Amazon-backed Anthropic is taking a route that looks far more stable than the mass-audience path OpenAI is pushing.

Both companies build advanced AI systems used in chatbots, coding tools and image generators, but their strategies for making real, steady revenue are not the same. Anthropic is focused on business clients who use AI for work that saves time and lowers costs, while OpenAI is focused on everyday users replacing search queries with chatbot conversations.

Naturally, this difference has produced two very different business outcomes.

OpenAI has massive visibility thanks to ChatGPT, which it says pulls more than 800 million weekly users. That has pushed OpenAI to a $13 billion annual revenue run rate.

But the company itself says only around 30% of that comes from businesses. The rest is consumer subscriptions through a $20 per month Plus plan and a $200 per month Pro plan, alongside a slower free tier, a strategy that depends on huge scale continuing and is extremely costly to maintain.

Anthropic’s enterprise focus changes its revenue dynamics

Anthropic reported that about 80% of its revenue comes from corporate customers.It recently said it had around 300,000 business clients.

These customers use Anthropic’s Claude models for tasks such as writing code, reviewing legal language and accelerating billing workflows. These are job functions that directly affect budgets and time, making it clearer for companies to justify paying for them.

A July survey from Menlo Ventures, which has invested in Anthropic, found that the company held 42% market share in coding applications, compared to OpenAI’s 21%.In broader enterprise AI usage, Menlo estimated Anthropic at 32% versus OpenAI at 25%.Anthropic is also close to OpenAI in revenue terms despite having far fewer active users.

The company is already at a $7 billion annual run rate and expects to reach $9 billion by the end of the year. That means its revenue per user is much higher because those users are businesses, not casual chatbot users.

Both companies rely on major tech backers for compute and infrastructure. OpenAI works closely with Microsoft, which integrates its models into products across Office and enterprise software.

Anthropic receives support from Amazon and Google, giving it access to large-scale computing resources and corporate customer channels.

Corporate demand for Anthropic is clear. In September, Microsoft announced that Claude would be added to its Copilot software suite, and it happened despite Microsoft already being deeply tied to OpenAI, showing that enterprise users specifically requested Claude and its performance in work scenarios.

OpenAI’s revenue path remains unclear as it courts mass users

OpenAI’s consumer-first approach has not settled on a fully reliable business model. Subscriptions alone are unlikely to cover the extremely high cost of training and operating advanced models.

The most obvious additional revenue stream would be advertising, but ads in chatbot responses pose major problems. Unlike search ads, chatbot conversations do not have natural placement for paid suggestions.

Users are unlikely to accept brand messaging inside personal chat responses. OpenAI would also have to compete with Google, whose advertising infrastructure is far more established.

OpenAI is trying to expand among businesses as well, but there are concerns about whether its consumer-driven brand aligns with what corporate customers want. Recently, OpenAI said it would allow adults to have erotic conversations with ChatGPT.

The company has also pushed for a lighter regulatory approach. Even if OpenAI provides stricter controls in corporate settings, its reputation may raise hesitation among companies looking for consistent and predictable tools.

Performance comparisons also show the competitive AI space is changing. Vals AI, a company that evaluates language models on business tasks in finance, law and coding, ranked the newest version of Claude at the top of its benchmark.

Rayan Krishnan, the co-founder of Vals, said, “Anthropic is laser-focused on these agentic enterprise use cases and they’re playing a very competitive game with OpenAI right now.”

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.