On Wall Street these days, there’s a certain confidence that doesn’t shout. You read it in client notes that arrive a little early and hear it on morning calls. Oil remains above $100 due to the mess in the Strait of Hormuz, the S&P 500 is close to all-time highs, and technology continues to lead. It’s difficult to ignore how at ease the veterans appear as you watch it happen. People who experienced both 2000 and 2008 are returning to the AI industry; they refer to it as the “second wave” this time.
The phrase has become ubiquitous, appearing in research notes, CNBC chyrons, and casual remarks made by strategists who were previously wary. In general, Nvidia and the picks-and-shovels phase of the boom dominated the first wave. The second is more intriguing and messier. It concerns who sells the security software that prevents everything from collapsing, who builds on top of the models, and who actually uses the chips. Many older money seems to agree with Wedbush’s Dan Ives, who keeps saying that the next phase will be more about how the technology is deployed than how the chips are manufactured.
| The Second Wave AI Trade — Key Information | |
|---|---|
| Phase | The second wave of the AI boom, focused less on chip-making and more on real-world deployment, software, and enterprise artificial intelligence use cases. |
| Estimated 2026 Hyperscaler Capex | Around $700 billion across Microsoft, Alphabet, Amazon, Meta, and Oracle — roughly a 60% jump over 2025. |
| Major Players Driving the Trade | Nvidia, Microsoft, Alphabet, Amazon, Meta, Broadcom, and a widening list of cybersecurity and software names. |
| Key Veteran Voices | Dan Ives of Wedbush, Ed Yardeni, Jim Paulsen (formerly of Leuthold), Torsten Slok of Apollo, and Ryan Detrick of Carson Group. |
| Market Backdrop | S&P 500 near record highs; Nasdaq up sharply since late March; the benchmark index just posted its strongest month since November 2020. |
| Earnings Picture | About 84% of Q1 reporters posted positive earnings surprises, with tech earnings growth tracking near 30% for the year. |
| Main Risk | Concentration in a few mega-cap names, AI infrastructure overbuild, and unproven productivity gains — echoes of the late-1990s telecom buildout. |
| Time Horizon Veterans Are Watching | The next four to six earnings cycles, when AI capex must start showing measurable revenue impact rather than promise. |
The rotation itself is remarkable. Earlier this year, tech valuations fell to levels not seen since before ChatGPT was released in late 2022, according to certain metrics. Torsten Slok of Apollo noted that the sector’s price-to-earnings ratios had dropped from about 40x to 20x, which is the kind of information that attracts the interest of seasoned investors. They recall a time when purchasing technology at a 20x price felt careless. It feels almost conservative now. There’s a feeling that those who have been around long enough to notice an opening were filled by the war headlines.
This spring, you can practically feel the rebalancing when you stroll through any office tower in midtown Manhattan. Because Caterpillar is now selling power turbines to data centers and reaching record highs, junior analysts are, of all things, pulling up Caterpillar charts. As AI infrastructure plays, old industrials are being repriced. It’s a bit bizarre. A backhoe-and-bulldozer brand that shares Nvidia’s success. Instead of the early froth, that type of cross-pollination typically indicates the middle innings of a true boom.
The doubts haven’t gone away, though, and they probably shouldn’t. Former Leuthold strategist Jim Paulsen has been dissecting the productivity narrative. Paulsen’s voice is powerful because he doesn’t use it frequently. He discovered that the connection between innovation and widespread economic gains was weaker than people realized when he looked back at the dot-com era. Businesses made staff reductions under the pretext of “AI efficiency,” but the macro numbers didn’t change as the headlines claimed. It’s possible that a similar situation is currently taking place. It’s also possible that this time’s technology is truly unique. The fact that both can be true simultaneously contributes to the moment’s uniqueness.
Ten years ago, it would have been unimaginable for the hyperscalers to spend nearly $700 billion on capital expenditures this year. When some Wall Street analysts look at that number, they see the telecom expansion of the late 1990s, with fiber laid for demand that arrived years later than anticipated. Others observe that Alphabet, Microsoft, and Meta generate actual cash flow sufficient to pay their bills. It was dubbed a “show me” quarter by King Lip at BakerAvenue, and the term has stuck because it sums up the atmosphere. The expenditures must begin to justify themselves.
Even the cautious veterans seem to think that the second wave will reward selectivity instead of widespread exposure. Enterprise software, cybersecurity names, and the few cloud providers that have real clients paying for real workloads. Businesses that are “non-disintermediated” by AI—a term used to describe companies that AI supports rather than replaces—have been discussed by Daniel Newman of Futurum. As the upcoming quarters progress, that distinction will become increasingly important. Nothing has stopped the market from riding what Ed Yardeni refers to as the Roaring 2020s Express. No one wants to publicly state whether the second wave ends like the first or like 1999, but many are secretly placing bets on the former.
