Who Will Actually Cash In on AI?
Morgan Stanley said on Sept. 23, 2026 that the next winners from AI may be the companies that use it, not the ones that build it. Its researchers expect about $800 billion in AI spending on equipment and data centers in 2026, rising to $1.1 trillion in 2027. Morgan Stanley
The report says those users could win by making better decisions, cutting costs and protecting profit margins. Think of it like a highway: builders laid the road, now the payoff depends on who drives trucks on it.
That idea builds on earlier work in 2026. Morgan Stanley reviewed 3,600 stocks worldwide for the fifth time, and said investors now want proof that AI spending pays off, not just AI exposure. Morgan Stanley Research It separately estimated almost $3 trillion in AI building will move through the world economy by 2028. Morgan Stanley
A Morgan Stanley paper called 'The Paradox of a Placid Market Amid Semi-Euphoria' says bets that chip, hardware and AI power stocks will rise are near a 10-year high compared with bets they will fall. In July, Morgan Stanley published 'What This Choppy Market Is Telling Investors' about high hopes, shifts between Big Tech stocks, AI spending changes and how investors are positioned. Morgan Stanley
Strategist Michael Wilson expects a shift from chip stocks to hyperscalers, the giant cloud companies. His team said price strength in chip stocks is fading as investors move to laggards, or stocks left behind. Bloomberg The idea separates money tied to chip sales from money tied to longer-term cloud and software contracts.
In March, Morgan Stanley published 'AI Disruption Concerns Do Not Compute' about investors selling software and data stocks on fears AI would hurt them. Morgan Stanley In February, its investment arm added emerging-market picks shielded from AI risk. Bloomberg In January, Morgan Stanley named AI as a key driver of its 2025 big themes, including spread of tech and the future of energy. Morgan Stanley
In my view, that points to a split, not a rise for everyone. $800 billion going to $1.1 trillion keeps money flowing to sellers of equipment but puts pressure on buyers. Big cloud firms can handle it if customers keep using the tools and prices hold. Other companies must show higher profits without much more hiring or computing costs.
The broader context here is crowding. Bets on chips, hardware and AI power are near a decade high, so there is less room for good surprises there. Weakness in chips does not prove cloud or software will win. Proof must come from basic business tests: keeping customers, selling more per worker, and keeping the cost and energy use of each AI task down.
Looking at what this means for the next company reports, watch what turns into cash. Spending plans alone will not settle it. Watch profit after equipment wear-and-tear, cash left after paying for AI buildings and leases, and what companies say about using AI themselves. The 3,600-stock map is wide. Proof of payoff decides who counts.
