AI market correction: What IT leaders must know

Signs of a market correction

With AI valuations soaring, it is easy to draw comparisons to the dot-com bubble of the early 2000s.

"The difference between where we're at now to 2000 is there is an underlying business case for this," Yalowitz said. "The question is: Will we be able to get through the inertia to make that business case real in time? Will the markets have patience?"

Money is flowing into the market today, much of it using debt, Yalowitz noted. In June, Anthropic closed a $35 billion debt deal to pour into AI infrastructure expansion, according to Axios. OpenAI has a $4 billion credit facility. Oracle is using debt to fund its buildout, CNBC reported.

"When you have a dynamic with a massive amount of debt flowing into the market to fund infrastructure, the tipping point for a correction is what?" Yalowitz asked. "It's when the business that is promised because of that spend fails to materialize as quickly as the market would like it to." 

Companies such as Alphabet, Amazon and Microsoft are tapping their cash-flowing businesses to fund AI infrastructure, and they are feeling the squeeze, according to Yahoo! Finance. AI is gobbling up Capex at Meta, and the company cut 8,000 jobs earlier this year. Meta attributed the workforce reduction to efficiency and offsetting other investments, the New York Times reported.

As the billions keep flowing into AI investments, enterprise adoption and value creation are lagging. Last year, Gartner predicted that 40% of agentic AI projects will get the axe by the end of next year with rising costs, lack of business value and insufficient risk controls as the familiar culprits.

AI projects are suffering from failure to launch and scale, and the pricing model is changing. Tokens are expensive, and companies are blowing through them and driving up their costs. Some companies are spending $7,500 per employee on AI per month, according to Ramp Economics Lab. That dynamic could prompt enterprise customers to rethink how they use AI.

"While there might be an exceptional revenue case to be made that does support the valuation, if the underlying cost assumptions to deliver that revenue -- meaning compute cost per token isn't there -- that's where I think we start to see some compression in valuations," Yalowitz said.

Enterprises are also starting to better understand the risks associated with AI. In 2023, 12% of S&P 500 companies disclosed a material risk related to AI. In 2025, that number shot up to 72% of S&P 500 companies, according to Harvard Law School Forum on Corporate Governance.

"When you have to start claiming material deficiencies, now we're starting to talk Enron-level events," Doug Gilbert, CIO and chief digital officer at digital transformation company Sutherland, said.

Assets such as land, buildings and racks will continue to have value, even following a market correction, according to Yalowitz. But the market needs to see clear business value that justifies the money being spent.

"We pretty quickly are going to need to see CEOs shouting from the rooftops that AI has enabled them to have their sales teams own two times the number of accounts that they own today or drive revenue growth by X or Y amount," Yalowitz said.

What a correction could mean for IT leaders

An AI market correction will alter the vendor landscape. Some companies will flourish. Others will be bought or outright fail. However, if that change unfolds, enterprise IT leaders could find themselves dealing with the consequences. Reliance on a vendor swept up in a wave of consolidation or that goes out of business could upend AI projects that have already consumed considerable investment.

And the runway for enterprise AI investment is going to shrink. Boards, C-suites and investors are going to want concrete results, not open-ended experimentation and a crowded pilot graveyard.

"These budgets are going to be heavily scrutinized. I don't think anymore it's going to be a pool of money…just go be AI heat-seeking missiles to find problems and apply AI to it," Adam Field, chief AI officer at Tungsten Automation, a workflow automation software company, said. "It's going to be: Apply the right technology, but let's go look at each one of these programs individually like we used to do and bring things back to reality."

As AI investment becomes tethered to reality, IT executives may find themselves navigating the painful process of canceling and ripping out projects that they implemented with more enthusiasm than groundwork and risk management.

"If you did not build that observability in upfront, I foresee that we have a lot of projects being canceled," Gilbert said. "I also foresee that a lot of AI projects will now have to be ripped out because they're going to start saying this creates risk for the company."

Strategic opportunities

While enterprises will endure some of the painful aspects of an AI market correction, there are also strategic opportunities.

Enterprises may have greater leverage in negotiations as the vendor landscape shifts and they become more discerning about the kinds of partnerships that will drive value.

"Customers might pull back and just say: All right, we're going to just keep these people instead of trying to automate it because we can't predict a token," Field offered as an example. "There may be a little bit of a capacity there to negotiate and have some buying power."

But he anticipates that it may take time for negotiation power to tip in favor of enterprise customers. "Demand is still, I think, racing faster than GPUs can be built and data centers can be powered. So, maybe next year, 18 months, but probably not right today," Field added.

IT executives may also take advantage of AI talent. They could carve out a competitive advantage by building their AI talent internally. Talent that knows how to use AI can help companies be more competitive and productive, according to Yalowitz. "Support your AI superstars and the folks that are early adopters that will lead your organization through this change," he said.

Enterprise teams may also benefit from recognizing that they cannot reap all the benefits of AI without external help.

"We're starting to see a lot more startups in the delivery space, just experts to come in and consult," Field said. "There are millions, billions of dollars to be made in bringing in these AI experts."

How IT executives can prepare

Guiding enterprises through a market correction will require IT leaders to act now and develop long-term strategic positioning.

Immediate actions to take:

• Evaluate AI vendors. A market correction brings AI vendor risk to the fore. IT executives can conduct vendor risk assessments of AI companies to understand the potential impact on their enterprises.
• Define the right metrics. IT leaders need to define what AI value looks like in their organizations, which means defining and tracking specific metrics. "You could probably put together some metrics and say, 'We've saved this many hours, this many clicks, this many people,'" Field said. "We've not necessarily seen it result in better products yet or more revenue in most places." 
• Audit AI initiatives. How do current AI projects perform against those metrics? Are they delivering ROI? Are they creating risk that outweighs value?
• Focus on disciplined financial planning. Enterprises need an AI investment strategy grounded in reality, unclouded by hype. That means committing to investment with clear business use cases, tying spend to defined metrics and building flexibility into budgets to account for market changes.

Longer-term positioning:

• Focus on the fundamentals. Identify and invest in the necessary foundational work to give AI initiatives the best chance of success. "The companies that have put in the proper foundational work, governance work, observability work, guardrails around that AI, deployed it carefully, they're able to now move at such an extreme speed because they're just at the beginning," Gilbert said.
• Invest in change management. "You really have to think about the change management aspect of AI," Gilbert said. "It is not about taking an old crappy process and layering AI on top of it. It is about fundamentally rethinking the way that your company operates."
• Prove value. A successful IT executive strategy will hinge on proving value with AI, internally and to their customers. "It's now a C-suite-wide priority that includes the CIO and CTO to say, 'We need to be able to prove business value of this product at home so that we can share that story with our customers,'" Yalowitz said.
• Communicate with leadership. IT executives need to communicate effectively with their boards and C-suite peers to steer their enterprises through a market correction. "Executives need to act as that translator between what is being hyped and what reality looks like, and then plan accordingly," Gilbert said.

AI bubble predictions will rise and fall with the market. Whether there is a singular point at which that bubble bursts or it gradually deflates, AI is not going anywhere. The interim could be rocky, but there are opportunities for enterprises to secure leadership positions. As Yalowitz pointed out, "Every tech bubble we've had has actually led to positive things in the long run."

Carrie Pallardy is a freelance journalist with experience writing in cybersecurity, technology and healthcare. She currently covers a wide range of issues relevant to today's CIOs and IT leaders.