AI Bosses Cheer Rollout as Workers Fear Displacement

More than two-thirds of business leaders expect to be better off financially in five years because of artificial intelligence, compared with 27 percent of the U.S. work force. That 41-point gap lays out the split: executives see a payoff, while workers face possible displacement and trust those executives less.

Nearly three-fourths of leaders said AI had significantly changed how work gets done, compared with 58 percent of entry- and midlevel employees, according to a report released this month from the Wharton School. “We have corner offices and cubicles, driving the same economy, but living in completely different realities,” John Gerzema, C.E.O. of the Harris Poll, said in an email. He described “a gulf in hope between what leaders think A.I. could bring to companies and what employees feel it is taking away from them.”

Who Gets the Promised Gains

The Milken Institute-Harris Poll Listening Project found that business leaders were far more likely than workers to expect financial benefits from AI. The survey figures don’t offer a matching account of workers’ own financial gains; instead, the reported concerns include displacement and weaker trust in executives. The people deciding how to roll out AI and the people whose work changes aren’t reporting the same reality.

The data-center expansion behind the AI push has its own gap between corporate announcements and what can actually be built. Bernstein Research analyzed thousands of planned U.S. data centers and estimated that only about 35% of announced capacity is likely to become reality. The industry calls some questionable proposals “Bragawatts”: companies announce projects despite possibly lacking experience, electricity, construction permits or customers.

“Ghost projects” add another layer of paper demand: developers submit multiple requests for electricity connections in hopes that at least one site will work. In Texas, such requests have reached more than five times the state’s existing electricity supply. Even Amazon, described as one of the world’s most experienced data-center builders, is struggling to secure enough electricity for its AI equipment. The report said speculative projects may make the industry’s infrastructure challenges look larger than they are.

The Financing Behind the Buildout

Nvidia has used its investment-grade credit rating and financial resources to support AI-model developers, data-center companies and neoclouds. Its backstop guarantees aim to help finance projects at lower interest rates. As of its last quarterly financial filing in July, Nvidia had $36 billion in off-balance-sheet AI cloud agreements for neoclouds, plus a guarantee of up to $105 billion for chips, land and power at a proposed 10-gigawatt OpenAI data center in Ohio.

Nvidia also signed as a tenant for $20 billion in future data-center leases that it will later assign to third parties. For the vast majority of the deals, Nvidia won’t have to pay unless a participant defaults. Its largest single guarantee covers OpenAI’s planned facility in Pike County, Ohio, which could be the world’s largest data center and requires Nvidia hardware for at least 20 years.

Nvidia’s latest quarterly filing in August outlined more than $500 billion in off-balance-sheet commitments over the coming years, most maturing through fiscal 2029. That month, Jensen Huang announced a partnership with six large asset managers for a $500 billion financing push to treat compute infrastructure as an asset class. The agreement offered financial firms up to $125 billion in residual-value guarantees on GPU financing.

Nvidia also seeded $2 billion in Brookfield’s inaugural AI Infrastructure Partners Fund, which aims to raise $10 billion in total capital. Huang described the shift in August: “We used to build chips that we sell, and these are technology components that people buy and use. But now, Nvidia’s AI factory platform is really an investable asset.”

When the Bets Meet the Bills

The market’s appetite for these arrangements is under pressure. Shares of CoreWeave, Nebius and IREN, described as cash-burning companies in Nvidia’s portfolio, fell at least 30% this month. CoreWeave’s credit-default-swap spread rose above 11%, and the company has about $51.6 billion in total debt. Yields on bonds maturing in 2031 and 2032 topped 15%.

“If the demand for compute doesn’t meet these really super lofty expectations, we’re going to have a problem,” said David Trainer, founder and CEO of financial research firm New Constructs. Nvidia declined to comment. Huang told CNBC on Thursday that demand for neoclouds was “off the charts,” but also said, “What is missing in the middle is the working capital to fuel that growth.”

New Constructs estimated Nvidia’s net present value of off-balance-sheet liabilities in fiscal 2026 at more than $153 billion, an increase of more than $100 billion from the previous year. Moody’s said guarantees were the fastest-growing type of off-balance-sheet financing arrangement in 2026. David Gonzales, a vice president at the agency, said such guarantees weren’t counted as debt when rating corporate buyers, but were “something we’re constantly monitoring.”

The system’s builders are making enormous bets on demand, power and future payments. Workers, meanwhile, worry more about losing ground than executives worry about losing their promised gains.