NVIDIA

NVIDIA enters May 2026 with the most consequential earnings print of its history queued for Tuesday, May 20. CEO Jensen Huang's $1 trillion cumulative-revenue commitment for Blackwell, Blackwell Ultra, and Vera Rubin across calendar years 2025-2027 has been ratified by hyperscaler capex announcements: Microsoft $190B, Alphabet $180-190B, Meta $115B, Amazon $125B. The unique thing about this moment is not the AI thesis — it is that the demand for the moment has been priced in, while the supply ramp has not.

FY2026 closed in January with $215.94 billion in revenue (+65.47% YoY), a non-GAAP gross margin of 75.2%, and $96.58 billion in free cash flow. Management guided Q1 FY2027 to $78.0 billion — implying 77% YoY growth — and explicitly excluded all China data center compute revenue from that figure following the H20 export ban formalized in April 2025. The Vera Rubin platform, claiming 10× inference cost reduction over Blackwell and 4× compute efficiency, is targeted for H2 2026 deployment with AWS, Google Cloud, Azure, OCI, CoreWeave, and Lambda all confirmed as launch partners. The May 20 call is the first opportunity for management to quantify Rubin shipment ramps against the $1T backlog claim — and the first test of whether sequential growth can re-accelerate into the Rubin transition.

Michael Burry's NVDA position is the most-watched bearish bet in 2026 markets. After Scion Asset Management deregistered with the SEC in November 2025, Burry continued to disclose options activity through Substack. His confirmed put structure includes January 2027 $115-strike puts (premium $3.30) and earlier $100-strike puts, both deep out-of-the-money relative to the $215 May 2026 spot. As of May 7, 2026, he extended the bearish complex with new puts on iShares Semiconductor ETF (SOXX), Invesco QQQ, and Oracle. Burry's thesis is structural: he draws the Cisco Systems analogy from the dot-com infrastructure peak, focusing on hyperscaler depreciation accounting that he believes understates true AI workload economics.

On the institutional side, Goldman Sachs reported in April 2026 that hedge funds reduced NVDA exposure at the fastest pace in 13 years. CEO Jensen Huang's Form 4 filings show 64 sale transactions and zero open-market purchases over the past five years, all under pre-set Rule 10b5-1 plans. A July 2025 filing confirmed 225,000 shares sold across three sessions at $159-164. EVP Ajay Puri sold $148M and Director Mark Stevens divested over $100M between December 2025 and March 2026. Total insider sales since January 2024 exceed $3.3B; insider buys: zero. On March 20, 2026, 25 insider sale filings totaling $45.8M cleared in a single day, with two transactions running 2-3× larger than the seller's normal 10b5-1 cadence.

At $215 per share, NVDA trades at a forward P/E of 19.1× consensus FY2027 EPS of $8.23-$8.45 — historically compressed relative to its 10-year average of 61.6× and the lowest forward multiple since the pre-ChatGPT era. AMD trades at 22-24× forward earnings on materially lower growth (low-30s percent versus NVDA's 70%+); Broadcom at 25-27× with AI networking exposure but no full-stack GPU platform; Intel below 20× on a turnaround story with negligible accelerator revenue. The Wall Street consensus across 49 Barchart-tracked analysts averages $268.80, implying 25% upside; MarketBeat's aggregate sits at $275.25, while WallStreetZen's 29-analyst composite is $279.90. Cantor Fitzgerald holds the Street-high at $300; Goldman Sachs and Morgan Stanley each at $250. Of 60 actively covering firms, 57 rate Buy or Strong Buy, 2 Hold, 1 Sell. The compressed multiple reflects either a market priced for Rubin slippage — or an asymmetric setup if the May 20 print confirms execution.

1. May 20, 2026: Q1 FY2027 earnings call — first quarter to fully reflect the ~$8B annualized H20 revenue gap, first opportunity for quantitative Vera Rubin ramp guidance, and the highest near-term volatility event for the stock
2. August 2026 (estimated): Q2 FY2027 earnings — first quarter where Rubin revenue should begin to register meaningfully alongside Blackwell Ultra, validating or invalidating the $1T cumulative-revenue claim against shipped ASP data
3. H2 2026 (Q3/Q4 FY2027): Commercial availability of Vera Rubin-based instances at AWS, Google Cloud, Azure, OCI, CoreWeave, and Lambda — the 10× inference cost-reduction will be the key margin lever for defending mid-70% datacenter gross margins

NVDA at $215 is the cleanest mega-cap setup I track right now. The forward P/E of 19× sits at a tighter discount to the S&P 500 multiple than at any point since the 2022 bear market — and that's against 70%+ revenue growth, $96B free cash flow, and the most-anticipated Q1 print of the year. My add-trigger is a confirmed beat on May 20 with revenue above $80B and non-GAAP gross margin holding above 72%, combined with any quantitative Vera Rubin shipment disclosure. That combination would validate that the China hole is genuinely plugged by Rubin-cycle hyperscaler spend. I would not add ahead of the print given Burry's escalating put complex (now SOXX and QQQ alongside NVDA) and Goldman's institutional de-risking flag. If Q1 misses on either revenue or gross margin, the stock retests the $185-190 zone where I would build a full position. The thesis breaks if Rubin shipments slip into 2027.