Nvidia (NVDA) vs Microsoft (MSFT)

NVIDIA designs and sells the AI accelerator GPUs that nearly every modern AI workload requires. Microsoft purchases NVIDIA chips at massive scale (estimated $30+ billion annually for Azure) and packages them into cloud services and Copilot products. The customer-supplier relationship makes the pair one of the cleanest in mega-cap equities: NVIDIA captures hardware margin, Microsoft captures software margin, both win when AI capex grows.

The revenue compositions reflect the stack position. NVIDIA Q4 fiscal 2026 revenue was $68.1 billion (90 percent data center). Microsoft Q2 fiscal 2026 revenue was $81.3 billion (Azure $90 billion annualized, productivity software $30 billion+, balance gaming and other). The data point for AI exposure: roughly $250 billion of NVIDIA annualized revenue versus roughly $90 billion of Azure annualized revenue. NVIDIA has 2.8x larger AI-direct revenue scale than Microsoft Azure.

NVIDIA briefly overtook Microsoft as the world's most valuable company on June 18, 2024, the first time in nearly two years. Microsoft had been the most valuable company for several months in 2024 after briefly overtaking Apple in January. The June 2024 NVIDIA crossover was a watershed moment: a chipmaker became more valuable than the company that bought its chips at scale.

NVIDIA held the top position briefly before Microsoft regained it through July 2024. The two companies traded the top spot multiple times through Q3 2024 before NVIDIA decisively pulled away in Q4 2024 to Q1 2025 as data center revenue acceleration overwhelmed Microsoft's Azure trajectory. By April 2026, NVIDIA at $5.06 trillion has 1.61x the market cap of Microsoft at $3.14 trillion. The gap is the largest in technology peer history; previous mega-cap leadership transitions (IBM-Microsoft 1990s, Microsoft-Apple 2010s) saw smaller eventual gaps.

The two companies are linked through capex flows. Microsoft's $110 to $120 billion fiscal 2026 capex is largely spent on NVIDIA accelerators and supporting infrastructure. Estimated $30 to $40 billion of MSFT capex flows directly to NVIDIA. The capex relationship has produced parallel growth: as MSFT capex has expanded from $40 billion in fiscal 2024 to $110+ billion in fiscal 2026, NVIDIA data center revenue has tripled.

The customer-supplier symbiosis has limits. Microsoft has been investing heavily in custom AI silicon (Maia accelerators) to reduce NVIDIA dependence. As of early 2026, Maia accelerators are deployed in Azure for select internal workloads but represent less than 10 percent of MSFT AI compute. Long-term, MSFT's strategic goal is to reduce NVIDIA dependence to 50 percent or less of AI compute. The transition is gradual; markets currently price NVIDIA's dominance through at least 2028.

From November 2022 (ChatGPT release) through April 2026, NVDA gained approximately 540 percent versus MSFT 75 percent. The 465 percentage point outperformance is the cleanest single mega-cap divergence in modern equity history. The path was nearly continuous: NVDA outperformed MSFT in 9 of 11 quarterly windows from Q4 2022 through Q3 2025.

The NVDA/MSFT ratio has expanded from 0.06 in November 2022 (NVDA $14 split-adjusted vs MSFT $235) to 0.50 in April 2026 ($208/$415), an approximately 8x relative move. The current ratio is the highest in NVDA's history versus MSFT. April 2026 has shown some compression (NVDA underperforming MSFT YTD by 5 to 7 percentage points) as AI capex narrative matures, but the structural NVDA outperformance remains the dominant pattern of the AI cycle.

NVIDIA reports quarterly with fiscal year ending in January (calendar Q4 fiscal year-end). MSFT reports with fiscal year ending in June (calendar Q2 fiscal year-end). The different fiscal years create offset earnings releases that produce sequential comparison points. NVIDIA Q1 fiscal 2027 (calendar Q1 2026 results) releases in late May 2026; MSFT Q3 fiscal 2026 (also calendar Q1 2026 results) releases April 30, 2026.

The pair therefore offers two distinct quarterly catalysts within roughly 4 weeks. The MSFT release first will reveal Azure growth and capex translation; the NVDA release shortly after will reveal whether the data center revenue translates the AI demand visible in Azure into NVDA-specific top-line growth. The April 30 to May 30, 2026 window will be one of the most consequential AI cycle data points in 2026.

Three NVDA-specific factors produce moves disconnected from MSFT. First, NVDA earnings releases: each quarterly release moves NVDA 5 to 15 percent typically with limited MSFT response. Second, China export restrictions: US export controls on NVIDIA H100/H200/B200 sales to China produced 5 to 10 percent NVDA-specific declines in Q4 2023, Q3 2024, and Q4 2024. Third, AMD competitive announcements: each major AMD MI300/MI325/MI350 product release produces 2 to 4 percent NVDA compression.

Three MSFT-specific factors produce moves. First, Azure quarterly prints: each cloud growth report moves MSFT 3 to 6 percent. Second, capex announcements: Q1 2026 capex shock drove 5 percent MSFT decline. Third, OpenAI partnership news: any developments in MSFT-OpenAI structure produce MSFT-specific reactions. The April 2026 environment has been dominated by NVDA Q4 fiscal 2026 underwhelm and MSFT capex concerns, producing tighter relative performance than the broader 2024-2025 NVDA dominance trend.

The 2024 to 2025 NVDA-vs-MSFT outperformance was nearly uninterrupted. NVDA reached approximately 30 percent of MSFT market cap by mid-2024 then expanded to 40 percent by year-end 2024 and 50 percent by mid-2025. The October 2025 NVDA peak at $126,198 (post-split equivalent $190) saw the ratio reach approximately 60 percent.

From October 2025 through April 2026, the ratio has compressed slightly to 50 percent. The compression reflects three factors. First, NVDA forward growth deceleration (Q4 fiscal 2026 +73 percent YoY versus +200+ percent in fiscal 2025). Second, MSFT Azure acceleration (39 percent constant currency in Q2 fiscal 2026) catching up to historical NVDA dominance. Third, broader hyperscaler capex maturation as customers digest the 2024-2025 build-out. The compression suggests NVDA dominance peak may have occurred in October 2025; mean reversion of the ratio toward 30 to 40 percent over multi-year horizons is consistent with how prior mega-cap leadership transitions have evolved.

The most underappreciated structural feature of the NVDA-MSFT pair is the customer-supplier risk transmission. If NVDA loses pricing power (through hyperscaler custom silicon adoption, AMD competition, or AI workload efficiency improvements), MSFT benefits with lower input costs. Conversely, if NVIDIA maintains pricing power, MSFT capex efficiency suffers.

For pair trading: long NVDA / short MSFT works when NVIDIA pricing power is strong (current configuration); short NVDA / long MSFT works when hyperscaler custom silicon (Microsoft Maia, Google TPU, Amazon Trainium) gains share. Markets currently price NVDA holding 70 to 80 percent of hyperscaler AI compute through 2027 to 2028. Any acceleration of custom silicon adoption would re-rate the pair toward MSFT outperformance. Q1 to Q3 2026 hyperscaler commentary will be the key data points; Microsoft's ratio of Maia versus NVDA deployment in Azure is the cleanest leading indicator.

NVIDIA fiscal 2026 free cash flow was approximately $80 billion (revenue $215.9 billion, FCF margin 37 percent). Microsoft fiscal 2026 estimated free cash flow is approximately $70 to $80 billion (revenue trajectory $300+ billion annualized, FCF margin 23 to 27 percent). The two companies have similar absolute FCF generation despite NVDA being 1.61x larger market cap.

The FCF parity reflects MSFT's broader business diversification and lower capex intensity in software products. MSFT's FCF is more diversified across products and customers; NVDA's FCF is concentrated in AI accelerator sales to a handful of hyperscalers. From a quality-of-earnings perspective, MSFT FCF has been more durable historically. NVDA's FCF surge is more recent (fiscal 2024 FCF was approximately $25 billion, growing to $80 billion in fiscal 2026). The FCF growth durability is the central question for NVDA-vs-MSFT trading: if NVDA FCF holds at $80+ billion through 2028, the current valuation gap is justified; if it normalizes lower as competition intensifies, MSFT's more durable cash flow profile becomes more attractive.

For practical use: track the NVDA/MSFT market cap ratio. April 2026 ratio is approximately 1.61x ($5.06T / $3.14T). The ratio peaked above 1.7x in October 2025 and bottomed at 0.06x in November 2022 (NVDA at $14 split-adjusted vs MSFT $235). The 27x range in the ratio over 3.5 years is the most extreme mega-cap valuation move in modern history.

For pair trading: long NVDA / short MSFT captures pure AI infrastructure exposure with hedged platform risk. The trade benefits from continued NVDA pricing power, hyperscaler capex acceleration, and limited custom silicon adoption. Short NVDA / long MSFT captures the customer-supplier rotation: benefits from custom silicon adoption, AI capex deceleration, and software monetization phase taking over from infrastructure phase. Position sizing should account for NVDA's higher volatility (38 percent annualized vs MSFT 22 percent). The May 2026 simultaneous earnings releases (MSFT April 30, NVDA late May) are the dominant near-term catalysts.