Morgan Stanley analysts say despite growing investor anxiety, major tech giants will achieve returns on invested capital between 25% and 50% from their massive trillion-dollar artificial intelligence spending.

  • Tech hyperscalers are deploying an estimated combined $1.4 trillion in capital expenditures to build out AI infrastructure. 
  • Morgan Stanley outlines frameworks demonstrating that GPU rentals and model-enabled APIs can generate incremental returns on invested capital.
  • Wall Street remains cautious over short-term cash flow pressures and heavy spending guidance, creating volatility for tech stocks. 

Severe capital expenditure plans by tech industry hyperscalers like Amazon.com Inc. (AMZN) Alphabet Inc. (GOOGL,GOOG) and Microsoft Corp. (MSFT) will ultimately produce substantial financial rewards despite recent market hesitation, according to a research report from Morgan Stanley.

Led by analyst Brian Nowak, the Morgan Stanley team projects that major cloud providers—investing a combined $1.4 trillion into AI infrastructure—are positioned to achieve long-term Return on Invested Capital (ROIC) between 25% and 50%.

Addressing ongoing concerns among market participants, Nowak wrote that Morgan Stanley remains "bullish on the long-term ROIC from these investments" and introduced three core frameworks to demonstrate how the unit economics scale effectively over time.

AMZN stock marginally slipped 0.2% on Tuesday; GOOGL and MSFT stock added about 2.2% and 1%, respectively.

Generative AI Led Profitability 

Morgan Stanley’s model outlines three distinct generative AI revenue drivers that underpin their positive outlook, including hyperscaler GPU rental business, model-enabled APIs and,third-party infrastructure. 

By analyzing per-hour pricing, chip utilization rates and underlying costs such as hardware depreciation and energy usage, Morgan Stanley estimates that renting out graphics processing units (GPUs) can deliver 60% to 70% incremental EBIT margins and an ROIC of 25% to 40%. 

In addition, Morgan Stanley sees tech deployment of model APIs on third-party capacity is projected to yield about a 25% ROIC, with profitability driven by product innovation and compute efficiency (token throughput). 

Nowak noted that while continuous investment is required to maintain product innovation and throughput, the fundamental unit economics support the elevated spending. 

Short-Term Stock Volatility Vs. Long-Term Capex Strategy

While Morgan Stanley maintains a bullish multi-year view on capital efficiency, Wall Street's immediate reaction to the massive capital expenditures by Microsoft, Amazon, Meta, and Alphabet (Google) has been marked by skepticism, heavily impacting equity valuations.

Alphabet and Meta shares have declined 16% and 8%, respectively, on a year-to-date basis purely on account of uncertainty that capex plans could generate viable returns to justify the expenses.

Morgan Stanley's report concludes that despite near-term headwinds and stock price volatility, the scarcity value of AI computing infrastructure and high operational leverage will ensure that these massive investments yield strong long-term shareholder value. 

For updates and corrections, email newsroom[at]stocktwits[dot]com. <