Anthropic May Never Catch OpenAI. But It’s Already 40% as Big.

For years, the artificial intelligence race sounded like a one-company show: OpenAI launches ChatGPT, the internet loses its mind, businesses scramble to “do AI,” and every executive suddenly discovers a deep spiritual connection to productivity software. But while OpenAI was busy becoming the default name in generative AI, Anthropic was quietly building something harder to ignore: a fast-growing enterprise AI business that may not need to beat OpenAI in consumer fame to become enormous.

The headline number is striking. Based on reported annualized revenue figures, Anthropic has been roughly 40% as big as OpenAI by revenue run rate at key points in the race. That does not mean Anthropic has 40% of OpenAI’s users, brand awareness, infrastructure, or cultural gravity. It means something more specificand possibly more important for investors and business buyers: Anthropic has turned Claude into a serious revenue engine.

OpenAI still has the brand advantage. ChatGPT is practically a household verb. It has massive consumer reach, a deep partnership history with Microsoft, and one of the most valuable AI ecosystems in the world. But Anthropic has found its own lane: enterprise AI, coding agents, safety-focused positioning, and a product experience that many professionals describe as thoughtful, precise, and unusually good for complex work.

In other words, Anthropic may never catch OpenAI in the way Instagram never became Facebook and Shopify never became Amazon. But that does not make it small. Sometimes the second giant is still a giant. And in AI, the second giant may be minting revenue faster than anyone expected.

What Does “40% as Big” Actually Mean?

When people say Anthropic is “40% as big” as OpenAI, they are usually talking about annualized revenue or revenue run ratenot total valuation, users, model quality, or cash reserves. Revenue run rate is a forward-looking business metric. It takes a recent sales pace and annualizes it, basically asking: “If the company kept earning at this current speed for a full year, what would revenue look like?”

That is useful, but it is not magic. A run rate can rise quickly when enterprise customers suddenly consume more tokens, buy larger subscriptions, or move from pilots to production. It can also exaggerate momentum if a company has one unusually hot month. In AI, where large customers may pay through usage-based pricing, a few major deployments can make the graph look like it drank five espressos and joined a rocket launch.

Still, the comparison matters. OpenAI reportedly crossed more than $20 billion in annualized revenue in 2025 and later topped $25 billion by early 2026. Anthropic, meanwhile, was reported around $9 billion in annualized revenue by late 2025 or early 2026. Depending on which exact numbers you compare, that puts Anthropic at roughly 36% to 45% of OpenAI’s size by run-rate revenue. That is where the “40%” idea comes from.

For a company founded in 2021, that is absurdly fast. Most software companies spend years trying to reach $100 million in annual recurring revenue. Anthropic has been discussed in billions. The AI boom has compressed what used to be a decade of enterprise software scaling into a few dizzying quarters. It is exciting, a little terrifying, and probably giving every CFO in Silicon Valley a resting heart rate of 140.

Why OpenAI Still Has the Lead

OpenAI’s advantage begins with distribution. ChatGPT became the first AI product millions of people used for homework help, coding, writing, brainstorming, customer support drafts, travel planning, spreadsheet formulas, and the occasional suspiciously polished resignation email. That consumer familiarity matters because enterprise adoption often starts with employee behavior. Workers use a tool at home, bring expectations into the office, and eventually the company procurement team has to turn “everyone is using it anyway” into a managed contract.

OpenAI also has a broad product stack. ChatGPT, API access, enterprise plans, coding tools, multimodal models, image generation, voice features, agents, and deep integrations with business workflows all help create a platform effect. The company is not selling only a chatbot. It is trying to become the operating layer for AI-assisted work.

Then there is the Microsoft factor. Microsoft’s relationship with OpenAI helped put generative AI inside the enterprise conversation at massive scale. Even as the partnership evolves, Microsoft’s cloud infrastructure, distribution channels, and enterprise credibility have helped OpenAI move from viral consumer product to boardroom priority.

OpenAI also owns mindshare. In technology markets, mindshare can become a moat. Developers build around what they know. Executives ask for what they have heard of. Journalists write about the company that defines the category. Recruiters, investors, and customers all respond to the gravitational pull of the market leader. OpenAI has that gravitational pull.

Why Anthropic Is Growing So Fast Anyway

Anthropic’s growth story is not about becoming a louder ChatGPT. It is about becoming indispensable in serious work environments. Claude has developed a strong reputation among professionals who care about long-context reasoning, document analysis, coding, writing quality, and enterprise-grade reliability. That may sound less flashy than consumer virality, but enterprise software has always rewarded products that quietly save expensive people expensive time.

Claude Code has become one of Anthropic’s most important growth engines. Coding is a perfect AI wedge because software teams can measure output quickly. Did the agent understand the codebase? Did it fix the bug? Did it write usable tests? Did it save a developer two hours or create a disaster shaped like a pull request? The feedback loop is brutally clear.

Anthropic has leaned into that. Claude Code works across developer environments and is designed to help with real codebase tasks, from debugging to refactoring. For teams drowning in maintenance work, migration projects, test coverage, and internal tools, that is not a novelty. It is a potential productivity layer.

Anthropic has also benefited from a safety-first brand. The company was founded by former OpenAI employees and has long emphasized responsible scaling, model behavior, and enterprise trust. For highly regulated industries, that positioning matters. Banks, law firms, healthcare companies, insurers, and government-adjacent organizations do not just ask, “Which model is smartest?” They ask, “Which vendor can survive legal review, security review, procurement review, and one very intense meeting with compliance?”

The Enterprise AI Race Is Not the Consumer AI Race

Consumer AI is about habit. Enterprise AI is about budgets, controls, integration, security, governance, and return on investment. A consumer might choose an AI assistant because it feels friendly. A company chooses one because it can connect to internal systems, respect permissions, reduce repetitive work, and not leak confidential data into the digital soup.

This difference helps explain why Anthropic can look smaller and still be financially huge. OpenAI has massive consumer reach, but consumer scale does not automatically translate into enterprise dominance. Enterprise buyers often test multiple models. They may use OpenAI for general productivity, Anthropic for coding or document-heavy workflows, Google for cloud-native AI features, and open-source models for cost-sensitive internal systems.

In that world, the winner may not take all. The market may become multi-model, with companies routing different tasks to different providers. A legal research workflow might prefer one model. A coding workflow might prefer another. Customer service might use a cheaper, faster model. Strategy teams might use the most capable reasoning model available that week. AI procurement is starting to look less like buying one software suite and more like managing a portfolio of digital brains.

That is good news for Anthropic. It does not need every consumer to say “Claude” the way they say “ChatGPT.” It needs enough high-value enterprise use cases where Claude performs well enough to justify serious spending.

The Valuation Question: Brilliant Growth or Bubble Math?

AI valuations are now so large they can make normal startup math look like a lemonade stand ledger. Anthropic has raised huge funding rounds and attracted major backing from cloud and investment giants. OpenAI has also raised at historic scale and is discussed as one of the most valuable private technology companies on earth.

But revenue is only one side of the story. Frontier AI companies burn extraordinary amounts of money on compute, research talent, inference costs, data infrastructure, safety testing, and product development. Training and serving advanced models requires chips, energy, data centers, and engineering teams that cost more than some national infrastructure projects. “We have great revenue” is impressive. “We have great revenue after paying for the supercomputer mountain behind it” is the harder part.

That is why investors care not only about who has the biggest revenue number, but who can convert revenue into durable margins. If Anthropic’s enterprise customers keep growing and its products command premium pricing, the company can justify a massive valuation. If usage costs remain high and competition forces prices down, the math gets more complicated.

OpenAI faces the same pressure. Its consumer reach is enormous, but serving hundreds of millions of users is not cheap. Free users create influence, data, and habit, but paying users fund the machine. The strategic question for OpenAI is how to turn global adoption into durable profit without slowing product growth. The strategic question for Anthropic is how to keep enterprise momentum without being crushed by compute costs or copied by larger rivals.

Why “Catching OpenAI” May Be the Wrong Goal

The phrase “catch OpenAI” assumes there is one race with one finish line. That is probably too simple. AI is becoming a stack of different markets: consumer assistants, developer tools, enterprise agents, cloud APIs, robotics, scientific research, healthcare workflows, education tools, search, advertising, productivity suites, and custom business automation.

OpenAI may remain the most recognized AI company in the world. Anthropic may become the preferred AI provider for certain enterprise and coding-heavy workflows. Google may dominate AI inside search, Android, Gmail, and cloud. Microsoft may win by embedding AI into the software businesses already live in. Meta may push open models into broader developer ecosystems. The “AI race” is not a 100-meter sprint. It is more like a citywide marathon where everyone keeps changing shoes, roads, and the definition of running.

Anthropic does not need to own the whole city. It needs to own profitable neighborhoods. Coding agents are one. Complex knowledge work is another. Enterprise AI governance could be a third. If Claude becomes the model that teams trust for high-value work, Anthropic can build a business that remains huge even if OpenAI stays bigger overall.

Examples of Where Anthropic Can Win

1. Software Engineering Teams

Software teams are ideal early adopters because they already work in structured digital environments. Claude Code can assist with bug fixes, code review, documentation, tests, and refactors. A developer who saves one day per week creates measurable value. Multiply that across a team, and suddenly the subscription cost looks less like an expense and more like a bargain hiding in a hoodie.

2. Legal and Professional Services

Law firms, consulting firms, accounting teams, and corporate strategy departments spend enormous time reading, summarizing, comparing, and drafting documents. Claude’s reputation for handling long context and careful writing makes it appealing for these use cases. In professional services, saving time on document-heavy work can directly improve margins.

3. Financial Analysis

Finance teams need models that can analyze reports, summarize trends, compare scenarios, and assist with research. Accuracy, traceability, and risk controls matter. Anthropic’s enterprise positioning gives it a credible opening with buyers who want powerful AI but are allergic to chaos.

4. Internal Knowledge Workflows

Large companies are full of internal documents nobody wants to read but everyone needs to understand. Policies, manuals, contracts, support tickets, product specs, technical docs, and meeting notes can become searchable and usable through AI assistants. This is less glamorous than viral chatbots, but it is exactly where enterprise AI can become sticky.

The Risks Anthropic Still Faces

Anthropic’s momentum is real, but the company is not gliding on a frictionless cloud made of venture capital and good intentions. It faces serious risks.

First, OpenAI is not standing still. OpenAI has the talent, capital, distribution, and urgency to fight back in coding and enterprise AI. If OpenAI improves its developer tools and enterprise packaging, Anthropic’s advantage may narrow.

Second, Google is becoming more aggressive. Google has unmatched infrastructure, research depth, consumer reach, and enterprise cloud relationships. It can compete on model capability, price, and integration across products used by billions of people.

Third, AI pricing may fall. As models become cheaper to run and open-source alternatives improve, customers may become more price-sensitive. If enterprises route routine tasks to cheaper models, premium labs will need to prove they are worth the extra cost.

Fourth, the market is still young. Today’s leader in a specific workflow can lose ground quickly if a rival releases a better model, a cheaper API, or a more convenient product. AI loyalty is not like database loyalty. Switching costs exist, but they are not always permanent.

Experiences From the Front Lines of the Anthropic vs. OpenAI Debate

For people actually using these tools at work, the OpenAI-versus-Anthropic debate is less like a sports rivalry and more like choosing the right coworker for the task. OpenAI often feels like the universal generalist: fast, familiar, broad, and easy to introduce to almost anyone. Anthropic often feels like the careful specialist: strong with long documents, thoughtful in tone, and particularly useful when the task requires patience rather than fireworks.

A common experience among knowledge workers is that ChatGPT becomes the first stop because it is already part of their daily rhythm. They use it to draft emails, outline presentations, clean up messy notes, brainstorm titles, explain confusing topics, and generate quick first drafts. It is the AI equivalent of a Swiss Army knife that somehow also took a management consulting course.

Claude, on the other hand, often enters through heavier workflows. A team might try it when they need to summarize a long PDF, compare multiple documents, review dense technical material, or produce a polished memo. The experience can feel different: less like asking a quick assistant and more like handing work to a patient analyst. That distinction matters because enterprise buyers do not always choose the most famous tool. They choose the tool that makes a painful workflow less painful.

Developers tell a similar story. A casual user may judge an AI assistant by how well it writes a birthday message or explains a recipe. A developer judges it by whether it understands a messy codebase, follows instructions, avoids breaking tests, and proposes changes that survive review. In this environment, Claude Code has become a serious contender because coding agents are not judged by charm. They are judged by output.

Business leaders also experience the rivalry through procurement. One department may already be paying for ChatGPT. Another may request Claude for coding. A third may want Gemini because the company is already on Google Cloud. Suddenly, the CIO is not choosing “the AI.” The CIO is managing an AI vendor map. That is where Anthropic’s 40% scale becomes meaningful. It shows the company is not a niche curiosity. It is large enough to be part of serious enterprise planning.

The most practical lesson is that the best AI strategy is rarely emotional loyalty to one brand. Smart teams test multiple models, measure outcomes, track costs, and match tools to tasks. They ask which model reduces review time, which one produces fewer errors, which one employees actually use, and which one passes security requirements. That experience-driven approach is much healthier than treating AI companies like superhero franchises.

So, will Anthropic catch OpenAI? Maybe not in consumer awareness. Maybe not in total users. Maybe not in cultural dominance. But in the daily experience of many businesses, Anthropic has already crossed the most important threshold: it is credible, useful, and expensive enough to matter. That is how a company can be “only” 40% as big and still feel like one of the most important players in technology.

Conclusion: The AI Race Is Bigger Than One Winner

Anthropic may never become OpenAI. That is not an insult. OpenAI has a once-in-a-generation consumer brand, huge usage, powerful partners, and a head start that changed the entire technology industry. But Anthropic has done something almost as impressive: it has built a business large enough to make the AI race feel genuinely competitive.

The “40% as big” figure captures a deeper truth. Anthropic is not merely chasing attention; it is capturing enterprise dollars. It is proving that the AI market will not be decided only by the chatbot with the most famous name. It will be decided by workflows, trust, integrations, developer productivity, cost control, and measurable business value.

OpenAI may remain the face of consumer AI. Anthropic may become one of the strongest names in enterprise AI. Both can be true. In fact, that may be the most likely future: not one winner, but several giants dividing the market by use case.

For businesses, the lesson is simple. Do not ask only which AI company is biggest. Ask which AI tool solves the problem in front of you. The answer may be OpenAI. Increasingly often, it may be Anthropic. And next quarter, because this industry apparently refuses to move at a normal human speed, it may be someone else entirely.