NVIDIA captures $690 billion of AI spending, with the remaining 499 S&P 500 companies sharing $340 billion.

Original author: Zhao Ying

Source: Wall Street Journal

Morgan Stanley has mapped out the full flow of AI capital expenditures for 2027 for the first time, revealing a highly concentrated profit structure: of the $1.58 trillion spent by hyperscale cloud providers, approximately two-thirds of the S&P 500’s受益 amount flows to NVIDIA alone, while the remaining 499 components collectively receive only about $339 billion.

Morgan Stanley strategists Nick Lentini and Mike Wilson, along with over twenty industry analysts, jointly released the report “Tracking the $1.6 Trillion AI Capital Flow: Mapping the Revenue Landscape Created by the Capital Expenditure Boom,” which for the first time traces individual capital expenditures by hyperscale cloud providers directly to specific companies’ revenue statements. The report’s key conclusion is that AI capital spending will directly drive approximately 25% of S&P 500 sales growth, with AI-related companies accounting for 78% of S&P index earnings growth in 2027.

The report also reveals a long-neglected hidden risk: hyperscale cloud providers have disclosed over $1.1 trillion in uninitiated lease commitments, which constitute "synthetic capital expenditures" not reflected in any cash flow statement under capital spending, yet are economically equivalent to debt financing. Morgan Stanley credit strategists anticipate that 2027 will be the year with the largest free cash flow gap for hyperscale cloud providers, when debt financing pressures will rise significantly.

The core of Morgan Stanley’s report is a Sankey diagram that traces the $1.58 trillion in 2027 spending by nine major buyers—Alphabet, Amazon AWS, Microsoft, Meta, Lease Pool, SpaceX, Oracle, CoreWeave, and Nebius—layer by layer to semiconductors and servers, networking and optics, data center infrastructure, storage, and international suppliers, ultimately landing on specific stock tickers.

Among the nine major buyers, semiconductors and servers represented the largest single spending category, absorbing $672 billion, with the vast majority flowing to NVIDIA. The second-largest category was "international" suppliers, totaling $511 billion, encompassing Samsung, SK Hynix, TSMC-related infrastructure, and Taiwanese ODM manufacturers—all of which are not part of the S&P 500.

The report also notes that the map links $1.6 trillion in capital expenditures to approximately $1.7 trillion in supplier revenue, with the difference arising from disruptions caused by private emerging cloud providers and Bitcoin miners on customer allocations, timing discrepancies, and supplier price increases not yet reflected in capital expenditure forecasts. Morgan Stanley’s interpretation is that the next round of capital expenditure increases is already embedded within supplier models.

Of the approximately $1.03 trillion in revenue absorbed by the S&P 500 from AI capital expenditures, 67% flowed to NVIDIA, amounting to about $690 billion—equivalent to 44% of the entire $1.58 trillion capital expenditure budget landing on a single company’s income statement. Excluding NVIDIA, the entire AI capital expenditure boom contributed only 1.6% to the S&P 500’s sales.

Roughly $700 billion is attributed to NVIDIA, approximately $340 billion is shared among the remaining 499 S&P components, and another $550 billion has flowed out of the index to South Korean memory manufacturers, Taiwanese ODMs, and various private companies.

At the individual stock level, Morgan Stanley analyzed the proportion of each company’s 2027 revenue tied to AI infrastructure spending. NVIDIA leads at 91.7%, followed by Erock (EROC) at 91%, Vertiv (VRT) at 90%, Applied Opto (AAOI) at 83%, Broadcom (AVGO) at 81%, and Bloom Energy (BE) at 80%. Even Caterpillar (CAT) has 21% of its revenue linked to AI capital expenditure trends; Seagate (STX) and Western Digital (WDC) both reach 60%.

This structure means that for at least a dozen large companies, "AI capital expenditures" are nearly synonymous with "revenue"—upside boosts profits, and any adjustment to capital spending directly translates into earnings revisions, with no other business segments to absorb the impact.

One of the most important methodological adjustments in the report is the inclusion of leases in capital expenditure estimates. Morgan Stanley added $195 billion in capital expenditures for 2027 and $246 billion for 2028, using disclosed signed but not yet activated leases, analyst forecasts for new capacity additions, and the lease capital expenditure ratio per GW of new capacity.

This increases the total AI capital expenditure in 2027 from $1.39 trillion (cash capital expenditure) to $1.58 trillion, representing a year-over-year growth rate of approximately 60%; in 2028, it rises from $1.57 trillion to $1.81 trillion.

In terms of lease scale, Oracle leads at $57 billion, followed by Microsoft at $45 billion, Meta at $37 billion, Amazon at $28 billion, and Alphabet at $27 billion. As of the second quarter of 2026, the pipeline of signed but not yet activated leases has surged to $1.118 trillion, more than doubling in just two quarters, with Microsoft ($329 billion), Oracle ($288 billion), and Meta ($279 billion) as the primary signatories.

Morgan Stanley's qualitative assessment is that leasing data centers is economically identical to capital expenditures financed by debt, merely circumventing GAAP capital expenditure categories.

Morgan Stanley credit strategists Fernanda Lima and Vishwas Patkar expect free cash flow for hyperscale and emerging cloud providers to reach its most negative point in 2027, with a larger funding gap if lease payments are included.

Large-scale cloud providers have issued approximately $255 billion in unsecured bonds year-to-date, accounting for about 30% of their estimated 2026 capital expenditures. Amazon ($98 billion) and Alphabet ($81 billion) lead the pack, while overall yields have risen by 100 to 180 basis points since the beginning of the year, with Oracle experiencing the most significant expansion in its 10-year bond spread at 180 basis points.

Morgan Stanley's stress test shows that if an AA-rated hyperscale cloud provider issues $400 billion in bonds at a 7% interest rate, its interest coverage ratio would decline from 30.7x to a still comfortable 11.9x; however, if Oracle and SpaceX raise $120 billion in financing at an 8.5% interest rate, their combined interest coverage ratio would plummet from 4.0x to 1.8x.

The report explicitly states that emerging cloud service providers are the most reliant on debt financing, "almost entirely depending on debt issuance to fund their capital expenditure ambitions." Several days before the report's release, CoreWeave's stock price had already declined sharply, with the company corresponding to $47 billion in spending on Morgan Stanley's capital expenditure map.

Morgan Stanley expects AI capital expenditure growth to decelerate significantly from around 60% in 2027 to approximately 14% in 2028. Against this backdrop, the report identifies companies that maintain pricing power despite slowing capital expenditure growth, with power equipment leading the selection, followed by storage, hard drives, and cooling equipment.

Power supply is another structural constraint. Morgan Stanley’s proprietary “AI Power” model shows that even accounting for various rapid power supply solutions covering 34% of the required 97 GW, the U.S. will still face a net shortfall of 32 GW before 2028 and an expanded gap of approximately 72 GW before 2029. The report characterizes this as a timing mismatch rather than a demand collapse, but notes that storage, networking, optical, and ODM construction are most sensitive to power interruptions.

In terms of downside risk, Morgan Stanley used unusually direct language: the current market is driven by a "fear of missing out," with buyers paying high prices today because they expect even higher prices tomorrow; once supply catches up with demand, this logic will reverse. The report cites the 2022 NAND memory cycle as an example—where shipments declined by only about 5%, but prices fell by 20% to 30%—and concludes: "We believe the market is currently underestimating the risk of simultaneous declines in both volume and price."

Morgan Stanley's report provides, for the first time, a dollar-by-dollar map of AI capital spending flows, revealing that approximately 25% of S&P 500 sales growth and 78% of profit growth rely on a $1.6 trillion capital expenditure budget—of which two-thirds of the index-wide benefits are concentrated in a single company, with an increasingly large portion supported by leases that do not count as capital expenditures and debt requiring refinancing.