1. Headline & intro
Anthropic’s latest growth spurt isn’t just a good quarter; it’s a stress test of the entire AI hype cycle. When a supposed "number two" in foundation models suddenly looks like the better value, investors are forced to confront a question they’ve been avoiding: what is a trillion‑dollar AI company actually worth? In this piece, we’ll unpack why some OpenAI backers are quietly shifting allegiance, what Anthropic’s momentum reveals about the business models behind large models, and how this reshapes risk for enterprises, startups, and regulators on both sides of the Atlantic.
2. The news in brief
According to TechCrunch, citing reporting from the Financial Times, OpenAI’s $852 billion private valuation is starting to attract skepticism from parts of its own cap table. The company is said to be reorienting around enterprise customers while facing intensifying competition from Anthropic.
Anthropic’s annualized revenue reportedly jumped from $9 billion at the end of 2025 to $30 billion by the end of March 2026, driven largely by demand for its coding tools. That puts its current private valuation of around $380 billion at a much lower revenue multiple than OpenAI’s.
One investor backing both companies told the FT that justifying OpenAI’s last round requires believing in a future IPO above $1.2 trillion. In contrast, Anthropic is framed as the "relative bargain." Secondary-market trading seems to support that view: demand for Anthropic shares is described as intense, while OpenAI stock is selling at a discount. OpenAI CFO Sarah Friar, however, pointed to the company’s record $122 billion raise as proof that investor confidence remains high.
3. Why this matters
This story is not really about OpenAI versus Anthropic. It’s about whether the economics of foundation models can support the kind of valuations usually reserved for dominant operating systems or global payment networks.
Anthropic’s $30 billion annualized revenue against a $380 billion valuation implies a high but at least recognizable growth‑tech multiple. OpenAI’s $852 billion valuation, by contrast, seems to assume not only that it wins the model race, but that it becomes the default AI infrastructure layer for the global economy – and then defends that position against both Big Tech and open source for years.
Some investors are deciding that is simply too much platform risk for the price. Anthropic looks, in comparison, like a concentrated bet on one of the most capable teams in the field, at a discount to the market’s current OpenAI narrative.
The product mix also matters. Anthropic’s recent surge is reportedly powered by coding tools – an area where value can be realized quickly and measured clearly. Development teams either ship faster and with fewer bugs, or they don’t. That makes it easier for CIOs to justify large contracts and for investors to underwrite growth.
OpenAI, meanwhile, has spent years as the consumer-facing brand of the AI revolution. Enterprise revenue is now the focus, but that requires different sales motions, longer procurement cycles, and deep integration with existing IT stacks. In other words: a slower, messier path to the kind of predictable revenue streams that justify a near‑trillion valuation.
The immediate implication: capital may start pricing a duopoly rather than a runaway winner. That could make life more expensive for OpenAI when it comes to secondary sales or future rounds – and more comfortable for Anthropic as it pitches itself as the more reasonably priced leg of the AI barbell trade.
4. The bigger picture
Underneath the OpenAI–Anthropic rivalry is a broader transition: the market is moving from a "wow, this is magic" phase to a "show me the unit economics" phase.
In the early ChatGPT era, investors and enterprises were willing to believe that whoever had the largest and most advanced model would automatically win outsized profits. That fit the mental model of search, social, or mobile operating systems: one or two players dominate, network effects do the rest.
But AI infrastructure behaves differently. Models rapidly diffuse through papers, open‑source releases, and talent poaching. Cloud providers commoditize access via APIs. Enterprises hedge by using several providers at once. And regulators are increasingly uncomfortable with single‑vendor dependencies.
We’re also seeing a more complex competitive map. Google DeepMind, Meta, xAI and others are not sitting still. Cloud giants like Microsoft, Amazon, and Google have every incentive to keep model providers in check so that value doesn’t concentrate entirely in one layer of the stack. Meanwhile, open‑weight models and tools – from Llama‑style architectures to European initiatives – are eroding the moat around proprietary giants.
Against that backdrop, the idea that one independent lab can be worth close to a trillion dollars looks less like rational pricing and more like a derivative of zero‑interest‑rate thinking lingering in a higher‑rate world.
Anthropic’s rise therefore acts as a correction mechanism. It reminds the market that even in AI, competition exists, switching costs are not absolute, and pricing power can be challenged. It also demonstrates that there are multiple viable paths to big AI revenues: not just chatbots, but deeply embedded B2B workflows such as code generation, security tooling, and verticalized assistants.
5. The European and regional angle
For European companies, the shift in investor sentiment is more than Silicon Valley gossip – it directly affects procurement, compliance, and strategic risk.
CIOs in Frankfurt, Paris or Madrid are already juggling several constraints: the EU AI Act, GDPR, sector‑specific rules in finance and health, and increasing pressure to reduce dependence on a handful of US tech giants. The prospect of an AI stack dominated by a single, ultra‑expensive US vendor has always been politically and operationally uncomfortable.
The emergence of a stronger Anthropic – alongside European players like Mistral AI and a growing open‑source ecosystem – gives buyers more leverage. If OpenAI is priced and perceived as an inevitable monopoly, regulators will feel compelled to treat it like one, with correspondingly heavy oversight under tools like the Digital Markets Act. A more balanced duopoly, plus credible European alternatives, is easier to square with the EU’s competition doctrine.
There’s also a capital‑flows angle. If global investors start to accept that no single US lab can capture all AI value, appetite for regional champions improves. That benefits not only high‑profile model startups in Paris or Berlin, but also smaller applied‑AI companies in places like Ljubljana, Zagreb or Barcelona that build on top of multiple model providers.
Finally, European corporates tend to be more risk‑averse and price‑sensitive than their US peers. Anthropic’s positioning as the "better‑priced" hyperscaler‑agnostic alternative may resonate strongly with banks, manufacturers and public‑sector bodies trying to build AI capabilities without locking themselves into one US vendor at any price.
6. Looking ahead
The most likely outcome over the next 12–24 months is not that OpenAI collapses or Anthropic "wins" outright, but that valuations converge toward something compatible with real enterprise software economics.
Expect three developments:
Harder questions from boards and CFOs. Enterprise buyers will increasingly ask whether they truly need to standardize on a single model provider. Multi‑model strategies – mixing Anthropic, OpenAI, cloud‑native and open‑source models – will become the default for large organizations.
Pressure on margins and pricing. As Anthropic and others compete aggressively in coding tools and enterprise platforms, per‑token and per‑seat prices will face downward pressure. To defend valuations, providers will need to move up the stack into higher‑margin offerings: agents, workflow automation, vertical solutions.
Regulatory differentiation. Europe in particular will reward vendors who can demonstrate transparent training data, controllable behavior, and strong governance. If one of the big US labs stumbles on safety or compliance, that will immediately reshape enterprise demand in regulated sectors.
Investors, meanwhile, will keep using the secondary market as a reality check. If OpenAI stock continues to trade at a discount to its last round while Anthropic clears at or above its stated valuation, boards will be forced to revisit their assumptions about who is really "number one" – and whether that ranking even matters in a multi‑model world.
7. The bottom line
Anthropic’s ascent is less a coup and more a correction: a signal that the market is no longer willing to pay any price for the OpenAI story. For enterprises and regulators, a more balanced AI landscape is healthy, increasing choice and reducing single‑vendor risk. For investors, it’s a reminder that even in a transformative technology wave, entry price still matters. The real question now is not which lab wears the crown, but which business models around AI can compound profitably once the hype premium finally burns off.
