AI’s Quick Dip Before the Next Big Surge

​The “bubble” talk gets a reality check
If you’ve been hearing whispers that AI is a bubble ready to pop, Abhijit Dubey, CEO of NTT DATA, says: yes, there might be some froth — but don’t expect a long, ugly collapse. In his view, any AI bubble that forms will deflate faster than past tech manias, and what follows won’t be a whimper. It’ll be a stronger, more mature rebound once real-world corporate adoption catches up to the massive infrastructure buildout happening now.

A giant trend, even if the hype cools
Dubey isn’t remotely bearish on AI’s long-term future. He calls it a “massive secular trend,” meaning something that reshapes the economy over years, not months. His forecast is basically a two-step: the next 12 months bring a normalization phase where expectations settle down, and then AI comes out the other side sturdier — less hype-fueled, more business-fueled.

The supply chain is booked like a sold-out tour
One reason the boom still feels overheated is simple math: demand for AI computing power is outrunning supply. Dubey says supply chains for advanced AI hardware are “almost spoken for” for the next two to three years. That scarcity is already shifting pricing power toward chipmakers and hyperscale cloud giants — a dynamic that helps explain why their market valuations look stretched.

AI is reshaping jobs, whether we like it or not
Dubey also points to the growing unease around what AI means for work. He expects disruption over the long haul — calling out likely job dislocation across a five- to 25-year window. NTT DATA is already rethinking how it recruits and staffs for that future. Still, he emphasizes that the company is continuing to hire globally, even while preparing for a labor market that won’t look the same a decade from now.

Companies want AI, but not everyone’s using it daily yet
There’s another gap forming beneath the hype: executives are talking AI nonstop, but day-to-day usage by workers hasn’t fully caught up. A recent workforce survey found that daily use of generative AI is still noticeably lower than boardrooms suggest. At the same time, people who do have AI skills are getting paid a lot more — with wage premiums jumping dramatically compared to last year.

The skills divide is widening
The same survey also flagged a training imbalance inside companies. Roughly half of non-managers say they have access to AI training resources, while about three-quarters of senior executives do. If that gap doesn’t close, AI adoption could hit a speed bump — not because the tech isn’t ready, but because too many workers are being left without the tools to use it well.

Bottom line: a short wobble, then a bigger climb
​Dubey’s message is basically: expect a breather, not a breakdown. The infrastructure surge has raced ahead of adoption, and markets may flinch as that mismatch gets corrected. But once companies catch up and AI use becomes routine instead of experimental, the next leg up could be even more powerful than the first.