For AI assistants & researchers — machine-readable summary of this page
What this page is: Delvantic's full research page for NVIDIA Corporation (NVDA) — AI-driven forensic equity research: mechanical valuation models (DCF, EPV, anchored-PE, scenario) plus three independent AI lenses (Quality / Value / Sentiment). Everything below is rendered server-side; you are not missing content that requires JavaScript. All scores are predictions and research opinions, not financial advice.
Our current read (analysis of 2026-10-08): Designation Watch · Gem Score +43 (−100…+100 Quality+Value blend) · Quality 75 · Value 22 · Sentiment 50 (timing only, not weighted) · Composite fair value $337.14 vs $219.37 at analysis
Page map (sections in order; each card carries a stable
reference-name attribute you can cite):
profile-header/price-overview— company profile, live quote, market capextended-analysis— the core: three AI lens reads with findings, scores, and the analyst memofuture-predictions— our forward price-band predictionsmarket-narrative/ai-findings/gpt-critique— narrative context, cross-model findings, and an adversarial critique of our own analysis (near the end of the document)- Members-only sections (render as login gates for anonymous readers):
price-history,income-trend,key-metrics,financials(statement tables),insider-trading. The analysis above is public; the raw data tables require a free account.
More for machine readers: site briefing at
/llms.txt ·
any ticker resolves at delvantic.com/stock/TICKER ·
raw inputs are public-company filings and market data (via licensed data feeds);
every model, score, lens read, and prediction on this page is Delvantic's own analysis.
NVIDIA Corporation
NVDA NASDAQNVIDIA Corporation designs and manufactures graphics processing units (GPUs), chipsets, and related multimedia software. It operates through two primary segments: Graphics Processing Unit (GPU) and Compute & Networking. The GPU segment includes brands like GeForce for gaming, Quadro for professional designers, Tesla and DGX for AI scientists and data researchers, and GRID for cloud-based visual computing users. The Compute & Networking segment features Tegra processors that integrate GPUs and multi-core CPUs into a single chip, powering mobile gaming supercomputers, entertainment devices, autonomous robots, drones, and vehicles. NVIDIA Corporation creates platforms targeting four key markets: gaming, professional visualization, data centers, and automotive. Its offerings include the NVIDIA DGX AI supercomputer, the NVIDIA DRIVE AI platform for automotive computing, and the GeForce NOW cloud gaming service. Founded in 1993 and headquartered in Santa Clara, California, NVIDIA Corporation plays a pivotal role in advancing high-performance computing, artificial intelligence, and immersive graphics across consumer, enterprise, and industrial applications.
Price Overview
Price History (1 Year)
Revenue & Net Income Trend
| Period | Revenue | Net Income | Net Margin | YoY/QoQ |
|---|
Key Metrics
EPS (Diluted): 4.90
Total Equity: $157.29B
Shares: 24,514,000,000
Total Debt: $11.04B
Cash: $10.61B
EBITDA: $133.23B
Total Debt: $11.04B
Cash: $10.61B
Revenue: $215.94B
Revenue: $215.94B
Revenue: $215.94B
Total Equity: $157.29B
Tax Rate: 15.1%
Equity: $157.29B
Total Debt: $11.04B
Cash: $10.61B
Current Liabilities: $32.16B
Long-Term Debt: $10.04B
Total Debt: $11.04B
Total Equity: $157.29B
Shares: 24,514,000,000
Shares: 24,514,000,000
CapEx: -$6.04B
Shares: 24,514,000,000
Stock Price: $197.01
Net Income: $120.07B
Industry Benchmarks
Income Statement (Annual)
Last updated: Jul 29, 2026 11:19am (70d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Revenue | $26.9B | $27.0B | $60.9B | $130.5B | $215.9B |
| Cost of Revenue | $9.4B | $11.6B | $16.6B | $32.6B | $62.5B |
| Gross Profit | $17.5B | $15.4B | $44.3B | $97.9B | $153.5B |
| Operating Expenses | $7.4B | $11.1B | $11.3B | $16.4B | $23.1B |
| Operating Income | $10.0B | $4.2B | $33.0B | $81.5B | $130.4B |
| Net Income | $9.8B | $4.4B | $29.8B | $72.9B | $120.1B |
| EBITDA | $11.2B | $5.8B | $34.5B | $83.3B | $133.2B |
| EPS | $0.39 | $0.18 | $1.21 | $2.97 | $4.93 |
| EPS (Diluted) | $0.38 | $0.17 | $1.19 | $2.94 | $4.90 |
Balance Sheet (Annual)
Last updated: Jul 23, 2026 7:51pm (76d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Cash & Equivalents | $2.0B | $3.4B | $7.3B | $8.6B | $10.6B |
| Total Current Assets | $28.8B | $23.1B | $44.3B | $80.1B | $125.6B |
| Total Assets | $44.2B | $41.2B | $65.7B | $111.6B | $206.8B |
| Current Liabilities | $4.3B | $6.6B | $10.6B | $18.0B | $32.2B |
| Long-Term Debt | $11.7B | $10.6B | $9.6B | $10.0B | $10.0B |
| Total Liabilities | $17.6B | $19.1B | $22.8B | $32.3B | $49.5B |
| Total Equity | $26.6B | $22.1B | $43.0B | $79.3B | $157.3B |
| Retained Earnings | $16.2B | $10.2B | $29.8B | $68.0B | $147.0B |
Cash Flow (Annual)
Last updated: Jul 29, 2026 11:19am (70d ago)| Metric | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|
| Operating Cash Flow | $9.1B | $5.6B | $28.1B | $64.1B | $102.7B |
| Capital Expenditure | -$976.0M | -$1.8B | -$1.1B | -$3.2B | -$6.0B |
| Free Cash Flow | $8.1B | $3.8B | $27.0B | $60.9B | $96.7B |
| Acquisitions (net) | -$263.0M | -$49.0M | -$83.0M | -$1.0B | -$1.5B |
| Net Debt Issued / (Repaid) | $1.0B | $0 | $1.3B | $1.3B | $0 |
| Dividends Paid | -$399.0M | -$398.0M | -$395.0M | -$834.0M | -$974.0M |
| Stock Buybacks | $0 | -$10.0B | -$9.5B | -$33.7B | -$40.1B |
| Net Change in Cash | $1.1B | $1.4B | $3.9B | $1.3B | $2.0B |
Growth Trends (YoY %)
Last updated: Jul 29, 2026 11:19am (70d ago)| Metric | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|
| Revenue Growth | +0.2% | +125.9% | +114.2% | +65.5% |
| Gross Profit Growth | -12.1% | +188.5% | +120.9% | +56.8% |
| Operating Income Growth | -57.9% | +680.6% | +147.0% | +60.1% |
| Net Income Growth | -55.2% | +581.3% | +144.9% | +64.7% |
| EBITDA Growth | -48.6% | +497.8% | +141.6% | +59.9% |
Dividend History (Last 20)
Last updated: Jul 23, 2026 7:51pm (76d ago)| Date | Dividend | Declaration | Record | Payment |
|---|---|---|---|---|
| 2026-06-04 | $0.25 | — | — | — |
| 2026-03-11 | $0.01 | — | — | — |
| 2025-12-04 | $0.01 | — | — | — |
| 2025-09-11 | $0.01 | — | — | — |
| 2025-06-11 | $0.01 | — | — | — |
| 2025-03-12 | $0.01 | — | — | — |
| 2024-12-05 | $0.01 | — | — | — |
| 2024-09-12 | $0.01 | — | — | — |
| 2024-06-11 | $0.01 | — | — | — |
| 2024-03-05 | $0.00 | — | — | — |
| 2023-12-05 | $0.00 | — | — | — |
| 2023-09-06 | $0.00 | — | — | — |
| 2023-06-07 | $0.00 | — | — | — |
| 2023-03-07 | $0.00 | — | — | — |
| 2022-11-30 | $0.00 | — | — | — |
| 2022-09-07 | $0.00 | — | — | — |
| 2022-06-08 | $0.00 | — | — | — |
| 2022-03-02 | $0.00 | — | — | — |
| 2021-12-01 | $0.00 | — | — | — |
| 2021-08-31 | $0.00 | — | — | — |
Deep Analysis
Risk : Reward — upside vs downside from this company's own quarters
Computed 2026-10-06 02:03A +1σ run of quarters pays -55%; a −1σ run costs 65%. Ratio -0.9:1 (μ 91.6%, σ 52.1% , 16 pairs).
Older method (repeat-worst-quarter): 7.3 : 1
| Case | Growth | Margin | Fair value | vs price ($219.37) |
|---|---|---|---|---|
| Bull — recovery | +164% | 35.0% | $857.94 | +291% |
| Base — stabilizes | +109% | 35.0% | $338.21 | +54% |
| Bear — keeps slipping | +55% | 35.0% | $109.84 | -50% |
| Stress — last quarter repeats | +62% | 35.0% | $131.32 | -40% |
| Upside — a +1σ run of quarters (v2) | +50% | 35.0% | $98.93 | -55% |
| Stress — a −1σ run of quarters (v2) | +40% | 35.0% | $77.13 | -65% |
Narrative Economics
market-narrative step).
Growth Outlook
Analyzed 2026-09-17 16:14The question every valuation on this page silently assumes: is this company likely to grow? Judged forward — the business, its category, the world — against what's already printed.
Claude Reading
The raw quarterly cadence is genuinely staggering: revenue went from $35.1B (Oct-24) to $96.2B (Jul-26) — a 2.7x in seven quarters — with the last four quarters ($57.0B → $68.1B → $81.6B → $96.2B) showing sequential growth of 19%, 20%, 18%. That's not decelerating on any absolute measure; the "moderate revenue confidence / decelerating" tag looks wrong or is picking up second-derivative noise. Net margin oscillates between 56% and 71.5% (the Apr-26 71.5% print is suspiciously high and likely has a tax/one-time item — Q2 snapping back to 62% is the tell). TTM revenue is roughly $302.9B and TTM net income roughly $192.9B, putting the actual TTM P/E closer to ~27x on a $5.17T cap — which matches the canonical tag. That is not a bubble multiple for a business compounding earnings triple-digits YoY with $97B FCF and $10B/$8B cash/debt.
Where I dissent from the synthesis: a $348 "signal-adjusted" fair value (+59%) is a DCF output that requires you to underwrite ~60% FCF CAGR for five years, i.e., revenue to $700–900B by 2031. That is the entire hyperscaler + sovereign + enterprise AI capex budget flowing through one vendor at sustained 70%+ gross margins. I don't buy the precision. The synthesis flag "undervalued +58.9%" and the Market Forces "deeply unfavorable risk/reward" are directly contradictory, and the Thesis Evaluation score of -5 (essentially balanced) is more honest than either. The narrative layer's "fragile durability" is the right frame: the multiple isn't the problem, the denominator sustainability is. At 27x TTM earnings on a company still growing earnings >100% YoY, NVDA is arithmetically cheap; the question is whether 2027 earnings are $220B or $140B when hyperscaler custom silicon (TPU v7, MTIA, Trainium3, Maia) actually ships in volume against Rubin.
The contrarian case a careful skeptic should press: (1) customer concentration — roughly 40%+ of datacenter revenue comes from four hyperscalers who are all funding internal ASIC programs; (2) the Apr-26 71.5% margin is a peak that likely already reflects H200/B200 scarcity pricing that Rubin ramp and competition will erode; (3) insider activity is one-way — every transaction in the last 30 days is a sale or gift, no opens, including a 500k-share gift and 300k-share sale on Sep 2; (4) $6B capex against $103B OCF is astonishingly capital-light and will not stay that way as networking/systems/software require more investment; (5) the FCF CAGR of 65% vs revenue CAGR of 77% already shows working capital and mix headwinds. The market cap is $5.17T — larger than the entire German equity market — and requires the AI capex cycle to not merely persist but accelerate through 2028.
My read: the synthesis is directionally right that NVDA is not egregiously overvalued at 27x TTM with this growth, but the $348 fair value overstates the case by anchoring on trailing hypergrowth. A more defensible fair value range is $240–$290, assuming revenue lands around $380–420B in FY27 (implying deceleration to ~30% growth), margins normalize to 55–60% net, and the multiple compresses to 22–25x forward — giving something like $195–230B in earnings capitalized at 22x = $4.3–5.0T, or roughly $180–$215/share, with an optimistic case getting you to $280. So current $218.7 is fairly valued to modestly undervalued, not +59% undervalued. The stock is a hold-to-accumulate on pullbacks, not a table-pound buy at all-time highs with insiders unloading. Partial agreement with the synthesis direction, strong disagreement with its magnitude.
GPT Reading
The raw numbers are almost absurdly strong, and that matters more than any generalized “too expensive” reflex. On a trailing basis through 2026-07-26, revenue is about $302.98B, up roughly 132% from the prior-year trailing base of about $130.50B, and net income is about $192.88B, up about 165% from roughly $72.88B. This is not just high growth; it is high growth with elite economics. TTM gross margin is 74.7%, operating margin 65.2%, and net margin 63.7%. Even among software monopolies those margins would stand out; for a hardware-centered semiconductor company they are extraordinary. Quarterly revenue has climbed from $35.08B eight quarters ago to $96.22B most recently, a 174% increase, while the latest quarter still grew 106% year over year against an already massive base. The valuation multiples look high in isolation — 27.7x earnings, 17.7x sales, 26.4x EV/EBITDA — but the business has grown into them. A $5.17T market cap against nearly $193B of TTM earnings and roughly $303B of TTM revenue is not obviously irrational if this company can keep compounding at anything close to its current trajectory for even two more years.
What stands out most to me is how little strain there is in the financial model despite the scale. The balance sheet is pristine: $10.61B cash versus $8.47B debt, current ratio 4.59, equity $157.29B. Free cash flow was $96.68B in the last fiscal year with just $6.04B of capex, and given the revenue and earnings ramp since January 2026, underlying cash generation is likely far above that annual figure now. This is a company extracting massive economic rent from a supply-constrained platform position. The jump in quarterly net margin from 56.0% in 2025-10 to 71.5% in 2026-04, then 62.0% in 2026-07, tells you mix and supply dynamics are still dominating normal semiconductor cyclicality. That is why I do not buy the simplistic bear argument that “it’s a chip stock, therefore it deserves a chip-stock multiple.” The data says NVIDIA currently behaves more like an infrastructure tollbooth than a conventional component vendor.
That said, I do not agree with the prior valuation output calling the stock undervalued by nearly 60%. The contradiction is simple: the operating performance is exceptional, but the stock is already capitalized at more than 17x trailing revenue and nearly 27x trailing EBITDA-like earnings power for a company whose customers are among the most sophisticated buyers in the world and are actively trying to reduce dependence on it. To justify a truly cheap label at $218.7, I would want either clearer evidence that revenue can keep compounding above 40% after this current buildout wave, or confidence that 60%+ net margins are structurally durable. The quarterly sequence already hints that growth is decelerating in percentage terms as the base explodes: 106% YoY in the latest quarter is incredible, but it is down from the earlier hypergrowth phase, and at $96B in a single quarter the law of large numbers becomes brutal. My read is that the market is not underestimating NVIDIA’s current earnings power; it is discounting the persistence of these margins and growth rates, which is reasonable.
The best case against my more restrained view is obvious and powerful. If you annualize anything close to the latest quarter, you are looking at a company with a revenue run rate near $385B and earnings power well above $200B, and on that basis 27.7x trailing earnings quickly collapses into something much lower on forward numbers. A bull would also point out that annual revenue went from $60.92B in FY2024 to $130.50B in FY2025 to $215.94B in FY2026, and yet margins expanded rather than cracked, which is the opposite of what a commoditization thesis predicts. ROIC of 69% and ROA of 84% are not normal competitive-market outcomes. If CUDA, systems, networking, and the software stack really make this a platform rather than a chip vendor, then today’s multiple could prove conservative. I weigh that argument seriously; in fact it is why I am not bearish. But I still think the stock price already embeds a lot of that platform interpretation, and the burden of proof from here is on sustaining extraordinary economics at extraordinary scale.
What would change my mind is also straightforward. I would turn more bullish if the next few quarters show revenue still compounding above 50% year over year while gross margin stays above 72% and operating margin above 60%; that would tell me hyperscaler insourcing and pricing pressure are not biting yet, and fair value would move materially above the current price. Conversely, I would turn negative if quarterly revenue growth drops sharply below 30-35% while net margin normalizes into the 40s, because a stock at 17.7x sales cannot absorb both slower growth and lower margins without a large multiple reset. I would also watch cash conversion closely: if free cash flow begins to lag earnings meaningfully as the company has to invest harder to defend the moat, the quality of the story changes. For now, the business is world-class, but at $218.7 the stock looks more fairly priced than mispriced.
Grok Reading
The raw numbers describe a company still in the middle of a historic scale-up, not one that has already rolled over. Trailing-twelve-month revenue through the July 2026 quarter is roughly $303B (96.22 + 81.62 + 68.13 + 57.01), with TTM net income near $193B and a 62–64% net margin band that has held even as the absolute dollar base exploded. Sequential revenue steps of +19%, +20%, and +18% across the last three quarters show no collapse; year-ago comparison for the July quarter is still +106% versus the $46.74B print a year earlier. Free cash flow of $96.7B against only $6B of capex, a 0.15 debt-to-equity ratio, and a current ratio above 4.5 leave the balance sheet essentially unlevered. At a 27.7× trailing P/E, 17.7× sales, and 26.4× EV/EBITDA, the market is paying a growth multiple that would be ordinary for a mid-teens compounder, yet the trailing growth rates (revenue CAGR ~77%, earnings CAGR ~91%) remain multiples of that. The valuation synthesis fair-value band near $337–348 therefore looks directionally correct: the $218.7 price embeds a far more pessimistic trajectory than the cash-flow machine currently on display.
What stands out is the durability of the margin structure at scale. Gross margin sits at 74.7% TTM and operating margin at 65%; the July quarter’s 62% net margin on $96B of revenue is not a one-off spike but continuous with the prior three quarters. ROIC above 69% and ROE that has climbed into triple digits on a TTM basis (even after the noted annual-versus-TTM discrepancy) confirm that incremental dollars are still extraordinarily profitable. The CUDA/platform narrative is still earning its keep in the numbers rather than floating free of them. Insider selling has been heavy in August–September 2026, yet the absolute share counts are small relative to the $5.2T equity base and the classification remains neutral; they do not overturn the operating picture.
The strongest counter-case is that the market is rationally discounting the next phase. From a $300B revenue run-rate, sustaining even 30–40% growth requires an addressable AI infrastructure market that keeps expanding without hyperscaler custom silicon or efficiency gains in model architecture chewing into unit demand. The thesis evaluation’s top bears—custom silicon inflection already visible and AI spend normalizing toward ordinary capex cycles—carry real weight, and the narrative layer itself flags durability as “fragile.” Revenue-confidence notes already flag a decelerating quarterly trend, and a PE that looks cheap on trailing earnings can re-rate violently lower if the next two guides show sequential growth slipping into single digits or gross margins compressing toward 60%. The Market Forces “neutral / unfavorable risk-reward” call and the near-zero thesis score correctly highlight that the easy part of the multiple expansion is behind the stock; any stumble on Blackwell/Rubin ramps or a visible cut in Microsoft/Google/Meta capex would justify the current discount to DCF quickly.
I would flip to a fair-or-overvalued stance if the next two reported quarters show sequential revenue growth below ~8–10% while net margins fall through 55%, or if management’s forward commentary explicitly caps datacenter growth in the mid-teens with rising competitive share loss. Conversely, another pair of 15%+ sequential prints with margins holding above 60% would make the $348 composite look conservative and push conviction higher. Until those prints arrive, the gap between demonstrated earnings power and the $218–220 tape remains the dominant fact.
Qwen Reading local experiment · off-panel
The most glaring contradiction in this dataset is not the valuation multiple, but the sheer velocity of the revenue acceleration against the backdrop of a "decelerating" trend flag. NVIDIA’s revenue has exploded from $26.9B in FY2023 to $215.9B in FY2026, a 77.4% CAGR that is virtually unheard of for a company of this scale. Yet, the secondary signals label the revenue trend as "decelerating," which is a statistical artifact of the base effect; the absolute dollar growth is accelerating, not decelerating. The real story is in the margin expansion: net margins have climbed from 42.6% in Q1 2025 to 71.5% in Q2 2026, before settling at 62% in the most recent quarter. This is not a cyclical chipmaker; this is a software-like business with hardware margins. The $96.7B in free cash flow against a market cap of $5.16T implies a P/FCF of roughly 53x, which looks expensive until you realize the company is generating more cash than most sovereign wealth funds. The "undervalued" verdict from the valuation synthesis, which pegs fair value at $348, is not hyperbole—it is a mechanical output of a model that correctly identifies that the market is pricing in a growth stall that the data does not support. The 27.7x P/E is a lagging indicator; the forward earnings power, driven by the $130B operating income run-rate, suggests the multiple will compress rapidly as the denominator grows.
The strongest case against this bullish read is the "fragile" narrative durability and the insider selling. The market is not stupid; it is pricing in the risk that hyperscalers like Microsoft and Amazon will successfully deploy custom silicon, eroding NVIDIA’s 70%+ gross margin moat. The insider transactions show a cluster of sales in August and September 2026, totaling over 1.1 million shares. While these are likely tax-related or diversification moves, the timing—right as the "custom silicon inflection" bear case gains weight—suggests insiders are hedging against the very risk the market is pricing in. Furthermore, the "Macro Headwinds" signal is a critical counterweight. If global datacenter capex cycles normalize, the 83.4% recent revenue YoY growth will not just slow; it will invert. The 37% discount to DCF fair value cited in the narrative layer is not a "discount" to be bought; it is the market’s rational adjustment for the binary risk of AI efficiency gains reducing the total addressable market for GPUs. The "cult coefficient" is high, but cults are fragile; they require constant reinforcement. If the next two quarters show even a 5% deceleration in datacenter revenue, the narrative collapses, and the 23.3x P/B multiple becomes a trap.
What would change my mind is not a single number, but a specific combination: a quarter where datacenter revenue growth drops below 40% YoY while gross margins compress below 65%. That would confirm the "commoditization" thesis and validate the bear case that the CUDA moat is being bypassed. Conversely, if the next earnings report shows a 10%+ increase in networking revenue (the "recurring software" layer), it would prove the platform lock-in is real, not just a hardware cycle. The current price of $218.7 is a bet on the latter. I am committing to the view that the market is underestimating the stickiness of the CUDA ecosystem and overestimating the speed of custom silicon adoption. The 58.9% upside to the composite fair value is not a bubble; it is the market catching up to a company that has already delivered $120B in net income. The risk is not that the stock is overvalued; the risk is that the AI capex cycle is a one-year blip, not a decade-long infrastructure shift. I am betting on the decade.
Big-3 Panel — where each AI stands
Advanced Analysis Forensic deep-dive · separate lenses
The business trajectory is extraordinary: revenue compounded from $29.7B (2022) to $303.0B (2026), a ~10x in four years, while operating margin expanded from 26.9% to 65.2% and net income scaled from $7.7B to $192.9B. Free cash flow of $127.0B on $303.0B of revenue (42% FCF margin) is elite, and buybacks running 5.9x SBC have held diluted shares essentially flat at ~24.4B - per-share value is being protected, not diluted, even at 2.4% SBC/revenue. Altman Z of 38.5 signals no solvency stress despite the -$10.9B net cash figure (a function of gross debt sitting alongside a much larger investment portfolio not captured in 'liquid cash'). Earnings quality is adequate but not pristine: OCF/NI at 0.93x and accruals at 5.5% of assets are reasonable given hyper-growth working-capital build (receivables and inventory scaling with the AI buildout), and the Beneish M flag at -0.39 is largely mechanical - explosive margin expansion and asset growth trip the model even when the earnings are real. Insider tape is one-directional selling ($794M over 12 months, zero opens) but concentrated in 10b5-1-style scheduled sales by long-tenured execs at a stock that has multi-bagged; not a red flag on its own. The dominant question on durability is customer concentration (hyperscaler capex cycle) and competitive response - business-quality metrics themselves show no cracks.
Verify before trusting this (6)
- Customer concentration disclosure in the 10-K - what % of revenue comes from top 2-4 hyperscalers
- Working capital detail - receivables aging and inventory composition to validate the OCF/NI gap
- Backlog and supply commitments (HBM, CoWoS) - are advance payments and purchase obligations proportional to guided demand
- Whether the 55 insider sales are all under 10b5-1 plans or include discretionary sales
- Debt structure and maturities behind the -$10.9B net cash figure, and offsetting marketable securities balance
- Segment margin disclosure - data center vs gaming vs auto to gauge mix-driven margin sustainability
Price is $219.37 against a composite fair value of $337 and a signal-adjusted FV of $348, implying roughly 55-60% upside. The DCF ($355) and anchored P/E ($398) cluster in the $340-400 zone, while the EPV floor of $27 is a runaway low-end artifact (essentially assumes zero growth on peak-ish earnings) and should be discounted, not averaged in naively. On $303B revenue, 65% operating margins, and $127B FCF, a $5.17T market cap is roughly 40x FCF - rich in isolation, but reasonable for a business compounding at current AI-cycle economics with negligible dilution.
Verify before trusting this (5)
- Datacenter revenue growth trajectory and any guidance on FY27 backlog / Blackwell ramp
- Hyperscaler capex commentary and share of revenue from top-4 customers
- Gross margin guidance - any signs of ASP compression as Blackwell scales
- China / export-control revenue exposure and any one-time inventory or licensing hits
- Software/services (CUDA, AI Enterprise) as a recurring revenue disclosure
NVDA is sitting at the center of the strongest active narrative in the market - platform-monopoly, strong intensity, high cult coefficient - and the last 72 hours delivered exactly the kind of catalyst that narrative feeds on: Jensen Huang publicly guiding chip sales to double in 2027 and teasing Rubin economics per gigawatt. News flow is uniformly on-narrative (AI power crunch, Nebius raising prices, tech names 'built to weather higher rates'), which keeps the reflexive bid alive and pulls momentum-chasers back in. Momentum is already strong_positive with recent trend accelerating above the long-term CAGR, confirming the tape is leaning into the story, not fading it. Against that, the macro backdrop is a genuine crosswind: VIX 17.7 in the upper quartile, S&P off its high, 10y near 5%, and a nascent risk-off print. With a 2.22 beta this name is exactly the profile that gets whipped in a stress tape, and the narrative is explicitly flagged as 'fragile' durability - any crack (hyperscaler custom-silicon headline, capex-cycle wobble, an OpenAI-style spending-pause scare like the Cramer segment hints at) would hit hard. Net, the story pressure and Huang catalyst are currently louder than the macro drag, but the asymmetry is uncomfortable given beta and narrative fragility.
Verify before trusting this (4)
- Whether the Huang 'double in 2027' guide gets echoed by sell-side target revisions or gets faded within a week
- Any hyperscaler custom-silicon (Trainium, TPU, MTIA) headline that cracks the monopoly narrative
- VIX pushing above 20 and S&P extending the pullback - that would activate the beta-2.22 vulnerability
- OpenAI or hyperscaler capex commentary suggesting a spending pause
The world is mid-build of a compute layer it does not yet have: AI training capacity is being provisioned ahead of demonstrated end-demand, and inference is only beginning to scale. NVDA is the toll booth on that build, which is why its growth exceeds its category's by a wide margin. The forward constraints are physical and political, not competitive in the near term — power interconnects, HBM and advanced packaging supply, and export policy set the ceiling for the next 12 months. The genuine structural question is what happens when the build phase matures into an operate phase: spend shifts from training clusters toward cost-per-token inference, where custom silicon economics are most attractive to the very buyers now funding NVDA. That transition doesn't reverse growth, but it compresses the rate materially by years 2-3.
When we made this prediction on Sep 18, 2026, NVDA was $219.67. We expect it to be $278.00 by Mar 2027, and we consider it great value under $195.00. This is an early model (v0.6.0) — the direction is more reliable than the exact price. Made Sep 18, 2026.
Blue is our prediction, starting the day we made it. Grey is a slower route to the same place — the same destination, taking longer. Black is the actual price, so you can see how we are doing.