AI bubble? Nowhere near, says one leading brokerage
Wedbush is telling clients that the artificial intelligence boom is still closer to the start than the end, likening today to the mid 1990s internet build out rather than the 1999 dotcom bubble.
The broker thinks 2026 will be another strong year for technology shares, driven by heavy spending on AI infrastructure and the first real wave of monetisation. It expects the money being poured into data centres and related kit through 2025 to start showing up more clearly in revenues and profits next year.
Wedbush says it has seen about 20% of the AI related deals it tracks at the big cloud groups speed up in recent months, as large corporate customers try to get live projects into the market ahead of 2026. Chief information officers are still working out where AI fits into their organisations, but the bank believes this process will trigger a new phase of large scale deployments.
Capital spending by the large US technology groups alone is expected to reach between $550 billion and $600 billion in 2026. On top of that, Wedbush sees a “tidal wave” of AI investment from governments, global blue chip companies and buyers in Asia and the Middle East, with US platforms set to be key beneficiaries.
In its base case, the broker thinks technology stocks could rise another 20% in 2026 as this next stage of the AI build out takes hold. It expects the large US names to keep leading the market, followed by a second and third tier of beneficiaries in areas such as software, chips, networking and other “AI derivatives” plays tied directly to new use cases.
The note tackles the obvious question of whether this is all a bubble. Wedbush acknowledges the concerns, pointing to the sharp moves in AI exposed shares and drawing on its experience of the 1990s tech cycle. But it argues that this time is different in several important ways.
First, the consumer side of AI has barely started, in its view. Second, autonomous systems, including self driving vehicles, are only just beginning to roll out. Third, robotics remains largely a lab technology and has yet to hit mainstream adoption. Fourth, it estimates that fewer than 5% of US companies have gone “all in” on AI strategies. Finally, the rest of the world is even earlier in the adoption curve.
Taken together, Wedbush sees this as year three of a ten year AI investment cycle. It describes today as a fourth industrial revolution powered by massive technology capital spending, with “trillions” of dollars of expenditure still to come over the next few years.
The message for investors is straightforward. Despite bouts of volatility and loud warnings about excess, Wedbush believes the AI theme has much further to run. On its numbers, 2026 could mark an important turning point, as the focus shifts from building the plumbing of AI to cashing in on the services and applications that run on top of it.
