Morgan Stanley Advises a Barbell Strategy for the Next AI Wave

Morgan Stanley wants AI investors to hold two very different kinds of stocks at once. In a note published Thursday, a team of analysts led by Stephen Byrd recommended a "barbell" approach to investing in artificial intelligence — keeping the familiar AI hardware winners on one end while adding a new set of companies on the other that are only now beginning to reap the technology's benefits.

The argument rests on a simple observation: artificial intelligence is starting to benefit new types of companies, and that shift could force investors to adjust portfolios that have been built around a narrow group of AI names.

On one side of the barbell sit what the Morgan Stanley team calls "AI enablers." These are the companies at the core of the AI infrastructure build-out — chip makers chief among them. Byrd and his colleagues see no reason to abandon them. In their view, there is still room for AI hardware stocks to rise, even after the enormous run the sector has already delivered.

On the other side of the barbell sit the "AI adopters." These are businesses outside the traditional AI supply chain — the analysts point to companies in the transportation and real-estate sectors as examples — that are beginning to translate AI deployment into tangible business results.

One company the analysts highlight is Airbnb. According to Morgan Stanley, the short-term-rental platform has laid out a case for how AI is helping its business, making it an early illustration of what adoption looks like when it moves beyond the data center and into consumer-facing operations.

A broadening story

The recommendation marks a shift in how Wall Street's research desks are framing the AI trade. For much of the build-out, the investment case has concentrated on the suppliers of the technology — the semiconductor companies and other infrastructure players whose products power the models. The Morgan Stanley note suggests the analysts now see the beneficiary list widening into a much broader range of industries.

That broadening carries practical implications for portfolio construction. A barbell structure means investors hold both ends simultaneously rather than rotating out of one into the other. In practice, that means maintaining exposure to the chip makers and other enablers at the heart of the infrastructure build-out, while also branching into adopters in sectors such as transportation and real estate that are starting to realize AI benefits of their own.

The logic is that the two groups capture different phases of the same technological shift. The enablers monetize the build-out itself — the hardware, the compute, the equipment that makes the models run. The adopters monetize the application of that technology inside their existing operations, whether that means improving logistics, pricing, customer matching or, in Airbnb's case, the workings of a marketplace platform.

Why keep the hardware exposure?

The more notable half of the recommendation may be what Morgan Stanley declined to say. The analysts did not argue that the infrastructure trade is over. Instead, they wrote that AI hardware stocks still have room to rise — a judgment that the build-out continues even as the second wave of AI beneficiaries starts to form.

That dual stance distinguishes the note from calls for a wholesale rotation away from the sector's biggest winners. The message is one of addition, not subtraction: keep the enablers, add the adopters.

What it means for investors

For portfolio managers, the Morgan Stanley framework translates into a screening question. Which companies outside the obvious AI supply chain can already demonstrate, as the analysts say Airbnb has, a concrete case for how AI is helping the business?

The note, dated Thursday and led by Stephen Byrd, arrives as investors weigh how much of the AI story has already been priced into the infrastructure names and where the next leg of returns might come from. Morgan Stanley's answer is that the next AI wave will be broader than the first — and that portfolios built for it should be broader too.