Anthropic is heading toward what could become one of the most closely watched technology IPOs in history, but investors are already asking a difficult question:
Can the AI company maintain its extraordinary growth once it becomes a public company?
The developer of Claude has seen its annualized revenue run rate explode to about $65 billion, up from $47 billion in May and roughly $9 billion at the end of 2025. Anthropic’s investors expect revenue to exceed $120 billion by the end of 2026, according to the Financial Times.
Yet the company’s potential Wall Street debut is arriving at a moment when the AI market is becoming dramatically more competitive.
OpenAI is regaining ground. Chinese and open-weight AI models are becoming cheaper and increasingly capable. Businesses are gaining tools that allow them to switch between competing AI models more easily.
And that creates the central question surrounding Anthropic’s IPO:
Is its extraordinary revenue growth the beginning of a durable AI business — or a peak that investors may struggle to sustain?
Anthropic’s numbers are moving at extraordinary speed
Anthropic’s growth has been remarkable even by the standards of the AI industry.
The company said in May that its annualized revenue run rate had passed $47 billion. By the end of July, that figure had risen to approximately $65 billion, according to people familiar with the company’s financial performance cited by Bloomberg and subsequently reported by other outlets.
The distinction between revenue and revenue run rate is important.
A $65 billion annualized run rate does not mean Anthropic actually generated $65 billion in revenue during the year. It extrapolates recent performance over a full year.
Nevertheless, the acceleration has attracted enormous investor attention.
Anthropic’s May funding round valued the company at approximately $965 billion post-money after it raised $65 billion.
The company has also confidentially filed paperwork for a potential IPO, positioning itself among the first major frontier-AI companies preparing to enter the public markets. OpenAI has separately filed confidential IPO paperwork.
The valuation question is getting enormous
The most striking part of the IPO story is not simply that Anthropic wants to go public.
It is the potential size of the company when it does.
The Financial Times reported that investors familiar with the company believe potential post-IPO valuations could range from approximately $1.5 trillion to $4 trillion — an extraordinarily wide range that itself illustrates how difficult it is to value a company growing at such a rapid pace.
That range spans an enormous amount of potential market capitalization.
It also tells investors something important:
The debate is no longer primarily about whether Anthropic is growing. It is about how long that growth can continue.
OpenAI is fighting back
For much of the recent AI race, Anthropic’s Claude models have gained substantial traction with businesses, particularly in areas such as coding and enterprise applications.
But OpenAI has recently regained momentum.
The Financial Times reported that OpenAI began capturing a larger share of customer spending following the release of GPT-5.6, overtaking Anthropic in weekly spending on OpenRouter for the first time in more than two and a half years.
OpenAI has also introduced its Astra model, which the company says represents another advance in its model capabilities.
That creates a rapidly changing competitive landscape.
Anthropic’s revenue growth cannot simply be projected forward on the assumption that its current market position will remain unchanged.
Its competitors are improving at the same time.
Cheaper AI models are creating another threat
The competitive challenge isn’t limited to OpenAI.
The next pressure point could come from open-weight and lower-cost AI models.
Companies such as China’s DeepSeek and Moonshot, along with Meta, are developing models whose parameters can be made available for users and developers to customize.
Some of these models are increasingly approaching the capabilities of more expensive proprietary systems, according to the Financial Times.
That creates a potentially uncomfortable situation for Anthropic and other frontier AI companies.
If customers can obtain sufficiently capable AI from cheaper providers, they may become less willing to pay premium prices for the most expensive models.
And the easier it becomes to switch between models, the harder it becomes for one AI company to lock in customers.
Companies are already learning to switch models
One of the most revealing examples comes from Ramp, a corporate payments company.
According to the Financial Times, Ramp and some of its customers have begun using systems that route workloads between different AI models depending on the task.
That can make it easier for companies to choose cheaper models for certain workloads and reserve more expensive frontier models for tasks that require greater capability.
The result has already been significant for some users.
Ramp’s CEO said the company had reduced its AI spending by about 40% through this approach.
That does not necessarily mean overall AI demand is shrinking.
It could instead mean that customers are becoming more price-sensitive and more sophisticated about how they consume AI.
That distinction could become crucial for Anthropic.
The industry may be entering a price war
For years, the AI competition was dominated by a relatively simple proposition:
Build the most capable model and attract the customers.
Now the equation is becoming more complicated.
Companies have to compete on:
- Model capability
- Price
- Reliability
- Enterprise security
- Speed
- Computing efficiency
- Customer retention
- Specialized applications
- Distribution
The Financial Times quoted Ramp’s leadership describing the current environment as the first genuine period of price competition between major AI labs.
If that trend accelerates, the economics of the AI industry could change dramatically.
Rapidly increasing usage would still be valuable, but revenue growth could become less impressive if prices fall faster than demand rises.
Anthropic does have an important advantage
The concerns surrounding Anthropic’s IPO do not mean the company lacks competitive strengths.
Its enterprise adoption has been significant.
Ramp’s own AI Index showed Anthropic accounting for 42.4% of businesses using AI in June, compared with 39.5% for OpenAI in that particular dataset. The data also showed that adoption of open-source and Chinese model-serving platforms was increasing but remained much smaller.
Another analysis cited by the Financial Times found that Anthropic had retained a larger share of users after 12 months than OpenAI.
Approximately 22.5% of Anthropic users remained active after a year, compared with around 13.2% for OpenAI in the cited analysis.
That suggests Anthropic has developed meaningful customer loyalty even as the market becomes more competitive.
But retention alone does not settle the question of long-term profitability.
The biggest challenge may be turning growth into durable economics
AI companies face an unusual business-model problem.
Their products can attract enormous demand, but operating advanced models requires huge amounts of computing power.
Training and serving increasingly capable AI systems requires chips, data centers, electricity and networking infrastructure.
That means revenue can grow rapidly while costs also rise dramatically.
Anthropic’s $65 billion annualized revenue figure therefore needs to be considered alongside the enormous capital requirements of remaining at the frontier.
The company raised $65 billion in its latest funding round partly to expand compute capacity and continue its research and product development. Anthropic said the funding would also support safety and interpretability research.
For public investors, the critical question will ultimately be less about raw revenue and more about how much profit can be generated from that revenue after computing and infrastructure costs.
AI’s economics are still being figured out
The uncertainty extends across the entire industry.
OpenAI and Anthropic are generating revenue at unprecedented rates, while investors are simultaneously committing extraordinary amounts of capital to build computing infrastructure.
Reuters has reported that OpenAI is seeking additional funding at a valuation around $1.2 trillion, while the company faces enormous projected spending requirements as it expands its AI infrastructure.
That makes Anthropic’s IPO an important test for the entire AI investment cycle.
Public-market investors will eventually have access to financial disclosures that private investors have largely seen only through funding rounds and company updates.
The market will then have to determine how much of the AI industry’s extraordinary growth is sustainable.
The safety debate adds another complication
Anthropic has also positioned itself differently from some competitors on AI safety.
CEO Dario Amodei recently called for AI companies to slow the development of cutting-edge systems to give researchers more time to establish safeguards.
That creates a complicated business tension.
Slowing the development of the most advanced systems could potentially reduce training expenditures and allow additional safety work.
But if competitors continue moving rapidly, Anthropic could risk giving rivals additional time to catch up. The Financial Times reported that investors are weighing precisely this tension.
The debate is therefore not simply technological.
It is becoming a question of strategy, economics and shareholder expectations.
China’s AI industry is another variable
Anthropic also faces competition from China.
Chinese companies including DeepSeek and Moonshot are developing lower-cost models that are becoming increasingly competitive.
Ramp’s data suggests adoption of Chinese and open-source models among U.S. businesses remains relatively limited, but the share of companies using model-serving platforms increased from 4.5% in January to 5.8% in June.
Those figures do not demonstrate that Chinese models have displaced Anthropic or OpenAI.
In fact, Ramp found that most businesses using model-serving platforms also continued to use OpenAI or Anthropic.
But the direction of travel matters.
Customers now have more choices.
And more choices generally give buyers greater leverage over price.
Anthropic’s IPO could become a referendum on the AI business model
The potential listing comes at a fascinating moment.
Anthropic is growing extraordinarily quickly.
OpenAI is fighting back.
Chinese and open models are putting pressure on prices.
Enterprise customers are gaining more flexibility.
And investors are being asked to value businesses whose future revenues depend partly on a technology whose ultimate capabilities and economics remain uncertain.
That explains why valuation estimates are so wide.
A $1.5 trillion valuation implies one set of assumptions.
A $4 trillion valuation implies something dramatically different.
The difference ultimately comes down to expectations about growth, pricing power, customer retention, competition and profitability.
What investors will be watching
If Anthropic proceeds with an IPO, investors will likely focus heavily on several questions:
Can the $65 billion annualized revenue run rate continue climbing?
How much of the projected $120 billion year-end figure becomes actual recurring revenue?
How much pricing pressure will come from OpenAI and lower-cost models?
How expensive will it be to serve increasingly capable AI systems?
How concentrated is Anthropic’s revenue among its largest customers and cloud partners?
Can the company maintain enterprise customer loyalty as model switching becomes easier?
How will Anthropic balance its safety position with the competitive pressure to move quickly?
Those questions will determine whether today’s extraordinary growth becomes a durable public-company story.
The IPO may be the biggest test yet
Anthropic has already achieved something remarkable: it has gone from being viewed as a challenger in the AI race to a private company valued at nearly $1 trillion, with an annualized revenue run rate of approximately $65 billion.
But public markets operate under a different set of expectations.
Private investors can fund enormous growth based on long-term potential.
Public investors have to continually reassess revenue, margins, competition and cash requirements.
That makes Anthropic’s potential Wall Street debut more than another technology IPO.
It could become one of the clearest tests yet of whether the economics of frontier AI can support the extraordinary valuations now being placed on the industry’s leading companies.
Anthropic has built a $65 billion revenue machine at breathtaking speed. Now investors want to know whether that machine can keep accelerating — or whether the arrival of OpenAI, cheaper AI models and tougher price competition will finally force the industry’s growth story to slow down.
