Nvidia is working on fresh AI deals worth over $750 billion, accelerating a spending wave that skeptics worry inflates demand across the industry. The deals build on a pattern of massive infrastructure investments designed to support the AI boom, but investors are increasingly questioning whether these expenditures translate to real returns.

Nvidia’s strategy mirrors a broader tech sector problem. Companies are spending heavily on AI infrastructure — chips, data centers, models — but the return on that investment remains unclear. Critics call this “circular financing”: companies buy chips to build AI systems, then use those systems to justify buying more chips. No real economic output required.

The Market’s Skepticism

Chipmakers from Asia to Wall Street are feeling the pressure. SK Hynix dropped 13% and Samsung fell 10% in late July as doubts spread about AI spending sustainability. South Korea’s Kospi Index fell 9%, wiping out months of gains in the semiconductor sector.

The selloff signals that markets are starting to price in a reality: not every dollar spent on AI infrastructure will generate returns. Infrastructure spending without corresponding productivity gains is dead capital.

Nvidia’s Bet

Nvidia isn’t slowing down. The company is betting that this investment wave — whether from Nvidia itself or its customers — will eventually pay off through new applications, higher adoption, and competitive advantage in AI.

But timing matters. If the payoff doesn’t materialize within 18 months, expect more downward pressure on chip stocks and a broader recalibration of AI spending priorities.

What’s at Stake

The semiconductor industry has built its 2026 outlook on sustained AI demand. If that demand softens — either because of oversupply or underwhelming returns — the entire sector faces a correction.

The real test isn’t whether AI is valuable. It’s whether the current level of investment matches actual, measurable AI value creation.