Big tech companies are expected to spend about $725 billion on AI this year, according to Statista. Beneath the surface as billions of dollars flow into AI, concerns about an AI bubble continue to percolate.
Capital expenditures are eating up the cash flow of major companies like Meta. The company revealed in its first quarter earnings report that capital expenditures were $19.84 billion, leaving free cash flow of $12.39 billion. Capital spending was previously expected to fall between $115 billion and $135 billion this year. Meta currently expects capital expenditures to be in the range of $125 billion to $145 billion.
“This reflects our expectations for component price increases this year and, to a lesser extent, additional data center costs to support future annual capacity,” the company’s earnings release said.
According to Yahoo Finance, the company’s stock price fell about 10% after the earnings announcement.
Spending on AI continues to soar due to chip and data center prices and competition among companies to claim victory in the current technology race.
“We’re in a weird situation here where all the hype is a lot of money being thrown after a bad thing,” said Dave Nicholson, chief technology advisor at Futurum Group and program instructor at Wharton Executive Education.
CIOs are under pressure to get their companies on board the AI hype train, but they must stay on track to deliver value. As a result, enterprise technology leaders are increasingly required to recognize when spending becomes a liability and create a framework for sustainable and responsible investment.
What broader market conditions mean for companies
Nicholson expects to see more targeted AI investments in the coming years.
“We’re going to see within the next two years that not everyone is willing to sustain this level of investment because they can’t demonstrate the positive ROI that comes from it,” he said.
The market will become smarter in distinguishing between speculative and sustainable AI strategies and invest accordingly.
“If you were to draw a line 10 to 20 years from now, you would see continued upward growth,” Nicholson said. “The closer we get to that, the more severe the decline is going to be for individual organizations.”
As corporate boards observe the spending trends of large tech companies and market reactions, their view of AI investments may shift from a focus on spending growth to a focus on first-mover spending.
“We’ve definitely gone from fear of missing out to fear of failing,” Nicholson said. “The majority of established companies are recognizing that a fast-following model is the best path to pursue.”
Warning signs that spending on AI is becoming a liability
The story of AI investing has changed. The experiment is over and it’s time to deliver value. Market expectations may be driving that change, but there is still a lot of trial and error ahead.
“When you treat AI spending, pilots, and pre-production experiments as a research and development program, you’re as comfortable with the idea of spending some money to see what doesn’t work as you are to see what works,” says Alex Bakker, a renowned analyst and director of primary research at ISG (Information Services Group).
But that doesn’t mean companies can feel too safe. CIOs are facing tough questions, including when AI investments range from strategic necessity to corporate responsibility.
If capital spending continues to grow without measurable results, boards and investors will notice red flags. Vague promises of value without indicators of progress between “then and now” are not a solid foundation for further investment.
“There are countless examples of 18-month failures where millions of dollars were spent and nothing was accomplished,” Nicholson said.
Spending without cost control can lead to business failure. Costs will continue to rise and may outpace value or may rise without value being realized.
“Token generation can get out of hand really quickly. As soon as you have agents leveraging models on behalf of customers…If you think rising cloud costs are the problem, you haven’t seen anything yet on how to generate tokens,” Nicholson said.
CIOs need to be able to articulate the business case for AI and tie it to spend before technical debt starts accumulating, says Ravi Soin, CIO and CISO of work management platform Smartsheet.
“If you don’t have clear use cases and you can’t control your spending, you’re essentially accumulating debt disguised as innovation and speed,” he said.
There are multiple ways for companies to hide AI spending that is spiraling out of control. Layoffs are a big trend in the technology industry right now. Meta plans to cut 8,000 jobs on May 20, 2026. Janelle Gale, the company’s chief human resources officer, said in a memo to Meta employees that the layoffs are “part of our ongoing efforts to operate the company more efficiently and offset other investments we are making,” the New York Times reported.
Do these large-scale layoffs mean tech companies are significantly improving the efficiency of their AI, or are they doing an “AI wash”?Meta employee Arnab Gupta, who is waiting to see if he will be among the layoffs, claims in a post on X that these layoffs are the result of companies struggling. “These layoffs will continue until we learn how to use AI. Until we learn to transform AI tokens into outcomes rather than just inputs.”
Building a framework for sustainable AI investment
Managing the cost of AI is different in many ways from managing the cost of other technologies, but company leaders have extensive experience in making smart purchases and investments. How can they apply that experience to the fast-paced and novel world of AI?
Build, buy, or partner decisions
Big tech companies are building their own AI infrastructure, but that doesn’t mean all companies need to do the same. Companies must decide whether to build AI capabilities in-house, use external vendors, or take a hybrid approach. Each option has its own cost considerations.
For some companies, the upfront cost of building their own infrastructure may make sense. But Nicholson said these companies are not the majority.
“I advise students that the risk involved in trying to pioneer something really, really new is probably not worth it, unless you’re specifically trying to develop something to sell to others,” he said. “If you’re an Oracle shop and you want to deploy this in your database environment, let Oracle do it for you in just a few months.
understand the cost
ROI remains the most important measure of AI success. Companies are investing heavily in AI with the hope that it will be worth it in the end. But to get there, you need to truly understand your total costs.
“Positive ROI? First, understand the “I.” And that’s difficult in the token era,” Nicholson said.
Soin expects many organizations to be surprised by the total cost. “Licensing is the lowest cost of these AI capabilities,” he said.
CIOs must also consider costs associated with implementation, change management, system integration, and ongoing agile engineering. Companies that choose to build must also add in infrastructure costs.
Connect spend with value
CIOs need metrics to quantify the value of their AI investments. Large, ambitious, multi-year projects are proliferating. However, measuring the effectiveness of these efforts and justifying their ongoing costs is more difficult.
“This is a disciplined and limited demonstration of value effort that will protect the CIO’s job,” Nicholson said. “Small value proof tests can tell you a lot about what makes sense to invest in.”
Set a realistic timeline for ROI
Among the CIOs and CTOs Nicholson works with, 18 to 24 months appears to be a typical time frame to achieve ROI on AI investments. By setting and adhering to timelines, companies can avoid further spending on use cases that don’t work as intended.
“Be honest with yourself and your precautions,” Bakker says. “If you don’t see movement, don’t continue. Cut your spending, try again, and allocate your tokens and effort and human time and energy to new use cases.”
Monitor your spending in real time
Spending on AI requires disciplined oversight, not a set-it-and-forget budget.
“You need real-time visibility, not the quarterly checkpoints required by AI cost models,” says Soin. “Do we have the right framework and associated guardrails to trigger alerts and quarantines when cost overruns occur at that workload level?”
Stress test investment strategy
”[Look] What is overall AI spending as a percentage of that budget and how has it trended year-over-year? I think that establishes the scale and trajectory that the organization faces,” Soin said.
Is that percentage sustainable? Can it be adjusted? CIOs need to be able to consider how the company’s AI investment strategy can adapt to changes in vendor capabilities, business needs, use case outcomes, and company revenue.
Communication with the board of directors and investors
Corporate boards and investors are taking note of the massive investments in AI and the expectation that this technology will deliver measurable gains in efficiency and productivity.
“They’re being dragged along by global hype, and that expectation is placed on top of the people who work in the organization,” Nicholson said.
CIOs are on the front lines of meeting and managing those expectations. Bakker said the company is well-positioned to educate boards about AI and how it can drive business value, and to do that CIOs need to interact with boards as if they were investors. Investors understand the spend, they understand the risks. It is the CIO’s job to discuss alignment between the two with board members.
“If you’re a high-dividend company with very stable cash flow and a very stable business, your AI strategy might be sensitive to that,” Bakker said. “If you’re a hyper-growth company, you probably have to take more risks now.”
Opportunity for CIOs
Big tech companies like Meta have to pour billions into AI investments, but they still have a responsibility to their investors. Companies and their CIOs can monitor how the market reacts to AI spending trends as they continue to develop investment strategies.
CIOs can help companies develop intentional investment frameworks that connect AI to business value.
“In my opinion, the CIOs who are leading the way are not the ones spending the most on AI,” Soin said. “Show me a solution that has been implemented that has measurable value and a governance structure that allows it to scale responsibly.”
Carrie Paradis is a freelance journalist with experience writing about cybersecurity, technology, and healthcare. She currently covers a wide range of issues relevant to today’s CIOs and IT leaders.
