AI could turn out to be both the greatest technological revolution of the past few decades and one of the biggest investment bubbles in history.
Yesterday, Anthropic reported a net loss of $80.7 billion in just the first half of the year. To be fair, there’s no need to panic over the headline number itself. Around $72 billion of that loss came from an accounting revaluation of the company’s liabilities, not from Anthropic actually burning through that much cash. But something else is far more interesting.
The company has already committed hundreds of billions of dollars to computing capacity, chips, and data centers. And Anthropic is far from alone. Amazon, Microsoft, Google, and Meta are set to invest more than $700 billion in AI infrastructure this year. The scale has become so enormous that these investments are now supporting the U.S. stock market and accounting for a significant share of investment activity across the economy.
And this is where the U.S. is gradually falling into a very interesting trap. If tech companies keep spending at this pace, they’ll have to borrow more and more money and somehow prove that these massive investments will actually pay off. If they start cutting spending, the fallout will hit chipmakers, data centers, the energy sector, the tech companies themselves, and ultimately the stock market.
In other words, the AI boom is becoming too big to simply bring to a halt without serious consequences. The IMF already estimates that a sharp pullback in AI investment could leave the U.S. economy around 1.8% smaller by 2028 than it would be under the current baseline scenario.
Anthropic’s IPO in November could become the first serious test of this whole setup. Not long ago, the market was discussing a valuation of around $2 trillion for the company. Following the latest figures, some investors are now talking about roughly $1.5 trillion. At the same time, the business itself continues to grow at an extraordinary pace.
From a capital markets perspective, I see the main risk here as something other than AI itself. I have little doubt that it will fundamentally reshape the economy and the world. But the gap between the hundreds of billions being invested today and the point at which those investments start generating comparable returns could be much longer than the market currently expects. That’s precisely where the risk of today’s euphoria lies. Major technological shifts almost never happen quickly or without friction.
Yesterday, Anthropic reported a net loss of $80.7 billion in just the first half of the year. To be fair, there’s no need to panic over the headline number itself. Around $72 billion of that loss came from an accounting revaluation of the company’s liabilities, not from Anthropic actually burning through that much cash. But something else is far more interesting.
The company has already committed hundreds of billions of dollars to computing capacity, chips, and data centers. And Anthropic is far from alone. Amazon, Microsoft, Google, and Meta are set to invest more than $700 billion in AI infrastructure this year. The scale has become so enormous that these investments are now supporting the U.S. stock market and accounting for a significant share of investment activity across the economy.
And this is where the U.S. is gradually falling into a very interesting trap. If tech companies keep spending at this pace, they’ll have to borrow more and more money and somehow prove that these massive investments will actually pay off. If they start cutting spending, the fallout will hit chipmakers, data centers, the energy sector, the tech companies themselves, and ultimately the stock market.
In other words, the AI boom is becoming too big to simply bring to a halt without serious consequences. The IMF already estimates that a sharp pullback in AI investment could leave the U.S. economy around 1.8% smaller by 2028 than it would be under the current baseline scenario.
Anthropic’s IPO in November could become the first serious test of this whole setup. Not long ago, the market was discussing a valuation of around $2 trillion for the company. Following the latest figures, some investors are now talking about roughly $1.5 trillion. At the same time, the business itself continues to grow at an extraordinary pace.
From a capital markets perspective, I see the main risk here as something other than AI itself. I have little doubt that it will fundamentally reshape the economy and the world. But the gap between the hundreds of billions being invested today and the point at which those investments start generating comparable returns could be much longer than the market currently expects. That’s precisely where the risk of today’s euphoria lies. Major technological shifts almost never happen quickly or without friction.