Step Aside, ExxonMobil: 2 Foundational AI Stocks Now Pay Larger Dividends

Microsoft and Nvidia have surpassed ExxonMobil's more than $16.9 billion annual dividend outlay, paying approximately $27.05 billion and $24.15 billion annually, respectively. Nvidia raised its quarterly dividend 25-fold to $0.25 per share amid sustained AI-data-center GPU demand and supply constraints, while Microsoft has increased its dividend for 21 consecutive years. Microsoft's Azure revenue growth reaccelerated to 43% in fiscal Q4 following generative-AI integration, supported by roughly $183 billion in annual operating cash flow and $76.8 billion in cash and short-term investments.

Analysis

The dividend comparison is not a new earnings catalyst; it is evidence that AI leaders are transitioning from reinvestment-only growth vehicles into capital-return compounders. For NVDA, the payout remains economically immaterial relative to earnings power, so it neither creates a valuation floor nor changes the key debate: whether hyperscaler capex and accelerator pricing sustain through the next product ramps. A larger recurring distribution can marginally broaden the shareholder base, but an AI-infrastructure multiple reset will still dominate the stock’s return profile over the next 1-3 quarters.

MSFT has the more investable second-order setup. Its mature software cash flows can absorb AI infrastructure spending while preserving shareholder returns, unlike cloud competitors with less diversified profit pools. That makes MSFT relatively resilient if enterprise AI monetization takes longer than expected: Azure can retain workload growth while Office/Copilot pricing, seat penetration, and operating-margin commentary become the incremental catalysts over the next 6-18 months. The relevant risk is that AI capex rises faster than Azure gross-profit conversion, creating margin dilution despite revenue growth.

Contrarianly, dividend optics may encourage investors to underweight TSM’s strategic leverage. Incremental AI demand is constrained by advanced packaging and leading-edge foundry capacity, not simply GPU design demand; TSM has less demand-disruption risk than NVDA if customers diversify accelerators toward AMD, custom ASICs, or internal silicon. The near-term bottleneck trade is therefore better expressed through TSM than through a further expansion in NVDA’s already premium valuation, subject to Taiwan geopolitical-risk tolerance.

Key Decisions for Investors

• Maintain MSFT as the core AI long versus a basket of lower-quality SaaS (IGV) over 6-12 months. Add on evidence that commercial remaining performance obligations and Azure growth remain firm while operating-margin guidance is maintained; thesis fails if AI-related capex drives sustained margin-guide cuts or Azure decelerates materially for two consecutive quarters.
• Prefer long TSM / short NVDA as a 3-6 month relative-value trade after NVDA strength, sized modestly for Taiwan-tail risk. It captures advanced-node/packaging scarcity and custom-silicon diversification; stop if TSM’s AI-related capacity utilization or advanced-packaging commentary weakens, or if NVDA demonstrates renewed pricing expansion through its next ramp.
• Do not buy NVDA on dividend framing. Treat the next earnings report as a catalyst only if hyperscaler demand visibility, gross-margin trajectory, and Blackwell/Vera Rubin supply conversion support estimates; absent that evidence, use rallies to reduce directional exposure rather than add.
• Watch MSFT Copilot paid-seat adoption, Azure AI gross margin, and quarterly capex as the decision dashboard over the next 1-3 months. A favorable combination of accelerating monetization with stable margins supports multiple expansion; capex acceleration without monetization is the key warning signal.