OpenAI's investors are doing the math on a $1.2 trillion IPO bet, and some are wondering if they backed the wrong horse.
The Summary
- An investor who backed both companies said justifying OpenAI's latest round requires assuming a $1.2 trillion+ IPO valuation, making Anthropic's $380 billion valuation look like a bargain
- The safety-first AI lab is splitting its cap table: some investors are unhappy with Anthropic's Pentagon fallout, while others see the path to profitability getting clearer
- Anthropic positioned itself as the AI safety company, but politics and defense contracts are testing whether investors actually want that
The Signal
The valuation gap tells you everything about how quickly the AI race reshuffled the deck. OpenAI's recent funding round requires betting on a $1.2 trillion IPO to make the numbers work. Anthropic sits at $380 billion. That's a 3x difference for companies building nearly identical products with nearly identical model capabilities.
One investor put money in both. Now they're looking at the spread and reconsidering which bet pencils out. This isn't about model benchmarks or whose chatbot is smarter. It's about whether you're paying $1.20 for a dollar of future earnings or $0.38.
"Making Anthropic's current $380 billion valuation look like the relative bargain."
But Anthropic's discount comes with complications. The company built its brand on AI safety and responsible development. That brand is now colliding with the realities of government contracts and defense spending. Fortune reports investors are divided over Anthropic's dispute with the Pentagon, with some unhappy the company positioned itself as safety-first and is now dealing with White House tension.
This isn't a small split. These are investors who wrote checks expecting either a values-aligned company that wouldn't chase certain revenue, or a company that would ultimately chase all revenue like everyone else. Both groups can't be right. The safety positioning attracted capital from people who wanted an alternative to OpenAI's "move fast and commercialize everything" approach. Now some of those same investors are frustrated when Anthropic actually behaves differently.
Key investor tensions:
- Cap table includes both mission-aligned funds and growth-stage firms expecting defense contracts
- Pentagon fallout creates near-term revenue uncertainty while OpenAI signs government deals
- Lower valuation only matters if exit multiples hold, and government disputes complicate that path
The math gets interesting when you layer in government relationships. OpenAI has no public Pentagon problems. They take the contracts, ship the models, keep the revenue engine running. Anthropic's safety stance meant saying no to some of that money, or at least moving slower. Investors who wanted the safety angle are fine with that. Investors who wanted the $380 billion to grow into $1 trillion are not.
The Implication
Watch the next 12 months of Anthropic's enterprise deals. If they close big contracts outside of defense and government, the valuation gap tightens and the safety positioning becomes a differentiation advantage. If they struggle to hit revenue targets while OpenAI signs everything in sight, those second-guessing investors start looking prescient.
For builders and workers watching from outside the cap table, this is the agent economy's first real values test at scale. The companies building the infrastructure for Web4 are making trade-offs between speed, safety, and which customers they'll serve. Those choices will shape what kinds of agents get built and who gets to use them.