The Story
OpenAI and Anthropic may be heading toward the toughest phase of their rise, as expectations build around possible public listings at valuations near $1 trillion each. But a Financial Times report argues that convincing public market investors will be far harder than winning over private investors.
For years, venture investors have bet heavily on the future of artificial intelligence. Public market investors, however, will ask a simpler question: can these companies generate sustainable profits?
Details
- Costs are outrunning revenue: Even as generative AI revenue grows quickly, frontier model development requires enormous spending on data centers, chips, and energy. JPMorgan estimates that AI infrastructure spending could reach $5 trillion over the next five years.
- Model builders capture only part of the value: According to Exponential View, the generative AI sector generated about $110 billion in revenue over the past year, but foundation model companies captured only around 11% of that revenue, while most of it went to hosting and infrastructure providers.
- Allies are becoming rivals: The companies that backed OpenAI and Anthropic, including Microsoft, Alphabet, and Amazon, are increasingly building their own AI models and investing tens of billions of dollars in AI infrastructure, turning from partners into direct competitors.
- The edge is narrowing: As open-source models spread and cheaper Chinese models improve, companies have more options for routine AI tasks, putting pressure on the margins of the most advanced frontier models.
- Different strategies: The report says Anthropic currently has an edge with business customers, especially in real-world uses such as law and science, but faces political and regulatory challenges. OpenAI, meanwhile, is trying to strengthen ties with the Trump administration, including by offering the U.S. government a 5% equity stake.
Why it matters
The concern is not weak demand for AI. It is the gap between the scale of investment and the revenue generated so far. Companies are spending hundreds of billions of dollars on infrastructure, while revenues remain far below that level, raising questions about when the sector can become truly profitable.
Veteran investor Jeremy Grantham says the U.S. market is at historically expensive levels, while the Bank for International Settlements has warned that AI “exuberance” could end in a painful investment bust.
Background
Many analysts compare the current AI boom to the dotcom bubble of 2000. That era eventually produced giants such as Amazon and Google, but the Nasdaq still fell about 77% from peak to trough, and around 4,800 internet companieswent bust by 2003. The question is no longer whether AI will change the world, but which companies will still be standing when the hype fades and the profit test begins.
