancing real-world data connectivity.

The observation is made that performance, not price, is the prevailing consideration for companies when switching LLMs. This trend indicates a commitment to quality, even when faced with potentially lower costs.

This landscape might evolve with LLM maturity and performance parity, yet current advancements still sway enterprises to invest in the latest, high-performing models.

The Enterprise AI Landscape

According to the study, a notable transition is occurring with businesses opting to run models in live environments rather than solely focus on their development. Particularly startups spearhead this shift, with 74% indicating the majority of their operations are now live productions, while large firms follow with close engagement.

There's also a noticeable decline in open-source LLM usage, dropping to 13% from an earlier 19%. While some models like Llama remain influential, the broader adoption of cutting-edge proprietary models is evident.

In recent times, various open-source models have emerged from different developers, yet they haven't gained widespread traction due to performance lags compared to leading proprietary alternatives. Concerns also linger over some models originating from regions or companies viewed skeptically.

As the AI sector continues to evolve, predictions about its future remain tenuous. With rapid model enhancements and changes, building sustainable AI-driven businesses on contemporary foundational technologies appears promising, with questions still surrounding which will emerge as the ultimate standard-bearers.