The AI spending boom is forcing a harder look at the established technology budgets that still fund much of enterprise IT.

As companies spend more on infrastructure, AI systems and the costs that come with using them, older categories of technology services are coming under pressure. Noshir Kaka, senior partner at McKinsey & Company, described a redirection of enterprise technology budgets already underway—one that could squeeze companies tied to legacy work while opening a much larger market for those able to move quickly.

“The only thing you can say for certainty is that if you’re standing still, you’re pretty much dead,” Kaka said.

Newsweek’s recent “AI Impact Forum” webinar, “The Trillion-Dollar Question: Who Wins and Who Loses in the Services Economy?” examined that pressure as Dr. Ranjit Tinaikar, host of the series, spoke with Kaka about how AI is changing enterprise technology economics.

Kaka said the broad spending picture still looks healthy. McKinsey research has found that 72 percent of roughly 690 executives surveyed were increasing their technology budgets, with average increases of about 6 to 8 percent.

That growth does not flow evenly across the technology stack. Companies are spending heavily on data centers, infrastructure, AI tools and rising token costs. When business units want more technology investment than the overall budget can support, chief information officers have to find money elsewhere.

That often means looking at the legacy technology book.

“You’re going to have to find that extra cash from somewhere,” Kaka said. “And that’s what’s hurting the industry today in spades.”

The result is a market moving in two directions at once. According to Kaka, traditional services categories are weakening as companies redirect spending toward AI infrastructure, consolidate vendors or demand productivity gains from existing work.

“Every enterprise that I know on average is increasing the spend on technology,” Kaka said. “And yet, when you look at the trickle down that’s actually happening in services…services feels like it’s in a recession.”

AI returns are adding another layer of pressure. Kaka said McKinsey research shows only 13 percent of companies surveyed were scaling AI across the enterprise, while many others remained in pilots or limited deployment. In many cases, he said, companies may see operational improvements but still struggle to find the impact in a profit-and-loss statement or balance sheet.

The return problem often starts with the target. Narrow pilots may improve a task, while broader workflow redesign can change the economics of an entire business process.

In pharmaceutical manufacturing, for example, companies can use generative AI to write a cleaner report after something goes wrong on the shop floor. A broader use would apply AI earlier, predicting which batch is likely to fall short of quality standards before it is produced.

“You can absolutely write a good English report and save a little bit of money,” Kaka said. “You can reimagine how you do manufacturing entirely by stopping the line and not producing a batch that is likely to be defective.”

The difference between those approaches can be enormous. Kaka said companies in the second category can earn “five to ten times the return” of those focused on narrower use cases.

Legacy work may still shrink, even as AI opens new markets for technology providers. Kaka said AI gives services companies a new way to move into software markets by modernizing older systems that were previously too expensive to refactor.

Software companies, meanwhile, can push deeper into services with more flexible products. The boundary between the two categories becomes less stable as each side gains new ways to attack parts of the other’s market.

The opening also extends beyond traditional software and services. High-skilled labor across fields such as consulting, legal, security and business operations may become addressable to technology-enabled service providers as AI makes more knowledge work easier to package, automate or augment.

Tinaikar summarized the scale of that opportunity plainly: “We’re seeing a 15-trillion[-dollar] market where all of them compete with each other.”

Services CEOs, Kaka said, now face a two-part challenge: protect against pressure on annuity work while rebuilding for the new demand taking shape around it. That means reinventing offerings, changing go-to-market motions and reworking operating structures built for a different era.

“If you [take] a proposal that you gave to a customer 18 months ago and a proposal that you’re giving to a customer today, if they’re not dramatically different, there’s something that you are doing wrong,” he said.

Investors, Kaka added, should look for signs that companies are becoming attackers rather than defending existing revenue. Waiting for the market to settle may be the risk.

“Speed will win,” Kaka said. Companies that move faster in changing their spending base, focus areas, talent and go-to-market motions are “likely to capture an incredible, incredible opportunity that lies in front of you.”

You can sign up for the next “AI Impact Forum” webinar, “AI in Finance: From Individual Adoption to Enterprise Transformation.” This session will examine how agentic AI could reshape financial services, from research and analysis to client work, business models and enterprise transformation.

It takes place on June 25 and will feature Kevin Buehler, chief innovation officer at Rogo and senior partner emeritus at McKinsey & Company.