Experts say the market’s AI anxiety will ‘punish those chasing the froth’
Major technology stocks tied to artificial intelligence took a sharp downward turn Tuesday, rattling markets and raising concerns the sector’s billion-dollar promises may not be bearing fruit as quickly as hoped.
Shares of Palantir Technologies, the data-analytics firm widely viewed as an AI bellwether, plunged more than 9%, its worst tumble since March, after prominent short seller Andrew Left of Citron Research renewed his bearish stance. Other major names felt similar shocks, highlighting underlying investor doubts: Oracle, in the midst of aggressive AI investments and a strategic pivot that included mass layoffs in its cloud division, saw its shares drop nearly 6%. Chipmakers integral to the AI boom struggled as well: Advanced Micro Devices fell 5.4%, Arm Holdings lost 5%, and Nvidia, the sector’s dominant force, slid 3.5%.
SoftBank, whose outsized bets on AI have defined its recent strategy, dropped more than 7%—amplifying concerns about a broader tech correction and underscoring Wall Street’s uneasy relationship with the so-called next big thing. OpenAI CEO Sam Altman even admitted AI is in a bubble.
The abrupt sell-off echoes broader skepticism about the sustainability of sky-high valuations seen in AI-focused companies. But experts say that while investors are right to be cautious, the underlying technology isn’t going away—and this is a short-term drop during a long-term transformation.
What’s causing the current AI anxiety
Behind the market jitters, a recent report from MIT said approximately 95% of company generative AI pilot programs resulted in “little to no measurable impact” on revenue or profits. While a handful of startups have thrived, the vast majority of corporate efforts have stalled, caught in flawed enterprise integrations and learning gaps. The research, encompassing 150 executive interviews, 350 employee surveys, and an analysis of 300 public AI deployments, paints a sobering picture: Outside exceptional cases, generative AI projects have yet to justify the vast spending across the sector.
MIT’s lead author, Aditya Challapally, told Fortune failure may lie less in the underlying tools than in enterprise execution, citing issues around workflow adaptation and resource allocation. In contrast, nimble startups have rapidly scaled revenues—validating the potential of the technology when well integrated, but also highlighting a gulf between hype and reality for larger companies.
“There’s no doubt that when MIT reports a 95% failure rate in AI pilot programs, it’s alarming,” Mike Sinoway, CEO of AI-powered search software company Lucidworks, told Fortune. “But the problem has less to do with the underlying technology and more with how companies are approaching it.”
“In our own research, polling over 1,600 AI practitioners and leaders and validating this with bot analysis, we found 65% of teams are rolling out AI without the fundamental tech infrastructure in place,” he said. “Trying to build cutting-edge applications atop weak foundations is like building an F1 car on a go-kart engine—you simply won’t get results. So while a 95% failure rate might seem like a sign of a bubble, once organizations focus more on what AI actually needs to succeed, we’ll begin to see the traction everyone is expecting.”
Chase Feiger, CEO of Ostro, agreed current volatility is part of a typical tech cycle. “Talk of an AI bubble isn’t new,” Feiger told Fortune.
“Every major tech shift goes through a stage where hype runs ahead of business fundamentals,” he said. “Some companies are burning money on inference costs, offering ‘all-you-can-eat’ models that cost thousands to run but bring in only hundreds in revenue—a pattern reminiscent of Uber’s early years. That overinflation explains market caution, but the underlying technology isn’t overhyped. In health care, for example, AI is transforming drug development, patient care, and physician decision-making.
“The correction will come. But over the long haul, the winners will be those who prove AI delivers durable value in complex, high-stakes environments,” Feiger added.
Short-term froth, long-term transformation
Harvard professor Christina Inge told Fortune the duality at work is nothing new.
“Investors are right to be cautious,” she said. “Not every company claiming to be ‘AI-driven’ is creating real value; a lot of it is smoke and mirrors, with some tools amounting to incremental improvements on non-AI tech. Correction is inevitable, as history shows.
“But the technology isn’t going away. AI is already making a difference in health care, marketing, logistics, and finance. And we’re only scratching the surface. In the long run, I expect the impact of AI to rival the Industrial Revolution. There’s a lot of froth in the market right now, but the bigger story is just beginning. In other words: short-term bubble, long-term transformation.”
That view is echoed by Shay Boloor, chief market strategist at Futurum Equities.
“What we’re seeing isn’t a bubble, but the foundation of a new economy,” Boloor told Fortune. “There will be volatility—inevitable with a sector this hot—but the fundamental reality is every industry will be transformed by AI. Just look at Microsoft and Meta this quarter: Azure hit its biggest revenue numbers ever, Microsoft Cloud crossed $46 billion, and Meta monetized not just attention but intelligence, with 22% revenue growth and 38% profit growth, while spending $70 billion in CapEx. The demand is not hypothetical—it’s scaling now.
“We’re not at the peak of AI. We’re at an inflection point.”
Separating winners from pretenders
Siamak Freydoonnejad, co-founder of Sprites AI, which makes an AI-powered marketing agent, says, however, deciding whether or not we’re in an AI bubble “misses the point” entirely.
“Stock prices may have outpaced fundamentals, but inside enterprises, AI is already infrastructure,” Freydoonejad told Fortune.
“No one who’s seen campaign launch speed improve by 70% is going back to the old way,” he said. “Some vendors did slap ‘AI’ on legacy products to cash in, but those valuations will be corrected—and deservedly so. What matters is which firms are using AI not as a shallow trend but as the basis for their entire product. Real efficiency gains are showing up for companies embedding AI deeply in their workflows. The market is about to sort out those with substantive results from those selling only promises.”
Omar Kouhlani, CEO of Runmic, which uses AI to design revenue strategies for sales teams, told Fortune infrastructure spending reveals the true momentum.
“Big Tech just raised AI spending guidance to $360+ billion for 2025, up sharply from previous estimates. I watch those numbers more closely than day-to-day share price changes,” he said.
“This isn’t a rejection of AI, it’s a market becoming more selective,” Kouhlani continued. “The crash is separating real AI revenues from companies that only have AI PowerPoints. We’re not in another dot-com bust. The infrastructure is being built now, and expectations are adjusting faster than the technology itself.”
Usha Haley, the W. Frank Barton Distinguished Chair in International Business and professor of management at the Barton School of Business at Wichita State University, argues that cycles of bubbles and corrections are intrinsic to tech revolutions. “Historically, every breakthrough technology comes with bubbles,” Haley told Fortune. “AI is already delivering productivity gains, even as it erodes some jobs. We’ll see some correction and consolidation, but not a collapse. The strongest players will emerge into a changed landscape. Regulation and stochastic shocks could alter outcomes, but competitive environments—not monopolies—will point to future leaders.”
Fabian Stephany, a lecturer at the University of Oxford, sees evidence for both sides: “To some extent, yes, there is an AI bubble. But long-term fundamentals are exceptionally strong,” he told Fortune. “Many firms use AI for marketing more than substance, which has inflated valuations. Yet, stock-market gains this year are overwhelmingly linked to real advances in AI at companies like Nvidia, Meta, Microsoft, and Broadcom. Nvidia alone accounts for 26% of the S&P’s advance, underscoring real market transformation.
David Brudenell, executive director at Decidr, which builds an AI-powered operating system for businesses to automate workflows, told Fortune that “correction is necessary” as it “separates speculation from structural value.” And David Russell, global head of market strategy at TradeStation, agreed “oullbacks are normal after rallies stall.”
“Major players like Palantir and Microsoft failed to hold breakouts after strong earnings. That’s a sign the good news may be priced in,” Russell told Fortune. “Markets move ahead of fundamentals, but excessive prices punish those chasing the froth. In the weeks ahead, sentiment could shift to other macro factors.”
The expert consensus is clear: While stocks have pulled back, the fundamentals behind AI remain strong. Most believe the recent rout is an overdue market sorting—separating hype from reality, speculation from enduring value. Even MIT’s cautious findings are seen as a spur rather than a death knell.
Now, all eyes will turn to Nvidia, which reports quarterly earnings next week. But broadly speaking, what the market isn’t experiencing isn’t a sign of crisis, but a marker of growing pains.
For this story, Fortune used generative AI to help with an initial draft. An editor verified the accuracy of the information before publishing.