AI Layoffs May Be Inevitable as Spending Slows

“History tells us that at some point there will be layoffs. The only question is when and by how much,” says Sloan’s Gensler.

The current AI spending rate of $750 billion could remain “flat” or even decline next year. Alternatively, Gensler suggests that by 2028 or 2029, enough production capacity may exist for companies to begin cutting jobs. “It is almost certain that there will be job cuts,” he added.

Could an AI Market Crash Make the Industry Healthier?

Although layoffs may be inevitable, the future of the financial bubble surrounding AI—and the technology revolution built on top of it—could develop in several different ways.

Some Silicon Valley insiders are already rooting for an AI crash. Vijay Pande, a longtime venture capitalist, wrote in a recent blog post that “the upcoming crash may be the best thing to happen to this technology.”

The argument has some merit. A crash could make AI investment more rational, reduce the pressure to build multibillion-dollar data centers on every vacant lot and refocus investors on using the technology to create sustainable value.

However, investors and policymakers may need to be careful what they wish for. When the dot-com bubble burst in the early 2000s, hundreds of thousands of people lost their jobs, while companies of all sizes went bankrupt. The Silicon Valley and San Francisco economy suffered a devastating shock, and the broader impact contributed to the mild recession of 2001.

Hyperscaler Financing Raises Great Recession Comparisons

The Great Recession, which began in late 2007, was even more damaging to the economy and average Americans. It is sobering to compare the financial engineering that preceded that crisis with the methods now being introduced by today’s hyperscalers.

So-called special purpose vehicles, or SPVs, have returned. If Columbia University’s Van Nieuwerburgh is correct about the risks that hyperscaler investments pose to the wider economy, the consequences could be severe.

Technology Survived Previous Bubbles

Technology survived and even thrived after both recessions. The early 2000s remained a period of significant innovation and technological optimism despite the dot-com collapse.

The crash eliminated spending on low-quality technology and allowed investors to focus on more promising opportunities. It is no coincidence that each hyperscaler emerged from the crash or rose shortly afterward.

Fiber-optic infrastructure built during the communications bubble that ran parallel to the dot-com boom still forms the backbone of much of today’s communications infrastructure. Without it, companies such as Facebook, Amazon and Google would not exist in their current form.

Why Today’s AI Bubble May Carry Unique Risks

This time, however, the risks are different. Hyperscalers’ enormous financial commitments are tying the future of AI to the fortunes of massive data centers being built across the country.

That strategy depends on a deeply held belief in the scaling power of AI: the larger the system becomes, the more intelligent it gets. That may prove true, but it remains unproven—and represents a risky financial bet.

Several warning signs are already visible, including strong public opposition to new data centers, competition from inexpensive but powerful AI models and the rapid improvement of smaller, locally hosted models.

None of these trends necessarily points to a future dominated by frontier models housed in giant, billion-dollar data centers.

AI May Survive, but Massive Data Centers Face Uncertainty

The financial bubble surrounding hyperscaler spending will probably burst eventually, perhaps sooner rather than later. The consequences could be painful, but the technology industry may survive the shock.

Wall Street will survive. AI itself will survive, although it may look different and lose some of its current arrogance. The financial future and long-term usefulness of the large data centers supported by trillions of dollars in spending, however, are far less certain.