There’s not enough money in the world to pay for the AI boom
The biggest bet in the history of the United States keeps getting bigger and we’re all hanging on for the ride.
Technology companies recently launched a record scramble for cash to pay for data centers, computer chips and technology and equipment to power artificial intelligence. This boom is so expensive that even Silicon Valley giants are bleeding cash and companies are now racing one another to collect or borrow other people’s money for the estimated $10 trillion of AI-related spending in coming years.
Three eye-popping figures that emerged on Monday showed the scale of the AI cash grab.
Anthropic, maker of the Claude chatbot, started the clock on what is expected to be the largest initial public offering in history, raking in an estimated $100 billion in stock sales. But the company’s IPO paperwork warned that will just be the down payment necessary to fund the company’s AI ambitions. “We anticipate making significant and increased investments for the foreseeable future,” Anthropic said in its IPO document.
The same day, Meta said it would sell up to $95 billion in new stock after disclosing a few days ago that it was struggling to compete with other AI developers asking investors for cash. And less than five months after Elon Musk’s SpaceX held what stands as the current record largest IPO, the AI and rocket company finalized a complex deal to borrow $40 billion to pay for AI computer chips and equipment.
How to summarize the sentiment among AI firms right now? “Take it when you can,” Matt Orton, chief market strategist of investment firm Raymond James Investment Management, said about the AI companies’ cash grabs. “Everyone is going to need more.”
Not every company will be able to secure the cash needed to keep up in the AI race, Wall Street and global economic power players are now realizing. The global economy and the personal finances of just about everyone are tied to the outcome of the AI boom and its escalating costs.
In a widely cited draft research paper last month, Columbia University finance professor Stijn Van Nieuwerburgh estimated that it will cost more than $10 trillion through 2032 to build out AI including data centers and computer equipment.
Measured as a slice of U.S. economic output, the average yearly AI spending works out to more than the combined national commitments for electrifying American households in the early 20th century, building the interstate highway system and installing the backbone of the 1990s U.S. dot-com boom, he estimated.
AI companies are starting to hunt for funds in nearly every imaginable financial nook and cranny to meet those immense costs, Van Nieuwerburgh and other experts say. They’ve been borrowing hundreds of billions of dollars, selling fresh stock and cooking up more novel and potentially risky methods of financing. [Continue reading…]