Anthropic’s revenue growth is accelerating at a pace that is reshaping expectations for the artificial intelligence industry, with the Claude maker’s annualized revenue run rate surpassing $65 billion by the end of July, according to people familiar with the company’s financial performance cited by Bloomberg.

The latest figure represents a dramatic increase from roughly $47 billion in May and about $9 billion at the end of 2025. The acceleration is striking because Anthropic is now generating the reported revenue at a scale where sustaining the same percentage growth becomes progressively more difficult.

Investors are betting that the momentum will continue. The Financial Times reported that Anthropic investors expect the company to reach an annualized revenue run rate of between $100 billion and $120 billion by the end of 2026.

That would put Anthropic’s projected run rate at roughly 11 to 13 times its level at the end of last year, underscoring how quickly demand for advanced AI systems is translating into commercial sales. The figures also put Anthropic ahead of rival OpenAI on the latest reported annualized revenue figures. OpenAI’s annualized revenue has reached about $40 billion, according to recent reports.

The comparison needs some caution because the companies may calculate and report their revenue metrics differently. More importantly, an annualized run rate is not the same as revenue actually generated over the preceding 12 months. It extrapolates recent performance over a full year, meaning it can rise sharply when sales accelerate and can fall just as quickly if growth slows.

Even with that qualification, Anthropic’s trajectory is significant.

Enterprise AI Is Becoming A Major Revenue Engine

Anthropic’s growth underpins that the AI industry’s leading model developers are moving beyond consumer chatbots toward enterprise applications where companies are willing to pay substantially more for advanced models, coding tools, agents, and other AI services. This matters because enterprise customers can generate recurring revenue at a scale that is difficult to achieve through individual consumer subscriptions alone.

Anthropic has increasingly positioned Claude as a business and developer tool, particularly for coding and other complex professional workloads. As companies integrate AI into software development, customer service, research and internal operations, model usage can become embedded in day-to-day business processes rather than remaining an occasional productivity tool.

That creates the possibility of a much larger and more durable revenue base, provided customers continue increasing their AI usage.

But the revenue surge is also raising the stakes for Anthropic’s expected public offering.

Both Anthropic and OpenAI have filed confidential paperwork for potential IPOs. Anthropic is widely expected to reach the public markets first, potentially as soon as this autumn. Investors have told the Financial Times they expect Anthropic to seek a valuation of $2 trillion or more, potentially making it the largest IPO ever.

That would represent an extraordinary increase from Anthropic’s latest private valuation of $96.5 billion following its May funding round.

A $2 trillion valuation would mean investors were valuing Anthropic at more than 30 times its current $65 billion annualized revenue run rate. If the company reaches the projected $100 billion to $120 billion run rate by year-end, that multiple would fall to roughly 17 to 20 times revenue.

That distinction is important. A $2 trillion valuation becomes considerably easier to justify if investors believe the $100 billion-plus figure is not a peak but another step in a much larger revenue curve.

The public market will ultimately force investors to scrutinize that assumption.

Growth Is Only One Side of The Equation

Anthropic’s revenue figures are impressive, but they do not establish profitability.

Frontier AI companies face unusually high costs because serving advanced models requires large amounts of computing power. The more customers use Claude, the more Anthropic must spend on chips, data centers, networking, and electricity unless improvements in model efficiency and pricing allow revenue to grow faster than computing costs.

This creates a critical distinction between revenue growth and economic efficiency.

An AI company can increase revenue rapidly while still consuming enormous amounts of capital. The next stage of the market will therefore be judged increasingly on gross margins, cash burn, infrastructure commitments and the cost of serving each additional unit of AI usage.

That is likely to become one of the most important issues surrounding Anthropic’s IPO. Investors will want to know whether the company’s extraordinary revenue growth is accompanied by improving economics or whether rising sales are requiring a proportionate increase in infrastructure spending.

Anthropic’s reported lead in annualized revenue also changes the dynamics of its competition with OpenAI.

OpenAI remains one of the most valuable AI companies in the world, but Anthropic’s faster reported growth gives it a stronger position going into the public markets. The two companies are competing for enterprise customers, developers, computing capacity, and long-term relationships with major technology companies.

The rivalry also extends beyond the models themselves. Both companies need enormous amounts of computing infrastructure to serve customers. Their ability to secure chips and data-center capacity could become as important as model performance as AI adoption expands.

That creates an unusual capital cycle in which rising AI demand generates more revenue, which supports greater infrastructure investment, which allows companies to serve more customers and generate still more demand.

The sustainability of that cycle will be a central issue for public-market investors.

Anthropic’s potential listing is bigger than a single company’s debut. This is because a $2 trillion valuation would provide public investors with one of the clearest opportunities yet to determine how much they are willing to pay for a leading frontier AI company. The pricing of the IPO is expected to influence valuations across the AI ecosystem, from model developers and cloud providers to chipmakers and data-center operators.

A strong debut could reinforce the view that AI has developed into a massive commercial market capable of supporting technology companies at unprecedented valuations. But a weaker reception could have the opposite effect, particularly if investors conclude that the revenue growth of frontier AI companies does not justify the capital required to produce it.

Some analysts believe that’s what makes Anthropic’s accelerating revenue particularly important. The company is not entering the public markets with a conventional startup growth story. It is approaching an IPO with reported revenue at a scale that would place it among the world’s largest technology businesses if converted into actual annual revenue.

The challenge now is proving that the extraordinary growth is durable. Anthropic’s reported rise from a $9 billion annualized run rate at the end of 2025 to more than $65 billion seven months later demonstrates the explosive commercial demand for Claude and advanced AI services. But the next $65 billion will be considerably harder to generate than the first.

Anthropic’s plan to introduce an invisible watermark into text generated by its Claude AI models is facing growing criticism, with tech blogger John Gruber saying that the technology could compromise one of the most fundamental functions of a language model: choosing the best possible words for a user.

Gruber, who writes the influential Daring Fireball blog, criticized the proposal on Sunday, describing it as a “perversion of writing.” His objection goes beyond whether users can detect the watermark. He argues that any system that changes how Claude selects words to make its output identifiable is allowing an objective other than the user’s needs to influence the writing.

“The exact words we choose when writing matter,” Gruber wrote. “I want any LLM I use to choose the very best, most precise words at every single decision point.”

The move by Anthropic was necessitated by growing tension facing AI developers as governments seek reliable ways to identify machine-generated content. Developers are being pushed to make AI outputs traceable while also trying to ensure that those mechanisms do not degrade the quality of generated text.

Anthropic has sought to address those concerns by arguing that its proposed watermark would be effectively invisible and would not alter the quality of Claude’s writing.

The company initially described the feature as an “imperceptible watermark” embedded directly into Claude’s output. Anthropic has linked the technology to compliance with the European Union’s AI Act, which requires certain AI-generated content to be marked in a machine-readable manner.

The company said other major AI developers will face similar requirements and will need mechanisms capable of identifying AI-generated material.

Anthropic provided more detail about its approach in a blog post published Friday, saying the watermark would not involve hidden characters, unusual formatting or other visible alterations that could affect how text appears to readers. Instead, the system would make subtle changes to Claude’s word-selection process to produce a statistical signature that can later be detected.

Anthropic says that the changes would occur among words that are essentially interchangeable in a particular context.

For example, the company said that after generating a sentence such as “The weather today was cold and,” Claude might choose “gray” or “overcast.” Both words convey essentially the same meaning, allowing the model to alter its statistical pattern without materially changing the sentence.

That explanation, however, does not resolve Gruber’s central objection.

From his perspective, the issue is not whether readers can notice the difference or whether Anthropic can demonstrate that the resulting sentence remains understandable. The issue is whether watermarking introduces a competing objective into the model’s generation process.

“Within the constraint of executing inference quickly, and at a certain cost per token, I want the best words,” Gruber wrote. “The idea that anything other than my needs should factor into the generation of text for me is patently offensive.”

That criticism goes to the heart of how AI-generated writing should be evaluated. A language model normally optimizes its output according to a combination of factors such as relevance, coherence, accuracy, and the user’s instructions. Watermarking adds another consideration: whether a sequence of word choices contributes to a detectable statistical pattern.

Anthropic maintains that the trade-off is negligible because the watermarking system operates primarily when multiple words would produce effectively equivalent prose. Critics such as Gruber question whether an algorithm can reliably determine that two words are interchangeable in every context, particularly in writing where tone, rhythm, precision and connotation matter.

The distinction could become more important in professional writing. A journalist, lawyer, researcher or author may care about subtle differences between words that appear interchangeable to a statistical model. Even when two words communicate roughly the same idea, one may be more precise, natural, or appropriate for a particular audience.

The controversy has already generated concern among some Claude users, with reports of users cancelling subscriptions over the watermarking plan. Anthropic, however, told Business Insider that it had not seen a spike in cancellations.

The debate also exposes a difficult problem for the AI industry: machine-readable identification is easier to mandate than it is to implement without affecting the underlying technology.

Watermarking generated text could help publishers, educators and other institutions distinguish AI-assisted material from human writing. It could also make it easier to investigate the provenance of content circulating online, particularly as AI-generated material becomes more difficult to distinguish from human-produced work.

But the technology must contend with an important limitation. A watermark that depends on subtle word-selection patterns may become less reliable after text is edited, translated, paraphrased, or rewritten by another AI system. That means the practical value of watermarking may depend as much on how robust the detection mechanism is after content leaves Claude as on how accurately it works on untouched output.

For Anthropic, the challenge is therefore to satisfy emerging regulatory requirements without undermining the quality that users expect from Claude.

Gruber’s criticism is based on a fundamental concern: if watermarking influences even marginal word choices, users may reasonably ask whether the text is being optimized entirely for them or partly for the needs of the platform and regulators.