Mark Cuban: AI Data Centers Are Overbuilt, Could Become Pickleball Courts

Mark Cuban is throwing cold water on the AI infrastructure trade. On the latest episode of the All-In podcast, the billionaire said the buildout is running ahead of what future AI efficiency will actually require, and much of that capacity may end up repurposed.

The pickleball punchline

Cuban said that if AI models and data centers become more efficient over time, much of the computing capacity being built today will become redundant. His quip: there will be “a lot of data centers that are going to be turned into pickleball courts.”

He said hyperscalers like Meta and Alphabet, which are collectively pouring hundreds of billions of dollars into building data centers, are correctly anticipating that AI usage will rise over time, but he predicted “technological breakthroughs” will make AI cheaper and more power-efficient.

The fiber-optic parallel

Cuban pointed to the early 2000s, when telecoms firms laid far more fibre-optic cable than the internet needed at the time, and much of it sat unused for years. He expects a similar price-performance curve to hit AI compute, leaving hyperscaler capex ahead of real demand.

The Broadcast.com founder, who sold his internet-radio startup to Yahoo for $5.7 billion in April 1999, sounded the alarm on Big Tech companies plowing cash into AI infrastructure, borrowing heavily to fund those outlays, and pledging to keep spending for years to come.

Do you want to see how to make more plays? Do you want to find gains yourself?

Unusual Whales helps you find market opportunities through our market tide, historical options flow, GEX, and much, much more.

Create a free account here to start conquering the market with Unusual Whales.

Not a dot-com repeat, but concentrated pain

Cuban argued the AI boom is “not the traditional dot-com bubble” because there are fewer companies going public at “crazy valuations” with “no revenue, no traffic, no nothing,” and none of the retail-mania tells that defined 2000.

Instead, he said the damage would be concentrated: it “could just destroy a lot of VCs and a lot of funds and a lot of PE” that have gone all-in on AI infrastructure bets, while sparing “most people across the US.”

The market is running the other way

Super Micro Computer just disclosed over $60 billion in new orders and a record backlog in its latest SEC filing, guiding gross margin to 15-17%, doubled from a previous 8% outlook, and the stock rose 24%.

Bank of America is telling clients that wafer fab equipment spending will surge from $144 billion in 2026 to $250 billion by 2028, while OpenAI plans to dramatically increase compute spending through 2030, and Microsoft just signed a multi-billion-dollar agreement to rent computing capacity from Mistral’s European data centers.

Cuban is not alone: other well-known investors, including Michael Burry and Jeremy Grantham, have raised similar concerns about too much money going into AI infrastructure.

Options market and stocks to watch

Watch NVDA for any signs of hyperscaler order digestion, since the entire AI capex thesis flows through its GPU pipeline.

Watch META and GOOGL, the hyperscalers Cuban singled out as pouring hundreds of billions into data centers, for capex guidance and any hints of pacing.

Watch SMCI after its $60B backlog disclosure and margin guide, and MSFT as it leans further into third-party compute rentals with deals like Mistral.

For more market coverage, see other news here.

Want more market intelligence? Create your free Unusual Whales account for options flow, market tide, GEX, and the full toolkit.