Anthropic pitching investors on "$30 trillion in potential revenue" is the kind of slide deck that belongs in a museum of startup delusion, not a serious fundraising round.
This isn't a forecast. It's a SAM (Serviceable Addressable Market) fantasy — the kind where you assume you capture 100% of all economic activity if everything goes perfectly. Every AI lab has one. None of them are remotely defensible.
Meanwhile, $GOOGL and $MSFT have actual revenue, distribution, enterprise customers, and decades of infrastructure. Anthropic has a talented technical team and a balance sheet that burns cash at an alarming rate.
The gap between "we could theoretically dominate AI" and "we actually will" has widened, not narrowed — especially as capex requirements for training, inference, and scaling have become painfully clear.
This is the moment VCs start learning the hard way: there's a massive difference between funding a research org and backing a durable, profitable business. The AI hype cycle is about to meet the reality of unit economics, competitive moats, and actual margin profiles.
This isn't a forecast. It's a SAM (Serviceable Addressable Market) fantasy — the kind where you assume you capture 100% of all economic activity if everything goes perfectly. Every AI lab has one. None of them are remotely defensible.
Meanwhile, $GOOGL and $MSFT have actual revenue, distribution, enterprise customers, and decades of infrastructure. Anthropic has a talented technical team and a balance sheet that burns cash at an alarming rate.
The gap between "we could theoretically dominate AI" and "we actually will" has widened, not narrowed — especially as capex requirements for training, inference, and scaling have become painfully clear.
This is the moment VCs start learning the hard way: there's a massive difference between funding a research org and backing a durable, profitable business. The AI hype cycle is about to meet the reality of unit economics, competitive moats, and actual margin profiles.