Morgan Stanley: AI Adoption Phase Puts ROI, Margins in Focus
Morgan Stanley published 'AI Investing: Why Adoption May Drive the Next Wave' on Sept. 23, 2026, placing the next leg of AI returns with adopters rather than builders. Morgan Stanley Research expects roughly $800 billion in AI-related capital expenditures in 2026, rising to $1.1 trillion in 2027. Morgan Stanley
The article describes that next wave as likely rewarding companies that use the technology to improve decisions, lower costs and strengthen margins. Adoption is the filter. The emphasis is operating leverage from deployment, not incremental exposure to infrastructure demand.
That framing extends work from earlier in 2026. Morgan Stanley conducted its 5th global AI stock mapping across 3,600 stocks, and stated the market's focus is shifting from "AI exposure" to proof of ROI. Morgan Stanley Research Morgan Stanley Research separately estimated nearly $3 trillion of AI-related infrastructure investment will flow through the global economy by 2028. Morgan Stanley Scale the capex path against that stock universe and the question tightens. Who monetizes.
Positioning complicates the handoff. A Morgan Stanley document titled 'The Paradox of a Placid Market Amid Semi-Euphoria' states the ratio of long positions to short positions across semiconductors, IT hardware and AI power remains near the most bullish level of the past decade. In July, Morgan Stanley published 'What This Choppy Market Is Telling Investors' about high expectations, Big Tech rotations, AI spending shifts and portfolio positioning. Morgan Stanley
Strategist Michael Wilson sees a market rotation from semiconductor stocks to hyperscalers. His team said momentum is fading in semiconductor stocks as investors shift toward laggards. Bloomberg The call separates AI beta tied to chip volumes from cash-flow duration tied to cloud, software distribution and enterprise contracts.
Disruption risk sits on the other side of that rotation. Morgan Stanley published 'AI Disruption Concerns Do Not Compute' in March about investors punishing software and data stocks on AI disruption worries. Morgan Stanley In February, Morgan Stanley Investment Management added emerging-market bets insulated from the artificial intelligence boom. Bloomberg In January, Morgan Stanley identified AI as a key catalyst behind its 2025 mega-investment themes including tech diffusion and the future of energy. Morgan Stanley
In my view, the sequence points to dispersion, not a broad re-rating. Buildout capex of $800 billion moving to $1.1 trillion sustains revenue for the value chain while raising the bar for buyers of that capacity. Hyperscalers can absorb the spend if utilization and pricing hold. Enterprise adopters face a different test. Margin expansion must arrive without proportional headcount or compute-cost creep.
The broader context here is crowding against rotation. A long/short ratio near a decade high in semis, hardware and AI power leaves less room for positive revision in the early-cycle winners. Fading momentum there does not automatically validate hyperscalers or software. It shifts the burden of proof to ROI metrics that are harder to standardize across sectors: net revenue retention, sales productivity, inference cost per task, energy intensity.
Looking at what this means for how to read the next prints, focus on conversion. Capex guidance alone will not resolve the debate. Watch gross margin net of accelerated depreciation, free-cash-flow conversion after AI infrastructure leases, and disclosure around internal deployment. The mapping of 3,600 stocks gives breadth. Proof of ROI decides weighting.
