IT leaders face budget pressure as AI spending grows

It was the splurgiest of times, it was the thriftiest of times—both can be true of IT budgets right now, depending on the price tag attached to every exec’s favorite two-letter buzzword.

While some companies are starting to get sticker shock from all their AI spending, investment in the technology continues to be a top priority for IT leaders. That means fewer resources left for more quotidian line items like infrastructure maintenance or legacy software contracts.

A recent Goldman Sachs survey of CIOs found that 42% of respondents—up from 35% in a previous survey last November—now expect AI to exceed 10% of their budgets in the next three years. Meanwhile, overall IT spending dipped slightly, and funding for two-thirds of AI inference costs is coming from budget reallocation, according to the survey.

“There may be some crowding out of category spend not tied to AI,” Goldman analysts led by Gabriela Borges wrote in the note. “[And] there may be some hesitancy on committing to larger IT projects before the impact of AI on core budget categories is clear.”

Funneling back to AI

For Mike Arrowsmith, chief trust officer at IT platform NinjaOne, the last six months in particular have been full of meetings and projects aimed at weaving AI into every aspect of the company’s IT operations.

“We’ve literally jam-packed all of our schedules, all of our available capacity to adopt some flavor of AI for the benefit of being able to ultimately come back to finance to ask for less,” Arrowsmith said.

Because IT is typically seen as a cost center rather than a revenue generator, there tends to be more pressure to run leaner and optimize for cost and tool usage, Arrowsmith said. But for NinjaOne, the productivity gains made with AI have yet to lead to much savings as the company invests more and more in the tech.

“A lot of those cost optimizations are just simply shifting to the AI providers themselves,” Arrowsmith said. “Any savings that we would have by eliminating legacy tooling, we’re just shoving into AI to be able to accelerate our adoption and use cases within those platforms.”

Straightened priorities

While AI companies will often boast that more efficient coding tools can speed up development work, there’s also the risk of creating more tech debt if a company is neglecting other parts of the tech stack. Duane Barnes, president of cloud services provider RapidScale, said when working with clients on AI transformation, prioritizing revenue-generating and mission-critical functions helps minimize new tech debt that might crop up.

“So if you happen to create an island of software somewhere creating tech debt inadvertently, it’s not going to drag the business down while it sits there,” Barnes said. “You could deal with it later.”

Colm Sparks-Austin, EY Americas tech consulting leader, said CIOs should take AI implementation as an opportunity to cut projects that aren’t serving the bottom line or consolidate processes that are redundant between different teams. He even suggested that AI itself could be used to spot these duplicative operations.

“A meaningful chunk of most portfolios is work no one would fund if they were starting today,” Sparks-Austin wrote in an email. “The AI moment is useful because it gives leaders air cover to make those calls. And once you’re operating around products, the conversation gets simpler. It’s not about whose budget gets cut. It’s whether the work serves the product. If it doesn’t, it stops.”

Barnes said it’s also important to analyze how AI is actually being used to automate processes within the org and whether the costs are worth it.

“Now with AI, you kind of have to look at it like your power bill. What did I use last month? How many people were using it?” Barnes said. “There’s dashboards and reports available, and all these tools, most people aren’t using them.”

Software prices

Meanwhile, the software platforms catering to functions like IT are adding more of their own AI features and raising prices to match that supposed added value, according to Mitchell Couper, VP of data and analytics at SpendHQ, which provides AI tools to track business spending. Rising prices may be part of the reason companies are reevaluating those contracts, while spending on cloud platforms and AI models is increasing, he said.

“Because we are in the spend visibility space…I can see that people are asking questions: ‘How do we get all this under control? And do we need a plan for optimizing this going forward?’” Couper said.

Barnes said it’s important for IT pros to understand the wider business context to be able to build tools that will replace some of these software functions going forward, beyond core vendors like customer relationship management platforms (CRMs) and billing packages.

“My advice to hiring leaders in businesses that have IT teams is, make sure that the people you hire have business context. That’s so frequently missing in IT,” he said. “They actually understand how your company runs, how it’s funded, how you make money, what makes your customers happy, so that what they’re really working on is helping you run your business, not running IT.”