Mark Cuban sounds alarm over OpenAI’s AI spending, ‘trillion dollars spent just to be an app’

Mark Cuban has joined the list of skeptics questioning the AI ​​spending boom, arguing that OpenAI is overspending on infrastructure with little certainty of generating adequate returns. The billionaire investor believes the ChatGPT maker’s ambitious trillion-dollar investment plan could backfire for two key reasons.

Cuban said that despite doubts that the investments will yield the expected rewards, companies are spending aggressively to maintain their leadership in artificial intelligence, noting that OpenAI is at the top of the list and is definitely overspending.

OpenAI raised $110 billion in its latest funding round, roughly three times the size of the largest IPO in history. The company plans to spend $1 trillion on infrastructure in the coming years.

“They’ll never get it. They’re just throwing this money away,” he said during an interview on the Big Technology Podcast. Here are two key reasons why OpenAI believes its ambitious spending could ultimately backfire:

Infrastructure gamble: Can OpenAI justify the spend?

First, whether OpenAI will be able to generate enough profits to justify the massive expenditure is a serious concern.

“There’s a lot of FUD being thrown around about AI spending,” Cuban said. “We’re going to spend a trillion dollars because we need all this data center capacity.”

Cuban argues that the math doesn’t make sense. As technology improves, computing power will become cheaper; This means that companies spending billions of dollars today may encounter expensive infrastructures that lose value much sooner than they expected.

“The ability to process is going to get faster and cheaper. And it’s going to happen faster and faster than people expect,” he said, adding: “I think a lot of the numbers they’re putting out are not going to come to fruition.”

Earlier this month, AI critic and researcher Ed Zitron also noted: “If OpenAI fails, it will be because LLM models are unprofitable.” They spend heavily on artificial intelligence infrastructure, but subscription business and advertising revenues do not cover these costs.

It’s too early to call the winners

Second, Cuban believes it is too early to assume that OpenAI will justify its trillion-dollar valuation. The AI ​​market is still taking shape, and it’s not yet clear who the long-term winners will be.

“We don’t know whether the business of the core models (ChatGPT, Gemini, Grok, Claude) will be like the streaming industry, where there is a single leader and a group of players making money, or like search, where there is effectively one company,” he said.

If you go all out and don’t win, “you just spent a trillion dollars being an app.”

Earlier this week, JP Morgan CEO Jamie Dimon also suggested that AI spending will eventually pay off, but not in the way we expect.

Pointing to previous trends, he said early major players such as Yahoo and Netscape faded away during the internet boom, while eventual winners such as Google and Facebook emerged later. “Will it pay off in the way you expect and on the timeline you expect? Absolutely not,” Dimon said.