✕

Markets

Insider Trading

Hedge Funds

Retirement

Opinion

Morgan Stanley Says These 8 Companies Will Spend $1.6 Trillion on AI Infrastructure in 2027

Page 1 of 3

Morgan Stanley still expects eight major companies to spend approximately $1.6 trillion on AI infrastructure in 2027. In its November 8 report, Buy the Drop: AI Investment Opportunities Post the US Midterms, the bank puts their combined capital expenditure at $1.641 trillion in 2027 and $1.760 trillion in 2028. The report maintains the aggregate spending outlook after the election, while warning that new requirements for data-center developers could push some construction into 2029.

The forecast predates the latest round of earnings updates. Morgan Stanley reiterated its rounded $1.6 trillion and $1.8 trillion estimates in a November 11 podcast, explaining that developers already hold permits or have started work on projects supporting 2027 construction. Its analysts expect about 5 gigawatts of U.S. projects, roughly 20% of the 2028 pipeline, to move into 2029. They identify potential approval delays in Texas and Ohio in the November 8 research note.

The bank now projects 19.9 gigawatts of new U.S. critical IT capacity in 2028, down 4% from 2027, followed by 38% growth in 2029. Its updated construction analysis also estimates a 44-gigawatt power shortfall through 2028. For related investment ideas, see our 10 Cheapest AI Data Center Stocks to Buy According to Billionaires.

Financing costs were already climbing before the election: Morgan Stanley ABS strategist Carolyn Campbell said in an October 16 podcast that yields on the highest-rated data-center bonds had risen from below 5% at the start of 2026 to nearly 7%. She explained that higher borrowing costs can squeeze returns when operators have already fixed rental income under long-term leases.

J.P. Morgan’s October 19 AI spending outlook estimated $900 billion of global AI capital expenditure in 2026, up from an earlier $736 billion estimate. Its researchers pointed to cloud revenue and contracted future revenue as support for continued investment, while identifying GPU and memory supply, permitting and public opposition as constraints. Our 15 Best AI Stocks to Buy and Hold for the Next 5 Years examines the broader investment theme.

Morgan Stanley’s November 8 report says financing remains available, although lenders could scrutinize new construction and specialized AI cloud providers more closely. It identifies tenant quality, access to electricity and contractual protections as important lending considerations. The bank also expects substantial AI-related borrowing in 2027 and argues that the largest operators retain advantages in scale and access to capital. In its podcast, the bank said existing data centers with stable cash flows could benefit from greater scarcity and stronger lease pricing. It expects political changes to affect project locations, timing and costs more than the decision to build at all.

Other brokerages have challenged the recent market pessimism. Reuters reported on November 9 that UBS considered investors overly pessimistic about political, credit and spending risks, while Jefferies described the November 5 selloff as “reactionary.”

Also read: Goldman Sachs AI Stocks: Top 12 Stocks to Buy.

4kclips/Shutterstock.com

Our Methodology

For this article, we selected the eight companies given in Morgan Stanley’s November 8, 2026 report, “Buy the Drop: AI Investment Opportunities Post the US Midterms.” We ranked stocks by their 2027 capital expenditure forecasts, counting down from the smallest spender at #8 to the largest at #1.

Note: These figures represent Morgan Stanley forecasts.

Page 1 of 3