COMPLETEDTier 1earnings-single

NVIDIA Q1 FY2027 Earnings — Blackwell Ramp Validation + Vera Rubin Supply Chain Test

Name: NVIDIA Q1 FY2027 Earnings — Blackwell Ramp Validation + Vera Rubin Supply Chain Test
Start: 2026-05-20
End: 2026-05-20

2026-05-20 → 2026-05-20·16 companies·10 predictions·Brier: 0.165

?NVDAdirect?TSMsupply-chain?SKHynixsupply-chain?MUsupply-chain?AMDcompetitive?AVGOcompetitive?SMCIsupply-chain?DELLsupply-chain?ANETcompetitive—MSFTsupply-chain—METAsupply-chain—AMZNsupply-chain?MRVLsupply-chain?VRTsupply-chain?COHRsupply-chain?CRWVsupply-chain

Scenario Comparison

Bull CaseS1

20%

Q1 revenue exceeds $82B (5%+ above $78B guide), Q2 FY2027 guided above $87B. Normalized GM ≥75%. Vera Rubin NVL72 cloud availability confirmed on track H2 2026. Networking revenue >$9B.

↑ 9↓ 211 companies

+NVDAstrong

+TSMmoderate

+SKHynixmoderate

+MUmoderate

+7 more

Base CaseS5

60%

Management addresses Vera Rubin production volume during call — either confirms KeyBanc's 25% cut (1.5M vs 2M target) or rebuts with higher volume targets. HBM4 supply commentary from Jensen/Colette.

↑ 1↓ 34 companies

−NVDAmoderate

−TSMmild

−SKHynixmoderate

+AMDmild

Bear CaseS4

15%

Q1 revenue <$76B or Q2 guided <$80B. GM <73%. Management acknowledges Vera Rubin ramp slower than planned. Customer capex commentary cautious. Export control impact larger than modeled.

↑ 1↓ 78 companies

−NVDAstrong

−TSMmoderate

−SKHynixstrong

−MUstrong

+4 more

All Scenarios

Positioning Suggestions

→ At ~$185 (18.5x CY2027E), NVDA is cheap on forward earnings IF growth sustains. Pre-earnings: hold full weight — risk-reward favors staying positioned given 8+ quarters of consecutive beats

→ Asymmetric setup: if Q1 beats and Q2 guide >$87B, expect re-rating toward $210-220. If Q1 misses (15% probability), $150-165 provides a strong reload zone on bear scenario

→ Consider hedging via AMD: if NVDA beats, AMD likely flat-to-down (competitive window narrows). If NVDA misses, AMD rallies on MI450 narrative. NVDA long / AMD short as pairs trade

→ Monitor SMCI as the highest-beta NVDA derivative — SMCI will move 1.5-2x NVDA's percentage move on earnings. If NVDA guidance strong, SMCI is the leveraged play

→ Post-earnings: if GM compresses below 74%, initiate thesis review — HBM cost absorption is the leading indicator of sustainable pricing power erosion vs temporary mix shift

→ VRT/COHR as infrastructure proxies: if NVDA confirms strong data center demand, infrastructure names lag by 1-2 quarters — consider adding VRT if not already positioned

Predictions

✓

NVIDIA Q1 FY2027 revenue meets or exceeds $78B management guidance

85%P1

If missed → bear trigger fires; trim position 10-15%, shift bear probability 15→25%

✗

NVIDIA Q1 FY2027 revenue exceeds $82B (5%+ above $78B guidance)

25%P2

If confirmed → acceleration validated; consider adding on pullback, raise base case EPS to $10.5+

✓

Q2 FY2027 revenue guidance exceeds $85B (sustaining 8%+ sequential growth)

40%P3

If confirmed → demand acceleration intact; bull scenario probability increases. If <$83B → deceleration, reassess

✓

Normalized gross margin (ex write-downs) reported at or above 75% for Q1 FY2027

65%P4

If below 74% for 2 consecutive quarters → pricing power erosion thesis activates; reduce conviction

✓

Networking/interconnect revenue exceeds $9B in Q1 FY2027 (above Q3 FY2026's $8.2B record)

45%P5

If confirmed → platform moat thesis strengthened; networking growing faster than compute validates vertical integration

✓

Management confirms Vera Rubin NVL72 shipping on schedule for H2 2026

75%P6

If delayed → Rubin bear trigger partially fires; AMD MI450 competitive window extends 1-2 quarters

✓

Jensen Huang or Colette Kress addresses Vera Rubin production volume targets (confirms or rebuts KeyBanc 25% cut report)

50%P7

If confirmed → supply-constrained narrative shifts to execution gap; downward revise H2 2026 Rubin revenue contribution

✗

Groq 3 LPU receives specific revenue or shipment commentary (not just roadmap mention)

20%P8

If disclosed → inference hardware category validated; $20B Groq acquisition starts generating measurable returns

✗

Management quantifies China export control revenue impact at >$8B annual run-rate foregone

35%P9

If quantified → clean read on non-China organic demand; removes an ambiguity that has weighed on sentiment

✗

NVDA stock moves more than 7% in either direction within 2 trading days of the earnings report

55%P10

If large move down → evaluate SMCI/VRT sympathy weakness as entry points. If large move up → hold, do not chase

Key Questions

Does Q1 revenue beat the $78B guide by enough to move the stock? Consensus at ~$78.8B means a modest beat is already priced in — need >$80B for a positive reaction
What is Q2 FY2027 guidance? This is the most important number — sequential growth rate from Q1→Q2 reveals whether Blackwell→Vera Rubin transition creates a demand pause or sustains acceleration
Does normalized gross margin hold at 75%? Any commentary on HBM4 cost trajectory — are suppliers raising prices 20-30% as analysts estimate?
Will management address the KeyBanc report of 25% Vera Rubin production cut? If so, is it confirmed, rebutted, or reframed?
What is networking revenue? Is it sustaining >$8B/quarter and growing faster than GPU revenue? Rising attach rate = platform moat deepening
Does Groq 3 LPU get any revenue or shipment commentary? First quarter it could appear in results — inference hardware category validation
How does management characterize the $1T order book? Firm POs vs demand projections? Conversion rate vs implied backlog drawdown?
Any update on CoWoS-L 4x reticle capacity — is TSMC hitting the 130-150K wafer/month target by end 2026?

Monitoring Checklist

Post-Event Results

Brier Score0.165

correct6/10

Trade Actions7

Actual Outcomes

• Q1 FY2027 reported after close on May 20, 2026: revenue $81.6B (+85% YoY, +20% QoQ) — beat $78.8B consensus by 3.5% and $78B guide by 4.6%, but landed just below the $82B (5%+) headline threshold.
• Data Center revenue $75.2B (+92% YoY) — beat $73.13B consensus.
• Net income $58.3B vs $42.9B consensus (~+36% above expectations).
• GAAP EPS $2.39 / non-GAAP EPS $1.87 vs $1.77 adj consensus.
• GAAP gross margin 74.9% / non-GAAP 75.0% — meets 75% guide; HBM4/CoWoS cost pressure NOT showing yet.
• Networking revenue $14.8B (+199% YoY, +35% QoQ) — massive blowout vs $9B prediction threshold and Q3 FY26's $8.2B record. Vera Rubin NVL72 bundled networking is the structural driver.
• Q2 FY2027 guidance: $91.0B ± 2% — way above $87B threshold and $87.36B consensus; non-GAAP GM held at 75.0% ± 50bps.
• Shareholder return: $80B additional buyback authorization + quarterly dividend raised from $0.01 to $0.25 (25x increase).
• Vera Rubin shipping: confirmed H2 FY2027 (H2 CY2026); Rubin CPX (8 exaflops) end-2026, Rubin Ultra (15 exaflops FP4) H2 2027.
• Jensen on Vera Rubin production volume: 'my sense is that we'll be supply-constrained through the entire life of Vera Rubin' — reframed KeyBanc's 25% cut as a structural supply constraint feature, not an execution failure. Did NOT confirm or rebut the 1.5M vs 2M number specifically. Emphasized 'all customers excited about Vera.'
• Groq 3 LPU: Huang said 'will be a niche product for some time' — positioning commentary, not revenue or shipment data.
• China: U.S. export controls 'effectively eliminated' Nvidia's ~15% China revenue share; ~10 Chinese firms incl Lenovo cleared for H200 but no deliveries yet. Q1 FY2027 guide explicitly excluded all China DC compute. No NEW quantification on May 20 call beyond what was already disclosed in prior quarters.
• Stock reaction: NVDA opened DOWN ~0.5% on May 21, 'beat but sell' — fourth-straight post-earnings slide. Total 2-day move stayed in low single digits, below the 7% prediction threshold.
• AMD reaction: Hit all-time high $452.50 on May 21 (high $452.55, low $431.62) — paradoxical strength as NVDA earnings boosted AI-trade confidence; Lisa Su's '35% CPU growth each year for next 5 years' commentary added.
• SMCI reaction: ~10% pre-market slump on a Wednesday session (likely 5/20 going into NVDA print or post-report).
• TSM, SKHynix, MU, AVGO, ANET, MRVL, VRT, COHR, CRWV, DELL, MSFT, META, AMZN: specific 5/21-5/22 closes not captured in this review.

Market Reaction

NVDA-0.5%AMD+0.0%

Thesis Updates Needed

NVDA·Bull Scenario 1 essentially actualized (rev $81.6B = 4.6% above guide just short of 5% threshold; Q2 $91B way above $87B; Networking $14.8B massive; Vera Rubin H2 2026 confirmed; $80B buyback + 25x dividend). BUT stock muted -0.5% open + 4th-straight post-earnings slide signals expectations are now at 'beat-but-sell' plateau. Thesis re-rate upward on growth trajectory but DOWNGRADE expected-return assumptions: $1T order book conversion >40% no longer surprises market. Jensen's 'supply-constrained entire life of Vera Rubin' is a doubled-edged statement — bullish for sustained pricing power, bearish for hyperscaler customer satisfaction over time. Risks: Networking growth +199% YoY may not repeat; Groq 3 LPU positioning as 'niche product for some time' walks back $20B acquisition expectations.

scenarioAnalysis.scenariosmonitoring.catalystWatchrisks

TSM·Jensen explicitly framed supply-constrained Vera Rubin entire-life — TSM CoWoS-L 4x reticle capacity IS the bottleneck. Bullish for TSM advanced packaging pricing power and capacity expansion ROIC. Q1 FY2027 $75.2B DC revenue + Q2 $91B guide = TSMC AI/HPC revenue trajectory well-supported through CY2026.

scenarioAnalysis.scenariosmonitoring.catalystWatch

SKHynix·HBM4 supply constraint validated as STRUCTURAL by Jensen ('supply-constrained entire life of Vera Rubin'). SKHynix is the primary HBM4 supplier — pricing power thesis strengthened through CY2027. Vera Rubin H2 CY2026 ship confirms HBM4 ramp timing; no demand pause risk.

scenarioAnalysis.scenariosmonitoring.catalystWatch

MU·MU HBM4 qualification status remains the variable — Jensen acknowledged HBM supply constraint but did not name MU vs SKHynix relative share. Memory cycle bull case strengthened (HBM4 demand pull > supply through Vera Rubin entire life). However, today's MU thesis Hold→Sell rebuild reflects valuation/cycle-peak concerns that NVDA print does not invalidate.

risksmonitoring.catalystWatch

AMD·Paradoxical bullish read-through: AMD hit ATH $452.50 on 5/21 as NVDA earnings boosted AI-trade confidence (NOT competitive pressure). 'Both NVDA and AMD can be right at the same time' frame strengthens. Lisa Su's '35% CPU growth each year for 5 years' commentary on 5/22 added independent catalyst. MI450 competitive window not narrowed by NVDA strength — instead validated.

scenarioAnalysis.triggerConditions

SMCI·Pre-market ~10% slump on a Wednesday (likely 5/20 or post-NVDA day) — even in a NVDA blowout quarter, SMCI traded as the higher-beta NVDA derivative with downside skew. Reinforces SMCI's vulnerability to 'beat-but-sell' pattern. Watch for entry opportunity if SMCI weakness extends below NVDA's percentage decline.

monitoring.catalystWatchrisks

ANET·Mixed signal: NVDA networking $14.8B (+199% YoY) validates AI networking TAM (positive for ANET), BUT shows NVDA's NVLink ecosystem capturing the lion's share of value in AI clusters. Ultra Ethernet competitive narrative still credible but NVDA's networking trajectory deserves direct attention in ANET's competitive risks section.

scenarioAnalysis.scenariosrisks

VRT·AI infrastructure buildout demand confirmed at record levels ($91B Q2 guide implies accelerating customer capex). VRT power/cooling backlog extension catalyst on track. No thesis flip but bull-case probability increases.

monitoring.catalystWatch

CRWV·Neocloud GPU procurement strategy validated by NVDA H2 2026 Vera Rubin schedule confirmation. CoreWeave's NVDA-first partnership well-positioned for the supply-constrained phase Jensen describes.

scenarioAnalysis.scenarios

Trade Recommendations

→ NVDA: Bull Scenario 1 actualized fundamentally but stock muted (-0.5% open, 4th-straight post-earnings slide). The 'beat-but-sell' pattern now signals expectations plateau. HOLD existing positions; do NOT chase above $220. Re-evaluate add triggers only on macro pullback to $180-185 range.

→ AMD: Hit ATH $452.50 paradoxically on NVDA's strength + Lisa Su's 5-year CPU narrative. Continue long; momentum + competitive window both supportive.

→ SKHynix/MU: HBM4 supply constraint structurally validated by Jensen. Memory pricing power thesis intact through CY2027. Long memory bias confirmed (MU's separate Hold→Sell call is valuation-driven, not cycle-driven).

→ TSM: CoWoS-L bottleneck confirmed as Vera Rubin entire-life constraint. Advanced packaging pricing power strengthens; capex ROIC improves. Long TSM remains.

→ ANET: Mixed read — NVDA networking $14.8B reveals NVLink's dominance in AI clusters. Hold rating advisable until Q2 ANET print clarifies Ultra Ethernet traction.

→ SMCI: Pre-market 10% slump despite NVDA beat = high beta downside even on good days. Watch for capitulation entry below $40-45 (depending on prior close) if NVDA derivative selling extends.

→ Cross-sector: The Vera Rubin supply-constrained framing through entire-life is a 2-year HBM/CoWoS bull cycle anchor for the supply chain. Equipment names (AMAT/LRCX/KLAC) implicitly benefited but already priced in following May 14 AMAT raise to >30% CY26.