Anthropic Strengthens Claude: Why New Corporate Customers Choose It Over ChatGPT

In corporate AI, a notable shift has begun: among companies making their first AI tool purchases, Claude from Anthropic took the lead over ChatGPT in March 2026. This is not yet a mass exodus from the market, but already an important signal about how businesses choose operational AI products.

Shift in Purchasing

According to the March 2026 Ramp AI Index, 73.3% of spending among companies making their first AI tool purchases went to Anthropic, while 26.7% went to OpenAI. Back in December 2025, the picture was reversed: OpenAI led with roughly a 60-40 ratio. In January, the forces nearly evened out, and by February the tide had turned. Importantly, this is not about the entire corporate market, but specifically about new buyers. Yet it is precisely in this segment that we typically see where future demand is shifting.

These figures do not prove that business has already massively abandoned ChatGPT. Large companies rarely change their stack all at once, and existing contracts and internal processes persist for a long time. However, such a shift in first-time purchases is a strong leading indicator. It shows that when teams make a new choice, they are increasingly looking not at the loudest brand, but at which tool integrates better into real, everyday work.

Why Claude Is Chosen

The main driver of Claude's growth is not abstract "intelligence," but practical usefulness for teams that write code, maintain documents, and build repeatable processes. Anthropic has bet on scenarios where the model should not simply answer in a chat, but maintain context over time, follow rules, and assist within a specific workflow. Hence the interest in Claude Code, work modes like Cowork, and skills in which a team can package regular actions.

• Cleaner and more maintainable code without unnecessary architectural exotica
• Stable work with long context and large documents
• Less drift in long sessions and better adherence to instructions
• More flexible pricing tiers for teams caught between $20 and $200

A separate factor is price. At OpenAI, many subscriptions show a sharp jump from Plus at $20 to Pro at $200, whereas Anthropic offers intermediate tiers: Max at $100 and $200 with different limits. For small teams this feels more accessible. That said, the API situation is more complex: GPT-5.4 is noticeably cheaper than Anthropic's flagship models on input and output tokens. So choosing Claude is not about minimizing cost, but about the feeling that result quality and predictable behavior justify the additional expense.

OpenAI's Response

OpenAI responds not only with brand, but with product strategy. On March 17, 2026, the company released GPT-5.4 mini and nano—models optimized for subagents, background tasks, and cheap bulk operations. The logic is clear: the large model handles planning, while lighter versions take on parsing, classification, repository traversal, and other narrow tasks. This is a strong bet on the infrastructure layer, where speed, cost, and scale matter, not just the quality of a single chat answer.

In pure benchmarks, OpenAI's positions remain strong, especially where speed and engineering efficiency matter. GPT-5.4 is slightly ahead in coding tests and real bug-solving, while Claude Opus 4.6 shows better performance in complex reasoning and long texts that humans evaluate. Hence the difference in perception: ChatGPT is often chosen where scaling economics matter, while Claude is chosen where a careful, multi-step executor is needed.

Sentiment in communities is also shifting, though this is more a signal than market proof. Around Claude, an image of a "work tool" is gaining popularity, while around ChatGPT there is increasingly discussion of censorship limitations, fatigue in long sessions, and loss of focus. Even a viral Reddit headline is telling, and on Anthropic's side there is already the reverse price of growth: the faster the audience grows, the more visible are the errors, limits, and access problems.

"ChatGPT refugees are already here."

What This Means

The corporate AI market is entering a phase where the winner is not simply the most famous or cheapest model, but the product that best fits into work processes. While OpenAI strengthens the economics and ecosystem of subagents, Anthropic is winning new corporate customers through convenience, stable behavior, and the sense of a "work tool," not just a smart chatbot.