Anthropic has attracted funding offers at $800 billion or higher, just two months after closing a $30 billion round at $380 billion, and the speed of that jump tells you more about private market psychology than it does about any particular product launch.
In September 2025, Anthropic raised $13 billion at a $183 billion post-money valuation. Five months later, it closed a $30 billion Series G led by GIC and Coatue at $380 billion. Two months after that, Bloomberg reported that investors were offering terms that would value the company above $800 billion. And by late April, secondary market platforms were showing transactions implying a valuation of roughly $1 trillion, with one trader telling Business Insider he had received an offer at $960 billion before it was snapped up by someone else. The company itself has reportedly resisted most of the recent funding overtures, at least for now. That almost does not matter. The bidding war around Anthropic has become its own story, separate from the revenue or the products, and it is the kind of story that tends to end one of two ways.
There is real substance underneath the numbers. Anthropic’s annualized revenue climbed from $9 billion at the end of 2025 to roughly $30 billion by early April 2026. AWS CEO Andy Jassy confirmed this week that his company holds a meaningful equity stake and disclosed $16.8 billion in pre-tax paper gains from that position in a single quarter. OpenAI, Anthropic’s most direct competitor, is trading at a slight discount to Anthropic on secondary markets despite being valued at $852 billion in its own most recent primary round. That is a significant inversion. It tells you something about how investors are currently reading the competitive landscape, even if the actual product gap between Claude and GPT is not as large as the valuation gap implies.
The company’s underlying business is genuinely strong. Anthropic is the dominant model provider for enterprise AI and coding workflows. Its enterprise API revenue is growing rapidly. Its partnership structure with Amazon includes the Project Rainier cluster, a dedicated training environment built on Trainium2 chips, which gives Anthropic compute access that most of its rivals have to bid for in open markets. The Claude model family has been well-received by engineering teams and developers who prize reliability and instruction-following over raw benchmark performance. These are real moats, not manufactured ones. But moats do not explain a valuation trajectory that has added roughly $800 billion in estimated enterprise value in under eight months.
The more honest interpretation is that investors are not valuing what Anthropic has built. They are pricing what they think Anthropic will become. The thesis goes something like this: frontier model companies will function as platform layers for the next generation of enterprise software, the way AWS became the platform layer for the last generation. If that is true, then a company with Anthropic’s model quality, enterprise reach and infrastructure lock-in could plausibly command a valuation that looks outrageous at current revenue multiples but reasonable on a ten-year forward basis. The bull case is not about the $30 billion annualized revenue number. It is about the software and agent layers that sit on top of Claude as the economy automates.
There is a version of that argument that makes sense. Anthropic’s coding capabilities in particular are attracting the kind of enterprise commitment that tends to produce long-term retention. A company that rebuilds its internal development workflow around Claude agents is not going to switch easily. The switching costs accumulate over time, in custom prompts, in fine-tuned models, in integrations and in the organizational muscle memory that forms around a specific tool. Anthropic is trying to get inside as many of those workflows as possible before the market consolidates. If it succeeds, the revenue base that follows could be very large and very sticky.
But the pace of valuation expansion is still unusual enough to warrant attention. The company raised at $183 billion, then $380 billion, then received offers at $800 billion, and now secondary buyers are paying implied prices near $1 trillion, all within about eight months. Revenue grew strongly over that period, but not nearly fast enough to justify the valuation growth in traditional terms. What is happening is that private capital markets have essentially lost their reference points. When a company triples in estimated value while its revenue doubles, and both events happen in less than a year, the price is no longer primarily a function of discounted future cash flows. It is a function of how badly the next buyer wants exposure to the AI cycle before the window closes.
The Bubble Test Moves Private
That is the more uncomfortable version of the story, because it raises a question that public market investors will eventually have to answer. If Anthropic pursues an IPO in late 2026 or 2027, as has been reported, it will arrive with a valuation that has been set by a small group of sophisticated private buyers who are motivated by both genuine conviction and competitive fear of missing out. The public market will then have to decide whether that valuation makes sense for a much larger and more diverse pool of investors. That test will be harder to pass if the valuation keeps climbing before the IPO rather than settling. The higher it goes in private markets, the more pressure there is on the public debut to justify a number that was set when access was scarce and urgency was high.
This is not a new dynamic in technology. It happened with Uber, with WeWork and with dozens of other companies that arrived at IPO with private valuations that had been pushed up by competition for allocation rather than fundamental analysis. Most of those stories ended badly for late-stage private investors and early public ones. A few of them ended with companies that did eventually grow into their valuations. The question for Anthropic is which category it belongs in.
The honest answer is that nobody knows yet, and the current valuation pace makes that uncertainty more consequential, not less. What is clear is that the bubble test for AI is moving into private markets rather than waiting for IPOs. The stress fractures, if they come, may show up in secondary market liquidity, in investor concentration and in the gap between paper gains and cash returns, before any ticker symbol exists. For the founders, the employees and the investors already inside the cap table, that is the number worth watching as much as the revenue growth. The price keeps moving. The question is whether it is moving toward something or away from it.
Also read: AWS just posted its fastest growth in four years and the Anthropic bet is paying off • AWS just posted its fastest growth in four years and the Anthropic bet is paying off • AWS is growing fastest when it needs to most and Anthropic is why