Tether Ceo Warns Of Four Cracks In Big Tech’s Ai Boom

What Happened

In a July 4 post on X, Ardoino said the AI infrastructure race contains four major “structural mismatches”. These are gaps between costs, revenues, investment timelines, and competition.

His warning comes as the world’s largest technology companies pour hundreds of billions of dollars into data centres, chips, and power capacity. The central question for investors is whether AI can generate enough revenue to justify that spending.

Big Tech companies are spending heavily now, while the profits from AI may take much longer to arrive. Data centres, GPUs, and power contracts require huge upfront investment.

Chinese hedge funds including Wealspring Asset and Shanghai Banxia Investment Management Center have warned that AI stocks may be in bubble territory. Wealspring reportedly called global AI stocks a “super bubble”, while Banxia said a possible trigger for a correction may already have appeared.

The concern is simple. AI has become a major driver of stock market performance, especially for large technology companies. If investors begin to doubt the return on AI spending, the impact could spread beyond the tech sector.

Some investors take a less negative view. They argue that today’s AI trade differs from the dot-com era because the largest companies funding the boom already have strong earnings and established businesses.

Morgan Stanley has also estimated that nearly $3 trillion in AI infrastructure investment could move through the economy by 2028.

Market Context

Tether CEO Paolo Ardoino has warned that Big Tech’s artificial intelligence spending boom may be built on weak economics, as concerns over an AI market bubble spread across global markets.

That makes growth look stronger than the underlying business model. If companies later raise prices, users may spend less. If they keep prices low, margins may remain under pressure.

This creates a gap between capital spending today and commercial returns in the future. The bigger that gap becomes, the more pressure companies face to prove that AI can become a durable source of revenue.

Open-source AI models are improving quickly and could weaken the pricing power of commercial AI providers. If cheaper or free alternatives become good enough, customers may resist paying premium prices.

Ardoino’s warning is part of a wider debate now moving through markets.

That level of spending gives AI a central role in corporate earnings, energy demand, chip demand, and credit markets.

Why It Matters

Ardoino argued that companies are charging too little for AI computing compared with the real cost of providing it. In simple terms, some AI services may look cheap because companies are subsidising usage to win customers.

AI Profits Could Take Longer to Realize

That matters because companies may need to replace expensive hardware before it has fully paid for itself. If demand slows or pricing falls, the economics become harder to defend.

That would make it harder for companies to recover the money they are spending on infrastructure. It could also reduce the revenue expectations that have supported high AI valuations.

AI Spending Could Hit Trillions

JPMorgan has projected that global AI-related spending could reach $5.5 trillion by 2030. At the same time, Alphabet, Amazon, Meta, and Microsoft are expected to spend up to $720 billion this year.

The Bank of England warned in October 2025 that AI-related valuations had moved close to levels seen during the dot-com bubble. It also said AI infrastructure may require trillions of dollars, with a meaningful share financed by debt.

Details

Businesses are Paying Too Less for AI

AI is Outdating Itself Fast

AI chips can become outdated within 3 to 5 years. Yet the debt and equity used to finance AI infrastructure often assume a much longer payback period.

The Industry Faces Intense Competition