There's a New AI Bubble No One Is Talking About: What Companies Can Do to Keep Up

"AI" written in three different colors and fonts embedded in colorful bubbles.

Famed investor Michael Burry is once again raising the alarm about an AI bubble, as return on invested capital slows and OpenAI declares a "code red" amid increased competition.

But the market bubble is only one side of the story. A bubble of equal size — and perhaps greater concern — is forming around AI "adoption," as board-facing messaging becomes increasingly detached from the realities of corporate AI implementation.

The growing market bubble around tech companies has long been a concern for investors and analysts, who see striking parallels between the valuations of today's AI companies and those of companies that were breaking records when the dot-com bubble burst.

From just $107.88 $24.99 for Kiplinger Personal Finance

Become a smarter, better informed investor. Subscribe from just ~~$107.88~~ $24.99, plus get up to 4 Special Issues

CLICK FOR FREE ISSUE

https://cdn.mos.cms.futurecdn.net/flexiimages/y99mlvgqmn1763972420.png

Sign up for Kiplinger’s Free Newsletters

Profit and prosper with the best of expert advice on investing, taxes, retirement, personal finance and more - straight to your e-mail.

Profit and prosper with the best of expert advice - straight to your e-mail.

But this emphasis on market prices has the corporate world ignoring the expectations bubble that's been growing right under their noses.

Fear of missing out (FOMO) and market-led urgency have businesses jumping at opportunities to show investors that they've adopted AI into their business operations. According to McKinsey, 88% of business leaders in 2025 reported regularly using AI in at least one business function — up from 78% the year prior.

While AI devotees see these numbers as evidence that AI is living up to its potential and transforming the economy, a deeper dive into the figures paints an alarming picture.

McKinsey identifies just 6% as "high performers" — those seeing 5% or more of earnings before interest and taxes (EBIT) attributable to AI. Meanwhile, only 39% of respondents report any enterprise-level EBIT impact from AI, with most of those seeing less than 5%.

Even for those only watching the market bubble, that should be a concern. Capital is being priced right now as if companies are seeing gains from their AI use. If companies fail to turn AI into an additive venture, that reality will eventually hit the market.

About Adviser Intel

The author of this article is a participant in Kiplinger's Adviser Intel program, a curated network of trusted financial professionals who share expert insights on wealth building and preservation. Contributors, including fiduciary financial planners, wealth managers, CEOs and attorneys, provide actionable advice about retirement planning, estate planning, tax strategies and more. Experts are invited to contribute and do not pay to be included, so you can trust their advice is honest and valuable.

Yesterday's data weakness, today's AI failure

The biggest problem in the AI market right now is that AI actually is as powerful as it promises to be. An effectively trained AI can (and will) transform the operations of a well-prepared business.

The issue is that most businesses aren't prepared.

While AI scores highly on the factors that make it an adoptable technology (relative advantage, observability and trialability), it scores weakly on the remaining two factors: Compatibility and complexity.

Actually building effective AI into workflows requires vast amounts of training, discipline and data that most companies simply don't have — because most companies have failed to adopt AI's precursor technologies.

For many businesses, cloud computing has been a failure. Regardless of the size or sophistication of a business's enterprise resource planning (ERP) or customer relationship management (CRM) systems, most companies never develop the discipline to ensure data consistency.

What's driving the AI implementation bubble is that those same companies are now trying to bolt AI onto an unstable foundation. And AI, unlike many past technologies, is unforgiving.

An ERP can limp along with messy data because humans are there to compensate; AI only scales the mess. Bad inputs don't just create bad reports — they create automated answers that move faster than people can catch.

The 'pilot' paradox

The AI implementation bubble is a result of this confluence of factors: FOMO driving leadership behavior, AI's seeming "adoptability" and underlying data immaturity at an organizational scale.

The race to adopt AI has led boards to be distracted by activity metrics, like the number of AI pilot programs a company is running. The problem with these quantitative metrics is that they don't actually say much.

It's like a restaurant bragging about the number of ovens they have. Unless the restaurant can also report the number of chefs they have, the quality of their ingredients and the menu, knowing the number of ovens doesn't actually say much about their ability to cater a dinner.

More critical than simply running pilots is knowing which pilots are actually ready to scale. The unfortunate answer for most businesses is that most aren't.

The majority of AI pilots fail to progress to enterprise-scale deployment — not because AI as a technology isn't ready, but because the companies running them aren't ready for AI.

The AI bubble that's forming is fundamentally an expectations bubble, as companies fail to meet AI's full potential. When this bubble bursts, the winners and losers will not be determined by who ran the most pilots, but by who has the right culture.

As urgent as the race to adopt AI feels, without the proper data infrastructure, process flexibility and AI literacy, companies cannot even begin to compete.

The difference between those who see value in AI and those who do not comes down to the idea of readiness and culture. It's not a matter of finding some novel use for AI, but of data discipline.

Companies with clean, structured data and the discipline and flexibility to adopt AI into their processes are the ones maximizing their returns.

One effective, scalable pilot with measurable outcomes is more valuable than a dozen scrapped test cases. Even when models don't behave as expected, digging in and diagnosing what's not working can help companies implement critical changes organization-wide.

The AI bubble isn't about the technology living up to the hype. AI is ready. It's the companies that need work.

The winners won't be the companies with the most pilots. They'll be the ones who built the foundations that make AI worth implementing at all.

Disclaimer

This article was written by and presents the views of our contributing adviser, not the Kiplinger editorial staff. You can check adviser records with the SEC or with FINRA.

TOPICS

Ryan is a partner at RubinBrown LLP, where he leads the AI & Data Practice, helping private equity portfolio companies and large organizations turn AI and data ambitions into measurable business value. With over 20 years of experience in technology and strategy, his work spans AI and data strategy and implementation guided by a proprietary methodology built on the principle that organizational capability matters more than technology selection.