NVDA trades against a final fair-value range of $223.19-$527.51, with the midpoint set by the accepted valuation synthesis rather than earlier draft model outputs. Fair value range: low $223, high $528, with mid-point at $377.
이 리포트는 아직 번역되지 않았습니다. 번역 대기열이 따라잡으면 몇 분 후에 새로고침하세요.
§1 개요
Composite fair value $377 with high case $528.
Implied upside of 71.9% to fair value.
Moat 9/10 · confidence 80/100 · Hyper-growth.
Trades below fair value with a meaningful cushion to the midpoint.
Fair value
$377
Margin of safety
+41.8%
Confidence
80/100
Moat
9/10
Educational research only - not investment advice, an offer, or a trade instruction. Confirm current data and do your own due diligence before acting.
$219.51Price
Low $223.19
Mid $377.33
High $527.51
NVDA trades against a final fair-value range of $223.19-$527.51, with the midpoint set by the accepted valuation synthesis rather than earlier draft model outputs.
CUDA ecosystem
4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL); retraining costs and workflow dependencies create multi-year switching barriers for AI training
Architecture roadmap
Blackwell → Rubin → post-Rubin generational cadence sustains price/performance premium; competitors are consistently one to two architecture generations behind
Bull thesis
Bull perspective: AI supercycle durability + CUDA lock-in + Blackwell/Rubin cadence sustain 73%+ gross margin and support $450–528 fair value over 24 months
Hyperscalers cut AI infrastructure spending 20–30% in FY2028 after capacity digestion and disappointing near-term AI ROI. Data center revenue plateaus at $280–300B by FY2028 vs $508B internal valuation cross-check. Revenue growth decelerates to single digits and the stock de-rates to a 20–22x mature semiconductor multiple.
FV impact
~$175–200 (-9% to -20% from current price)
Trigger
FY2028–2029
Custom ASIC displaces GPU in AI training workloads
10%· Low
Hyperscaler proprietary chips (TPU, Trainium, Maia) capture 25–30% of AI training compute within three years, compressing NVIDIA ASP and gross margins toward 65%. Pricing power erodes; operating margin falls 1,000bps from the FY2026 peak. The CUDA ecosystem advantage is circumvented rather than replicated.
FV impact
~$150–180 (-18% to -32%)
Trigger
FY2028–2030
Escalating US export controls eliminate remaining China exposure
10%· Low
Further restrictions beyond H20 ban eliminate all advanced AI chip sales to China, reducing addressable market by 15%+ and removing the incremental downside-cycle revenue floor. Geopolitical escalation accelerates customer dual-sourcing mandates globally.
FV impact
~$200–240 (-9% to +9%)
Trigger
6–18 months
Structural gross margin compression to sub-68%
8%· Low
Competitive pricing from AMD MI-series and hyperscaler custom silicon forces ASP reductions. Simultaneously, Blackwell production ramp yields unfavorable mix. Gross margin falls below 68% for two or more consecutive quarters, triggering multiple compression as internal valuation cross-check downgrades the quality premium.
FV impact
~$210–260 (-4% to +18%)
Trigger
FY2027–2028
모니터링할 조기 경보 신호
지표
현재
트리거 임계값
Data center revenue growth falls below 20% YoY for two consecutive quarters without management guidance upgrade
Monitor
Deterioration versus the report thesis
Gross margin declines below 70% for any single quarter without re-acceleration guidance
Monitor
Deterioration versus the report thesis
Hyperscaler aggregate capital expenditure guidance cuts exceeding 20% on any quarterly earnings cycle
Monitor
Deterioration versus the report thesis
Custom ASIC share of documented AI training deployments exceeds 15% based on public cloud provider disclosures
Monitor
Deterioration versus the report thesis
OCF/NI ratio falls below 0.70 or accruals ratio exceeds 0.15 in any trailing-twelve-month period
Monitor
Deterioration versus the report thesis
Inventory write-down exceeding $2B in any single quarter, confirming demand pull-forward hypothesis
Capital allocation should be evaluated against reinvestment needs, balance-sheet strength, and shareholder returns.
§4 경쟁 우위
Moat
Widening ↗
9
/ 10
Wide moat
Wide — 9/10
CUDA ecosystem: 4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL);WideCUDA ecosystem: 4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL); retraining costs and workflow dependencies create multi-year switching barriers for AI training
Architecture roadmap: Blackwell → Rubin → post-Rubin generational cadence sustaiWideArchitecture roadmap: Blackwell → Rubin → post-Rubin generational cadence sustains price/performance premium; competitors are consistently one to two architecture generations behind
Full-stack integration: NVLink, InfiniBand networking, CUDA software, and hardwaWideFull-stack integration: NVLink, InfiniBand networking, CUDA software, and hardware co-design cannot be replicated holistically by point-solution competitors
Sovereign AI demand: government AI mandates (Middle East, India, Europe) add a dWideSovereign AI demand: government AI mandates (Middle East, India, Europe) add a demand layer structurally independent of US hyperscaler capex cycles
해자의 원천
출처
강도
근거
CUDA ecosystem: 4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL);
strong
CUDA ecosystem: 4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL); retraining costs and workflow dependencies create multi-year switching barriers for AI training
Architecture roadmap: Blackwell → Rubin → post-Rubin generational cadence sustains price/performance premium; competitors are consistently one to two architecture generations behind
Full-stack integration: NVLink, InfiniBand networking, CUDA software, and hardwa
strong
Full-stack integration: NVLink, InfiniBand networking, CUDA software, and hardware co-design cannot be replicated holistically by point-solution competitors
Sovereign AI demand: government AI mandates (Middle East, India, Europe) add a d
strong
Sovereign AI demand: government AI mandates (Middle East, India, Europe) add a demand layer structurally independent of US hyperscaler capex cycles
Moat factor audit
Factor
강도
근거
Falsification trigger
intangibles
strong
CUDA ecosystem: 4M+ developer base, deep library stack (cuDNN, TensorRT, NCCL); retraining costs and workflow dependencies create multi-year switching barriers for AI training
Evidence that this structural moat driver no longer protects pricing power, customer retention, or excess returns.
scale economies
strong
Full-stack integration: NVLink, InfiniBand networking, CUDA software, and hardware co-design cannot be replicated holistically by point-solution competitors
Evidence that this structural moat driver no longer protects pricing power, customer retention, or excess returns.
Inference moat materially weaker than training: hyperscaler custom silicon captures increasing share of inference-tier workloads at 30–50% ASP discount
medium
Monitor
Monitor the threat against the report checkpoints.
AMD MI300X/MI400 and Intel Gaudi provide credible 30–50% price-per-flop discount for cost-sensitive training segments
medium
Monitor
Monitor the threat against the report checkpoints.
TSMC advanced node concentration: any manufacturing disruption or geopolitical event propagates directly to NVDA supply without domestic alternative
medium
Monitor
Monitor the threat against the report checkpoints.
§5 산업과 사이클
사이클 단계
Expansion
Semiconductor sector capex/revenue is 4.4% (industry median); NVDA's capex is principally supply chain investment, not PP&E-heavy. TSMC advanced node capacity is the binding constraint through at least FY2027, which structurally supports NVDA pricing power and margin. No late-cycle overinvestment risk is present in NVDA's own cost structure.
Hyperscaler AI procurement is lumpy — orders placed 12–18 months ahead of deployment create visible forward demand that can deteriorate rapidly. If model efficiency improves faster than expected (compute required per inference unit falls), upgrade cycles lengthen and near-term GPU demand contracts. Inventory build of $21.4B (+112% YoY) warrants monitoring as a potential demand pull-forward signal.
Forward earnings is the playbook-designated primary model for hyper_growth archetype (applicability score 100). It anchors on forward EPS $12.6457 (FY2027 internal valuation cross-check), validated by Q1 FY2027 actual GAAP EPS $2.39 annualizing to ~$9.56 (above internal valuation cross-check $8.57). The model uses explicit bull/base/bear EPS growth paths with scenario-weighted present values discounted at Ke moderate 11.55%. Despite 110.2% scenario sensitivity spread, the forward EPS anchor is the most relevant market-pricing lens for a visible hyper-growth compounder with 2-day-old earnings confirmation.
Model contribution bridge
Model
Role
Midpoint
Weight
Contribution
Reason
Forward earnings
Primary
$496.24
45%
$223.31
Reason
Primary anchor per hyper_growth playbook. Weight at lower bound of [0.45-0.65] hint range due to grade-D terminal growth and scenario probability assumptions that amplify the 110.2% scenario sensitivity spread. Forward EPS anchor of $12.6457 is validated by Q1 FY2027 actuals ($2.39 GAAP diluted EPS) and $91B Q2 guidance. The model embeds explicit scenario EPS paths: bull $765.26, base $500.65, bear $218.42.
Discounted earnings
Cross-check
$221.14
15%
$33.17
Reason
Highest reliability score (90) but anchored to trailing SBC-adjusted EPS of $6.52, which is stale relative to the forward EPS of $12.6457 (94% gap flagged by diagnostics). Gordon terminal math implies 13.77x vs the 30x terminal P/E assumption, creating a 54% gap that structurally depresses this model's output. The internal valuation cross-check EPS bridge (Y1 $8.57, Y2 $11.87) partially updates the trajectory but the trailing-EPS anchor still understates current earnings power. Weight at 0.15 (within [0.05-0.20] hint) serves as the conservative DCF bracket.
FCFF DCF
Cross-check
$328.35
20%
$65.67
Reason
Required for transformation-build capex phase (capex/D&A 2.13x, PP&E +65% YoY in Q1). Uses WACC 11.54% and enterprise value bridge, providing independence from EPS-based Ke sensitivity. The current-state FCFF growth-fade methodology (not an explicit revenue-down DCF) computes $328.35 midpoint with 65.6% terminal value concentration. Weight at top of [0.05-0.20] hint range reflects the capex-cycle relevance and the inventory build risk (AR $38.5B, inventory $21.4B) that FCFF captures better than pure EPS models.
PEG-adjusted peer
Cross-check
$314.25
15%
$47.14
Reason
Peer comparability score of 65 and dispersion of 3.68x across the 5-peer set (TSM, AVGO, MU, AMD, INTC) justify cross-check rather than primary status. INTC is a weak comp with negative earnings; TSM has anomalous EV/EBITDA. NVDA PEG of 0.50 vs peer median PEG of 0.71 implies undervaluation relative to growth-adjusted peers. The 25% premium applied in the peer policy is reasonable given NVDA's superior gross margin (74.9% vs ~58% peer median) and growth rate. Weight at 0.15 within [0.10-0.20] hint.
Owner earnings
Floor
$160.87
5%
$8.04
Reason
Explicitly labeled as floor value for high-growth companies with 5Y EPS CAGR well above 15%. The static owner-earnings framework (NI + D&A - maintenance capex at static multiples of 15-50x) materially undervalues a hyper-growth compounder. The 107.69% scenario sensitivity spread confirms the wide range is a model-design artifact, not valuation precision. Weight at top of [0-0.05] hint; provides downside anchor at $74.25 low.
Reverse DCF
Excluded
—
—
—
Reason
No dollar midpoint — this is an implied-growth diagnostic. Implied growth rate of 17.49% at current price $219.51 vs analyst growth baseline 34.86% reveals a 17.37pp growth gap. This confirms the market is pricing in substantial deceleration well below internal valuation cross-check and the analyst's forward path. Used qualitatively to support upside conviction but carries zero arithmetic weight per playbook [0,0] hint.
Dividend discount
Excluded
—
—
—
Reason
Excluded per hyper_growth archetype model applicability matrix. Dividend yield 0.02%, payout ratio 0.61% — dividends are negligible and not a meaningful capital return channel. The $0.63 midpoint is economically meaningless.
Multi-stage moat fade
Excluded
—
—
—
Reason
Excluded per playbook despite producing the highest midpoint ($671.03). The 12-year horizon (7-year Stage I + 5-year Stage II) with moat score 9 produces an optimistic valuation that is informative as a long-duration upside ceiling but should not anchor the composite given the grade-D terminal growth assumption that flows through the model. The bear reviewer's point about splitting training vs inference moat duration (7 vs 3-4 years) further weakens confidence in the undifferentiated moat horizon.
nav affo
Excluded
—
—
—
Reason
Excluded: archetype_not_reit. Not applicable to semiconductor companies.
ev revenue
Excluded
—
—
—
Reason
Excluded by the valuation engine: pre_profit_revenue_shares_or_discount_rate_missing exclusion condition. Not applicable given NVDA has substantial earnings.
residual income
Excluded
—
—
—
Reason
Excluded: archetype_not_financial. Not applicable to semiconductor companies.
Composite midpoint
$377.33
100%
The final valuation synthesis sets a $377.33 midpoint within a $223.19-$527.51 range. The published midpoint reconciles the accepted model weights: Forward Earnings 45% at $496.24; Discounted Earnings 15% at $221.14; Fcff Dcf 20% at $328.35; Peg Adjusted Peer 15% at $314.25; Owner Earnings 5% at $160.87. The range preserves supported upside while preventing any single outlier model or scenario value from controlling the published fair value.
NOPAT plus depreciation and amortization, less capex
Revenue growth
34.9%
Frozen FCFF DCF model input: revenue growth
WACC
11.5%
Frozen FCFF DCF model input: WACC
Terminal growth
4.0%
Frozen FCFF DCF model input: terminal growth
Enterprise value
$7.89T
Present value of explicit FCFF plus present value of terminal value
Net debt / cash bridge
$-67.76B
Debt minus cash; negative value means net cash
Equity value
$7.95T
Enterprise value less net debt
Diluted shares
24.22B
Frozen FCFF DCF model input: shares
Fair value per share
$328.35
Equity value divided by diluted shares
Terminal value concentration
65.6%
Present value of terminal value divided by enterprise value
Derived from accepted synthesis ledger and frozen valuation model inputs. Residual-cash rows are reinvestment-policy stress checks, not standalone FCFF DCF base values.
Metric
FY2027
FY2028
FY2029
FY2030
FY2031
Source
Revenue
$379.60B
$507.26B
$647.41B
$787.45B
$910.52B
Base revenue from the frozen FCFF DCF input, projected by accepted revenue-growth assumptions
Operating income
$235.35B
$314.50B
$401.39B
$488.22B
$564.52B
Revenue forecast multiplied by accepted operating-margin assumptions
Net income
$199.77B
$266.95B
$340.70B
$414.40B
$479.17B
Operating income forecast after accepted tax-rate assumptions
Residual cash after reinvestment policy
$-79.91B
$-106.78B
$-136.28B
$-165.76B
$-191.67B
Net income forecast after accepted reinvestment-rate assumptions; residual cash, not standalone FCFF DCF
Diluted shares
12.11B
6.06B
3.03B
1.51B
756.91M
Frozen diluted-share base projected by accepted share-count assumptions
EPS proxy
$16.50
$44.09
$112.53
$273.74
$633.06
Net income forecast divided by projected diluted shares
AI supercycle extends beyond 2030; sovereign AI + inference scaling add demand vectors
NVDA maintains >70% AI training share; gross margins sustain above 73%
Revenue exceeds $500B FY2028; EPS power reaches $14–15 FY2028; terminal at 32x
Bull case ties to the high end of the final fair-value range and requires better-than-modeled fundamentals or a durable improvement in the market multiple.
Base scenario
Base
Probability
50%
Return
+71.9%
Path target
$377.33
Revenue $380B FY2027 (internal valuation cross-check), decelerating at 6pp/year to 15% by FY2031
Terminal at 30x on FY2031 EPS; 17.5% margin of safety applied
Base case ties to the final fair-value midpoint and assumes the accepted model-weight synthesis remains intact.
Bear scenario
Bear
Probability
25%
Return
+1.7%
Path target
$223.19
AI capex digestion: hyperscaler spend falls 20–25% in FY2028; revenue misses internal valuation cross-check by 20%+
Custom ASIC captures 20% of training workloads; gross margin compresses toward 68%
Multiple compresses to 22x on $8–9 FY2027 EPS; fair value approaches $200
Bear case ties to the low end of the final fair-value range and assumes the key risk drivers in the bear-case section persist.
확률 가중 가격
$376.34
확률 가중 수익률
+71.4%
비대칭성
Probability-weighted scenario value is recomputed from the final bear, base, and bull valuation range so scenario math cannot retain draft fair-value targets.
§10 실적 의사결정 트리
비트
Add only on a guidance-backed beat if NVDA still trades below the refreshed fair-value midpoint; otherwise hold the core position.
트리거: Data center revenue sequential growth — target ≥15% qoq through fy2027; any quarter below 10% qoq triggers base-to-bear reassessment beats the base path, gross margin trajectory — two consecutive quarters below 70% is a structural moat deterioration signal stays intact, and management raises forward demand commentary toward the bull path near $527.51.
A beat matters only if it raises normalized outer-year cash generation, not just near-term multiple enthusiasm. Current price is $219.51 versus midpoint $377.33. Keep position sizing tied to valuation after the guidance reset.
부합
Hold and wait for the next data point unless valuation discipline improves materially.
트리거: Data center revenue sequential growth — target ≥15% qoq through fy2027; any quarter below 10% qoq triggers base-to-bear reassessment lands close to the base path and gross margin trajectory — two consecutive quarters below 70% is a structural moat deterioration signal remains consistent with the published midpoint near $377.33.
An in-line print confirms the base case but does not by itself justify raising exposure. The right move is to keep the thesis on watch and demand either better margin of safety or clearer acceleration.
미스
Reduce or pause new buying until the miss is isolated and the valuation floor is re-underwritten.
트리거: Data center revenue sequential growth — target ≥15% qoq through fy2027; any quarter below 10% qoq triggers base-to-bear reassessment misses the bear-case threshold near $223.19, or accounting/cash-conversion quality deteriorates alongside weaker guidance.
A miss is thesis-relevant when it weakens the moat, margin, or cash-flow assumptions behind the fair-value range. Treat that as a model-reset event rather than a routine volatility event.
Data center revenue sequential growth — target ≥15% QoQ through FY2027; any quarter below 10% QoQ triggers base-to-bear reassessment
Watch
Data center revenue sequential growth — target ≥15% QoQ through FY2027; any quarter below 10% QoQ triggers base-to-bear reassessment
Gross margin trajectory — two consecutive quarters below 70% is a structural moat deterioration signal
Watch
Gross margin trajectory — two consecutive quarters below 70% is a structural moat deterioration signal
Hyperscaler aggregate capex commentary during Q2 2026 earnings season (July–August 2026)
Watch
Hyperscaler aggregate capex commentary during Q2 2026 earnings season (July–August 2026)
Custom ASIC deployment disclosures from Amazon, Google, and Microsoft — quantified share of inference workloads is the key watch item
Watch
Custom ASIC deployment disclosures from Amazon, Google, and Microsoft — quantified share of inference workloads is the key watch item
Inventory turns — $21.4B Q1 FY2027 balance must begin unwinding within two quarters to confirm supply-positioning vs demand pull-forward
Watch
Inventory turns — $21.4B Q1 FY2027 balance must begin unwinding within two quarters to confirm supply-positioning vs demand pull-forward
§11 포지션 관리
포지션 사이징
Confidence 58/100 implies 60–70% of full-conviction sizing. Recommend 2.5–4% portfolio allocation at current levels ($219.51) vs 4–6% at maximum conviction (confidence ≥75). Size up toward full allocation only if Q2 FY2027 results resolve the inventory and margin uncertainty favorably.
확신 · High
Confidence 58/100 implies 60–70% of full-conviction sizing. Recommend 2.5–4% portfolio allocation at current levels ($219.51) vs 4–6% at maximum conviction (confidence ≥75). Size up toward full allocation only if Q2 FY2027 results resolve the inventory and margin uncertainty favorably.
Initiate at current price ($219.51). Add 30–40% more at $185–195 on any broad market correction. Maximum position near $160–170 if stock reaches 10% premium to owner-earnings floor ($160), representing maximum margin of safety scenario.
Initiate at current price ($219.51). Add 30–40% more at $185–195 on any broad market correction. Maximum position near $160–170 if stock reaches 10% premium to owner-earnings floor ($160), representing maximum margin of safety scenario.
청산 트랜치
레벨
행동
Thesis break
Trim 25–33% at $350+ or if forward P/E expands above 35x without corresponding estimate upgrade. Full exit triggers: two consecutive quarters of data center revenue growth below 15% YoY, or gross margin below 68%, or Q3 FY2027 guidance below $95B.
§12 투자자 관점
Bull view
Buy
The final Strong Buy view is anchored to the reconciled $377.33 fair-value midpoint and +71.9% upside/downside, with the debate centered on whether the moat and growth runway continue to justify the accepted model weights.
Base view
Buy
The final Strong Buy view is anchored to the reconciled $377.33 fair-value midpoint and +71.9% upside/downside, with the debate centered on whether the moat and growth runway continue to justify the accepted model weights.
Bear view
Sell
The final Strong Buy view is anchored to the reconciled $377.33 fair-value midpoint and +71.9% upside/downside, with the debate centered on whether the moat and growth runway continue to justify the accepted model weights.
§13 9개 항목 스코어카드
Valuation
9.0/10
1.00x
Management
8.5/10
1.00x
Balance Sheet
8.5/10
1.00x
Profitability
8.5/10
1.00x
Revenue Growth
9.0/10
1.00x
Risk Assessment
7.0/10
1.00x
Competitive Moat
9.0/10
1.00x
Earnings Quality
7.5/10
1.00x
Capital Efficiency
8.5/10
1.00x
Weighted Overall
8.4/10
§14 최종 권고
Action
적극 매수
NVDA is rated Strong Buy at $219.51 versus the reconciled fair value midpoint of $377.33, implying +71.90% upside/downside. Confidence is separately disclosed at 80/100.
At $219.51, the stock is measured against the final $377.33 midpoint and +71.90% upside/downside, so new capital should respect the 15% report margin-of-safety discipline.
Action steps
Maintain the Strong Buy stance while price remains near $219.51 versus the final fair value midpoint of $377.33.
Current price is already inside the 15% margin-of-safety entry zone below $320.73; size the position using the report risk and confidence limits rather than waiting for that same pullback.
Reassess after the next earnings update or if credit, capital, or rate-cycle assumptions change materially.
§15Your notes
Capture your own thesis, position size, and follow-up questions. Stays on this device, only you can see it.
Private notes
Visible only to you. Synced to this device.
0 chars
FAQ
NVDA — frequently asked questions
Based on our latest analysis, NVDA looks meaningfully undervalued. The current price is $220 versus a composite fair-value midpoint of $377 (range $223–$528), which implies roughly 71.9% upside to the midpoint.
Our composite fair-value range for NVDA is $223–$528, with a midpoint of $377. The range is triangulated across multiple valuation models (discounted earnings, forward earnings scenarios, peer multiples, and where applicable owner earnings or reverse DCF) and weighted by reliability for NVDA's archetype.
Our current rating for NVDA is Strong Buy with a confidence score of 80/100. NVDA is rated Strong Buy at $219.51 versus the reconciled fair value midpoint of $377.33, implying +71.90% upside/downside. Confidence is separately disclosed at 80/100. This is research for educational purposes, not personalized investment advice.
The top risks our latest report flags for NVDA are: AI capex cycle peak and digestion; Custom ASIC displaces GPU in AI training workloads; Escalating US export controls eliminate remaining China exposure. The single biggest risk is The biggest risk is that the bear-case drivers materialize: growth slows, margins compress, or competitive pressure reduces the fair-value range.
Our current rating for NVDA is Strong Buy, issued with a confidence score of 80/100 and a moat score of 9/10. The rating reflects the composite fair-value range ($223–$528) versus the current price of $220.
NVDA is classified as a hyper-growth stock. Archetype determines how every downstream parameter — discount rate, terminal growth, deceleration curve, terminal multiple, scenario probability weights, scorecard weights, and which valuation models are prioritized — is calibrated for NVDA.
