© Gorodenkoff / Shutterstock.com
Morgan Stanley‘s (NYSE: MS | MS Price Prediction) economics team just took its projection for artificial intelligence capital spending significantly higher, and the leakage math behind the headline number has become the more important story for US investors. On a recent episode of the firm’s Thoughts on the Market podcast titled “AI Spending: A New Engine for the Global Economy,” analysts revised their hyperscaler and AI-related CapEx estimates upward and walked through why a bigger topline does less for domestic GDP than the raw dollars suggest.
The Revised Forecast
The team’s own words captured the shift: “We were thinking a little over a trillion for 2027. Now we’re more like $1.2, $1.3 trillion, maybe as high as $1.4 trillion in 2028.” That trajectory sits alongside a Wells Fargo projection this week that top-four cloud service provider AI infrastructure CapEx alone will reach $1.1 trillion by 2027, with the bank hiking price targets on Alphabet (NASDAQ: GOOGL), Amazon (NASDAQ: AMZN), and Meta (NASDAQ: META) on the view that major cloud providers will pass higher AI infrastructure costs through to enterprise customers.
The Morgan Stanley figure is broader in scope because it captures the wider ecosystem: equipment makers, non-cloud infrastructure, and international operators. It is also consistent with Vanguard’s outlook work, which estimates the AI scalers alone will lay out $2.1 trillion in cumulative capital expenditure from Q1 2025 through Q4 2027.
Why 60% Leaks Out of the US Economy
The catch is composition. Roughly 60% of AI CapEx flows into “computers and peripherals, equipment spending categories that have a very, very high import content.” That imported hardware shows up on the wrong side of the trade ledger, which is why the US posted a $77.6 billion trade deficit in May 2026, the worst reading in a 12-month window that averaged a $60.8 billion monthly deficit.
Netting out the leakage, the Morgan Stanley team estimates “AI CapEx is probably contributing around 40 basis points to growth” this year, with a similar contribution expected next year. Against an economy the firm describes as growing “somewhere a little bit above 2% right now,” that matches the Bureau of Economic Analysis print of 2.1% real GDP growth in Q1 2026, driven partly by gross private investment of 7.9%. AI is meaningful at the margin without carrying the expansion by itself.
Asia Captures the Other Side of the Trade
The offset shows up abroad. Morgan Stanley notes AI spending is “fueling growth around the world, just not here in the US,” with semiconductor exports from Korea, Taiwan, and Japan growing by 90%. Global chip data supports the transmission mechanism. Worldwide semiconductor revenue reached $298.5 billion in Q1 2026, up 25.0% from Q4 2025. In March 2026, global semiconductor sales rose 79.2% year over year, while Asia-Pacific semiconductor sales totaled $86.2 billion, up 108.5% from March 2025. Taiwan’s IC industry logged NT$1,926.1 billion in Q1 2026 revenue, up 29.4% year over year.
Sustainability Questions Are Building
Not everyone thinks the current run rate holds. Palo Alto Networks (NASDAQ: PANW) CEO Nikesh Arora argued this week that token costs for enterprise AI must decrease by 90% within two years to achieve scalability, pointing to Uber (NYSE: UBER) having burned through its entire 2026 AI budget by April. Morgan Stanley itself flagged “potential continued volatility due to AI spending and capital expenditure uncertainties” in commentary on the KOSPI correction. Financing costs also matter, with the 10-year Treasury yield at 4.55% sitting in the 93rd percentile of its trailing 12-month range.
What to Watch
Investors tracking the domestic payoff should focus on three signals: the monthly US trade balance for the imported-equipment component, quarterly Asian semiconductor export data as a real-time proxy for hyperscaler orders, and enterprise AI unit economics. Corporate profits look healthy enough to fund the buildout, with total corporate profits reaching $4,426.5 billion in Q1 2026, up 12.8% year over year, and IT sector profits climbing to $352.5 billion.
Contact [email protected] for any questions or corrections.
