Third-Party Capital Leverages $500 Billion in AI Factory Investment, Revenue Sharing Opens Long-Term Upside

Overview

This report assesses the impact of NVIDIA joining with 6 institutional investment partners to mobilize up to $500 billion in potential funding and establish an independent compute financing platform. The funds are intended for neocloud providers and sovereign artificial intelligence projects. Investment partners will independently evaluate each project, while NVIDIA provides ecosystem support and may assume up to 25% co-investment or residual support. Morgan Stanley believes this structure, while expanding AI compute supply, helps ease market concerns about customer financing and circular demand, and may create long-term high-margin revenue through usage-linked revenue sharing.

Core views

The positives clearly outweigh the negatives. First, third-party institutions are responsible for primary underwriting, decision-making, and funding, making projects closer to independent commercial investments rather than demand driven unilaterally by NVIDIA. Second, new capital is expected to accelerate the construction of AI factories by sovereign and neocloud providers, strengthening customers' reliance on NVIDIA's platform. Third, the revenue-sharing mechanism may extend equipment sales into a long-term, recurring revenue source that is close to pure gross margin. Fourth, $500 billion is only potential investment capacity rather than committed spending, and actual deployment depends on project-by-project return assessments. Fifth, NVIDIA may still bear credit support risk, and rising ecosystem leverage will also intensify investment debates, but persistent compute shortages and roughly 10x growth in token demand make the overall outcome more likely to be positive.

Analysis framework

The report combines event and capital structure analysis, credit risk assessment, revenue-sharing sensitivity calculations, and P/E scenario valuation. The analysis first distinguishes third-party independent funding, NVIDIA equity investment, and credit guarantees, then estimates the impact of different graphics processing unit rental prices, deployment scales, and a 35% excess revenue share on revenue, gross margin, and earnings per share, and finally tests risk-reward under bull, base, and bear scenarios.

Methodology notes

• Earnings forecast frameworkMorgan Stanley ModelWare

  Forming revenue, gross margin, and earnings per share forecasts based on company operating drivers.

  The report uses Morgan Stanley ModelWare to forecast revenue, gross margin, inventory, and earnings per share, and uses calendar 2027 earnings per share of $13.08 as an important input for base-case valuation.

• Sensitivity analysisRevenue-sharing sensitivity analysis

  Calculating the impact of changes in rental prices, deployment capacity, and sharing ratios on incremental revenue and earnings per share.

  The model assumes that after revenue exceeds the project's breakeven price, NVIDIA receives a 35% share, and analyzes revenue, gross margin, and fiscal 2029 earnings per share upside under different hourly graphics processing unit prices and annual deployment scales.

• Valuation frameworkRisk-reward scenario valuation

  Forming bull, base, and bear target prices using different earnings and P/E assumptions.

  The base case values calendar 2027 earnings per share of $13.08 at about 22x, corresponding to $288; the bull case is $330 and the bear case is $160.

• Credit and capital structure analysisProject-level independent underwriting assessment

  Distinguishing third-party capital decisions, equity investments, equipment usage commitments, and direct credit guarantees.

  The report believes that third-party institutions assuming the main project screening and funding responsibilities can reduce concerns about circular financing, but NVIDIA's up-to-25% participation ratio and forms of support still need to be reviewed deal by deal.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• NVDA

  Direct research subject

  Strengths

  High market share in artificial intelligence accelerated computing, strong Blackwell and Rubin product cycles, full-stack capabilities across graphics processing units, networking, systems, and software, and potential new high-margin revenue sharing.

  Weaknesses

  High market share and high gross margins limit room for further near-term valuation expansion, while participation in customer financing increases credit and capital allocation complexity.

  Comparison

  Base-case valuation is about 22x expected calendar 2027 P/E, similar to the broad market, and at a discount to computing semiconductor peers such as AMD, AVGO, and INTC.

  Risks

  Funds are not actually deployed, insufficient project returns, credit guarantee losses, rising ecosystem leverage, customer in-house chips, intensified competition, export restrictions, and tariffs.

• AMD/AVGO/INTC

  Valuation and competitive reference

  Strengths

  Can benefit from expanding artificial intelligence infrastructure capital expenditure, and provide alternatives in accelerators, networking, or custom chips.

  Weaknesses

  The report believes NVIDIA still has leading advantages in product performance, software ecosystem, and system-level monetization.

  Comparison

  NVIDIA's base-case valuation is below the computing semiconductor peer levels cited in the report, but its market share and gross margin are already at high levels.

  Risks

  If NVIDIA's ecosystem and financing platform further strengthen customer stickiness, it may become more difficult for peers to capture artificial intelligence computing share.

Key data

• Potential investment scaleUp to $500 billionThis represents potential investment capacity and does not mean funds have been committed or will necessarily be fully deployed.
• Institutional investment partners6Partners will independently evaluate investment opportunities based on each project's own economics.
• NVIDIA potential participation ratioUp to 25%This may take the form of co-investment or a limited residual value support mechanism, with the specific structure determined by project.
• Existing partner lease guarantee exposure$3.5 billionFrom NVIDIA's latest 10-Q disclosure, existing direct credit exposure is mainly concentrated in a small number of smaller neocloud providers.
• Analyst ratingOverweightMorgan Stanley also names NVDA as its top pick in the semiconductor sector.
• Target price$288This implies about 32.4% upside from the August 10, 2026 closing price of $217.55.
• Valuation basisAbout 22x expected calendar 2027 P/ECorresponding to Morgan Stanley ModelWare's earnings per share forecast of $13.08.
• Revenue-sharing assumption35% of revenue above the breakeven priceUnder some deployment and pricing scenarios, potential upside to fiscal 2029 earnings per share can exceed 10%.
• Revenue growth assumptions82.0% in 2026, 52.4% in 2027The base case assumes artificial intelligence compute demand remains strong, and the Blackwell and Rubin product cycles maintain leadership.
• Scenario target pricesBull case $330; base case $288; bear case $160The main differences come from data center growth, the competitive landscape, product and software monetization, and valuation multiples.
• Global revenue exposureNorth America 50% to 60%; Europe (excluding the U.K.) 20% to 30%; Mainland China 0% to 10%The report also indicates exposure related to Asia Pacific (excluding Japan) as well as China and India.

Impact & implications

This financing arrangement may further transform NVIDIA from a supplier mainly relying on hardware sales into a capital coordinator and long-term revenue participant in the artificial intelligence infrastructure ecosystem. If third-party capital successfully expands compute capacity for sovereign and neocloud providers, NVIDIA can not only obtain incremental graphics processing unit, networking, and system revenue, but may also form high-margin recurring cloud revenue. However, this model will also increase market scrutiny of off-balance-sheet risks, credit support, customer balance sheets, and returns on artificial intelligence investment. Near-term share price catalysts depend on proof of independent financing for the first batch of projects and contract terms, while long-term value depends on compute utilization, rental prices, and the realization of revenue sharing.

Risks

• $500 billion is only nominal potential investment capacity, and projects may be delayed, scaled back, or ultimately not approved.
• NVIDIA may bear co-investment, residual value support, or direct credit guarantee risk, with specific losses depending on deal-by-deal contract structures.
• Leverage in the artificial intelligence ecosystem is rising, and some large cloud customers are already close to or in a free cash flow neutral or negative position.
• Artificial intelligence end demand or investment returns may fall short of expectations, potentially causing customers to sharply reduce graphics processing unit purchases.
• Compute supply catching up with demand more quickly may pressure data center revenue growth and graphics processing unit rental prices downward.
• AMD re-emerging as a strong competitor, or large cloud customers successfully developing custom chips, may erode NVIDIA's market share.
• Tariffs, export controls, and China market restrictions may have a greater-than-expected impact.
• High market share and high gross margins mean there is limited room for further near-term expansion in valuation multiples and profitability.
• Morgan Stanley has shareholding and investment banking and other business relationships with NVIDIA Corp.; investors should consider potential conflicts of interest.

What to watch

• The identities of the 6 investment partners, the scale of funding commitments, and the timing of the first projects.
• NVIDIA's actual funding ratio in each project, and whether the form of support is an equipment usage commitment, equity investment, or pure credit guarantee.
• Whether projects are truly independently underwritten by third parties, and whether NVIDIA's participation ratio can remain within 25%.
• The breakeven price, sharing ratio, utilization rate, duration, and accounting recognition method of revenue-sharing contracts.
• The financing costs, leverage ratios, and customer utilization rates of neocloud providers and sovereign artificial intelligence projects.
• Blackwell and Rubin supply, customer capital expenditure commentary, and growth in training and inference demand.
• The actual contribution of revenue sharing to revenue, gross margin, and earnings per share in fiscal 2028 to 2029.
• China market revenue recovery, export controls, and changes in tariff policy.
• Progress by AMD and cloud customers' in-house chips in performance, supply, and market share.