NVIDIA (NVDA) Research Report

NVDA NVIDIA Corporation 英伟达 NVIDIA エヌビディア NVDA

1. Core conclusions

Decision reference: 🟡 Neutral

The four-dimension total is +2, with a base tier of “Moderately Positive”; 3 risk signals were triggered (a valuation model dimension is overvalued / deeply_overvalued, fundamental growth or cash flow has deteriorated / cash conversion metrics have deteriorated, and broad-market breadth is clearly weak or trading volume is contracting). Under the rules, the rating is lowered by one tier, resulting in “Neutral.”

Here, overvalued / deeply_overvalued refers to the model’s overvalued / deeply overvalued range, not an assessment of company quality or a prediction that the stock price will fall.

NVDA’s main tension at present is that revenue and profitability remain very strong, but cash conversion, operating turnover, and some valuation dimensions have not improved in tandem.

• Fundamentals: The model score is 84.8, with a strong state. Latest-quarter revenue grew 105.85% year over year, but operating profit growth slowed from 147.4% in the previous quarter to 124.1%, and cash conversion also declined.
• Technical: The composite score is 65, with a neutral state; the short-, medium-, and long-term trend submodels are all bullish, but momentum and volume-price confirmation remain neutral.
• Valuation: The composite score is 51.7, with a fair_value state (the model’s middle range, composite score 40–60). This does not mean the stock is cheap; enterprise value valuation and growth-adjusted valuation diverge significantly.
• Market environment: The Nasdaq 100 composite score is 88, with a Bullish state, but contracting trading volume is a constraint.

The reference price is $229.28, dated October 9, 2026. The following sections discuss the company’s operations, current pricing, and market trends separately; they are not conflated into a single conclusion.

2. Fundamentals

2.1 Business growth is primarily driven by data centers

The latest financial report is for the second quarter of fiscal 2027, for the quarter ended July 26, 2026.

The company’s earnings release disclosed data center revenue of US$890 hundred million, up 117% year over year; this accounted for approximately 92.5% of quarterly total revenue (calculated from disclosed data). Therefore, understanding NVDA’s current results hinges on whether data center demand can continue to translate into actual deliveries, not just the news buzz around other products or new businesses.

The company also recently announced an open agent safety platform and an expansion of the CUDA-Q platform. These announcements indicate that the company is expanding its platform capabilities, but available data are insufficient to quantify these projects’ revenue contributions or customer switching costs.

2.2 Revenue and profit growth remain strong

Amounts are in USD 100 million; quarterly financial data are on a GAAP basis, and free cash flow is calculated by the platform as operating cash flow less capital expenditures.

Revenue growth rose from 85.23% in the previous quarter to 105.85%, but operating profit growth slowed from 147.42% to 124.10%. This is a slowdown in profit growth, not a decline in profit or a deterioration in revenue growth.

The following are rolling 12-month (TTM) operating scale and margins:

These margins support the fundamentals model’s strong state, but cannot replace a review of collection quality.

2.3 There is divergence within the fundamentals model

The “deterioration in the growth dimension” should not be interpreted as a decline in revenue growth: year-over-year revenue growth actually accelerated, while year-over-year growth in operating profit, net income, and cash flow slowed, leading the dimension-level trend label to deteriorate.

The main operating efficiency indicators use a comparison of the latest TTM with the TTM for the same period last year:

• Days sales outstanding: 61.43 days → 75.97 days.
• Days inventory outstanding: 109.62 days → 150.20 days.
• Cash conversion cycle: 104.65 days → 154.54 days.
• Operating cash flow / net income: 0.890 → 0.697.

Data-based interpretation: Profit growth has not translated into cash to the same extent, while collections and inventory turnover have slowed. Working capital usage is an explanation that warrants examination, but available data are insufficient to conclude that the cash flow change was entirely caused by any single item.

2.4 Financial safety and operating efficiency should be viewed separately

As of July 26, 2026:

• Current ratio: 4.59x.
• Quick ratio: 3.85x.
• Interest-bearing borrowings: US$33.366 billion, excluding lease liabilities.
• Interest-bearing borrowings / equity: 14.57%.
• Cash: US$22.443 billion; no short-term investment data that could be included in this item were obtained in this review.

The model still rates solvency as strong. Therefore, the current main constraint is operating efficiency and cash conversion, rather than weak solvency as assessed by the model.

Source: Second-quarter fiscal 2027 earnings release, August 26, 2026, and the corresponding SEC financial data.

3. Technical

Composite technical score: 65/100; model state: neutral; coverage: 100%.
The risk score is 50, with a normal risk level; the model did not identify technical divergence.

3.1 Trends are bullish, but the composite state remains neutral

The reference price of $229.28 is below the 5-day and 10-day moving averages, but above the 20-day, 50-day, and 200-day moving averages. This is supplementary context on price structure and does not change the original bullish labels for the three time frames or the overall neutral conclusion.

3.2 Momentum is still insufficient for comprehensive confirmation

• RSI: 52.96, model state normal. No overbought or oversold label.
• MACD: The model signal is bullish; MACD value 3.95, signal line 3.82, histogram value 0.127.
• Stochastic oscillator: K value 36.76, state normal; D value 51.87.
• ADX: 18.23; +DI is 29.75, −DI is 24.52. The model direction is bullish, but ADX strength is limited; direction alone is insufficient to conclude that price action is strongly trending.

3.3 Support from volume-price action and relative strength is limited

Trading volume on the day was approximately 0.78x the 20-day average; the volume-price model still labels it normal, with neutral confirmation, and did not identify distribution.

Relative to SPY and QQQ, the model returned “outperform” for both, but the composite relative strength label remains neutral. In other words, there is some local support, but not enough to raise the composite technical state to bullish.

Main support: Bullish trends across three time frames, bullish MACD, and no model-identified divergence.
Main constraints: Momentum, volume-price confirmation, and relative strength remain neutral.

4. Valuation

Composite score: 51.7/100
Model state: fair_value (the model’s middle range, composite score 40–60)
Coverage: 5/6 dimensions, 83.3%; 13/14 indicators
Calculation date: October 10, 2026; reference market data date: October 9, 2026

The user did not specify peers, so peer-relative valuation was not included in the calculation. This is not a complete six-dimension valuation and does not include a peer comparison selected independently.

4.1 Results by dimension

overvalued indicates the model’s overvalued range; deeply_undervalued indicates the model’s deeply undervalued range. Both describe valuation positioning relative to the model’s reference metrics only, not a buy or sell signal.

Other original multiples: price-to-sales 18.27x, price-to-book 24.18x. Overall, the stock is in the middle range; these multiples should not be interpreted as cheap.

4.2 PEG is unreliable and has been excluded by the model

PEG is 0.2315, but the model explicitly marks it as unreliable because EPS growth is excessively high and may be affected by one-time / non-operating gains or a base effect.

Therefore:

• PEG was not included in the score.
• The low PEG cannot be used as proof that the stock is undervalued.
• The growth-adjusted dimension’s score of 100 comes from the revenue growth-adjusted metric actually included by the model, not PEG.

This also shows that strong historical growth supports growth-adjusted valuation, but does not mean future growth will necessarily continue.

4.3 Meaning and limitations of historical percentiles

The model uses 20 quarters and returns:

• Historical P/E percentile: 60%.
• Historical price-to-sales percentile: 60%.
• Historical price-to-book percentile: 60%.
• Historical enterprise value / EBITDA percentile: 55%.

This is percentile analysis based on the distribution of historical quarterly valuation multiples. The model matches quarter-end closing prices to the financial statements for the quarter without adjusting for the reporting lag, so it is not a strict point-in-time backtest and is not suitable for validating a trading strategy.

Enterprise value / EBITDA uses EBITDA on the Yahoo basis and cannot be recalculated or replaced using the fundamentals module’s EBITDA of “SEC operating profit + depreciation and amortization.”

5. News and catalysts

5.1 SEC filings and company announcements

Filed financial results: The earnings release attached to the August 26, 2026 8-K disclosed second-quarter results and third-quarter guidance. Third-quarter revenue guidance is US$108 billion, plus or minus 2%; both GAAP and non-GAAP gross margin guidance are 74.0%, plus or minus 50 basis points. These are company expectations, not results already achieved.
Source: Earnings 8-K, earnings release.

Company announcements:

• September 28, 2026: Announced an additional US$150 billion increase in share repurchase authorization. Authorization does not mean repurchases have been executed, nor does it mean operating cash flow has grown.
  Share repurchase authorization announcement
• September 28, 2026: Announced an open agent safety platform. Available data do not quantify its revenue contribution.
  Agent safety platform announcement
• October 8, 2026: Announced an investment of US$1 billion to support the advancement of U.S. science over the next five years. The spending schedule and specific financial impact were not obtained.
  Science investment announcement

The above are company statements. No corresponding 8-K was obtained in the recent SEC window for this review, so “company announcement” should not be written as “verified in this SEC filing.”

Insider transactions: 16 filings were parsed in the 90-day window, with no truncation; open-market sales totaled approximately US$968 million, across 25 transactions, 9 of which were marked as 10b5-1 plan transactions. Planned sales, gifts, and tax withholding cannot directly be interpreted as bearish, and insider transactions do not automatically constitute major negative news.

Ownership filings: The July 20, 2026 13G disclosed that NVIDIA held 9.3% of Nebius Group, not that an institution held 9.3% of NVIDIA; it should not be misread as an increase in NVDA shareholder holdings.
The 13G filing

5.2 News leads

1. October 9, 2026: Firmus financing valuation dispute. The summary says investors have concerns about the valuation of this NVIDIA-backed data center developer; this was not verified against a filing in this review and cannot be used to quantify NVDA revenue or losses. Source type: original source unclear; 1 publication channel, 0 independent publishers; content level: summary.
   News item

2. October 10, 2026: Article on “physical AI.” An opinion article from Motley Fool, not a new order, company guidance, or sell-side rating. Source type: single source; 2 publication channels, 1 independent publisher; content level: multi-channel match, with republication not equivalent to independent confirmation.
   Opinion article

3. October 9, 2026: Investment management firm’s quarterly letter discusses AI demand. This is a retelling of an investor view and cannot be treated as an order or revenue fact. Source type: original source unclear; 1 publication channel, 0 independent publishers; content level: summary.
   News item

4. Verification of major single events this week: Supplementary searches found no major acquisition, litigation ruling, or guidance cut that could be independently verified; a media summary was found indicating that Cantor Fitzgerald reaffirmed its rating on October 5, 2026. This is an analyst opinion, not a new company operating fact; the search did not provide the number of independent publishers.

5.3 Analyst views

All of the following are others’ views, not this report’s price target or company guidance.

• Ratings: Finnhub monthly summary as of September 1, 2026, 39 days before the analysis date, and may be lagged. Among 69 analysts, 24 rated the stock “Strong Buy,” 41 “Buy,” 3 “Hold,” and 1 “Sell”; the rating trend was stable. The source audit status is warning, so this cannot be described as the latest rating distribution.
• Price targets: Current Yahoo snapshot, with no statistical date. The mean is $328.72, the median is $315, and the range is $180–515; relative to Finnhub’s October 9, 2026 reference price of $229.28, the platform returns upside of 43.37%. Price target dispersion is wide, and the sample size was not provided.
• Revisions: 31 events in the 90-day statistics, with 12 firms collectively raising price targets 16 times; the number of firms differs from the number of revisions. The latest revision date in this review’s statistics is October 1, 2026.
• Earnings expectations: Current-quarter consensus revenue estimate is approximately US$108.977 billion, with an EPS estimate of approximately $2.4733; this is a current snapshot with no statistical date, and the EPS basis cannot be directly treated as the same basis as the SEC quarterly GAAP figure.

5.4 Catalyst chain

6. Market environment

Benchmark basis: The stock is a Nasdaq 100 constituent, so the Nasdaq 100 (QQQ) is used.

The market model analysis date is October 10, 2026, and the market data series runs through October 9, 2026.

The current composite score consists of trend, momentum, and volatility; breadth and market regime classification are not included in the composite score. Therefore, a bullish composite reading and contracting volume can coexist.

Recent related themes:

• Broad market and inflation: A Reuters headline on October 9, 2026 said the market rose for the week and would then focus on earnings and inflation data; this review has only the headline and does not add specific data or causal claims.
  Report
• Interest rates: A Reuters headline on October 9, 2026 discussed rising yields and pressure in the bond market; the transmission to NVDA valuation is an inference only.
  Report
• AI and chip trading: CNBC discussed large trades in Micron and NVIDIA on October 9, 2026; the summary reviewed here does not provide trade details, so it cannot be used to determine the direction of the flows.
  Report

The number of themes reflects only news volume, not whether the news is positive or negative, and did not change any model score.

7. Cross-validation

Four-dimension direction: positive—fundamentals, market environment; neutral—technical, valuation; negative—none. There are 3 risk signals (a valuation model dimension is overvalued / deeply_overvalued, fundamental growth or cash flow has deteriorated / cash conversion metrics have deteriorated, and broad-market breadth is clearly weak or trading volume is contracting), the same results as in Section 10.

1. Operating growth and profitability support each other. High quarterly revenue growth and high TTM margins are consistent with the fundamentals strong state.
2. Profit and cash conversion diverge. Cash conversion declined from 0.890 to 0.697, while receivables and inventory turnover slowed, indicating that net income alone is insufficient.
3. The individual stock trend and broad-market direction are partly aligned. NVDA’s trends across three time frames are bullish, and the broad market composite is bullish; however, NVDA’s momentum and volume-price readings remain neutral, so the market state cannot be applied directly to the individual stock.
4. There is divergence within valuation. The overall valuation is in the middle range, while the enterprise value dimension is overvalued and the growth-adjusted dimension is deeply undervalued; no single dimension should be selected to represent the whole.
5. The explained EBITDA basis difference is present. EBITDA is US$233.894 billion in the valuation module and US$201.266 billion in the fundamentals module, a difference of 16.2%; the former is from Yahoo, while the latter is SEC operating profit plus depreciation and amortization, and they cannot be mixed.
6. The dates are not synchronized. Financial results, market data, and news have different time characteristics and span 76 days; they cannot be described as operating conditions on the same day, nor can recent news be considered reflected in the July financial results.

8. Main risks

1. Declining cash conversion. TTM operating cash flow / net income fell from 0.890 to 0.697, while quarterly free cash flow declined 55.96% quarter over quarter; if this continues, cash support for profit growth will be constrained.
2. Declining turnover efficiency. Days sales outstanding and days inventory outstanding rose year over year, and the cash conversion cycle lengthened; this is the specific basis for the operating efficiency model’s weak state.
3. Profit growth and revenue growth are not moving in sync. Year-over-year revenue growth accelerated, but year-over-year operating profit growth slowed from 147.4% to 124.1%; changes in the base and ongoing operating factors need to be distinguished.
4. A valuation sub-dimension is in the overvalued range. The enterprise value valuation score is 30, with an overvalued state; the overall middle-range valuation does not eliminate this constraint.
5. A low PEG could be misleading. The model has excluded PEG, which cannot be used to prove the stock is cheap; the growth-adjusted conclusion also depends on growth data already achieved.
6. Insufficient market volume-price confirmation. The broad market’s relative volume is 0.59x, with a contracting-volume state; NVDA’s own volume-price confirmation is also neutral.
7. Information and expectations are uncertain. Price targets are widely dispersed and ratings data are lagged; the Firmus news has not been verified against a filing. Customer investment returns, competition, and policy changes may affect demand, but sufficient material was not obtained in this review to confirm that they have caused specific financial losses; this point is an inference only.

These are narrative risks and do not separately increase the number of risk signals in Section 10.

9. What to watch next

9.1 Event calendar

The date of the next earnings report has not yet been formally confirmed by the company.

The platform estimates it based on historical SEC earnings 8-K dates:

• Forecast date: November 18, 2026.
• Forecast range: November 18–25, 2026.
• Confidence: Medium.
• Typical release timing: After market close.
• Basis: The release date for the same quarter last year and recent intervals between earnings reports.

Supplementary searches found third-party pages listing November 18 and November 17, respectively; the sources are inconsistent, and no announcement from the company for this report was obtained, so the date cannot be upgraded to confirmed.

Therefore, “next earnings report” below refers to the above forecast window, which has not been confirmed by the company.

9.2 Earnings review: Key metrics to track

① Core growth: Look at revenue and operating profit together

Latest figures and change: Quarterly revenue was US$962.21 hundred million, up 105.85% year over year; operating profit was US$637.34 hundred million, up 124.10% year over year. Year-over-year revenue growth accelerated, while operating profit growth slowed from 147.42% in the previous quarter.
Why it matters: Looking at both together helps distinguish whether business expansion continues to translate into operating earnings.
What to look for in the next earnings report: If revenue growth is realized and operating profit growth and margins remain supportive, operating expansion is still translating into profit; if the two continue to diverge, the cause needs to be examined.
Source and label: SEC financial data as of July 26, 2026 and 8-K earnings release, disclosed; third-quarter revenue of US$108 billion, plus or minus 2%, is company guidance.

② Data center revenue

Latest figures and change: Data center revenue was US$890 hundred million, up 117% year over year; approximately 92.5% of quarterly revenue (calculated).
Why it matters: This is currently the main source of revenue, and news about other businesses cannot substitute for this business’s actual growth.
What to look for in the next earnings report: If data center revenue continues to grow and drives realization of total revenue, this would indicate that core demand continues to translate into deliveries; if they diverge, the business mix needs to be examined.
Source and label: August 26, 2026 8-K earnings release, disclosed; financial transmission is an inference.

③ Expense pressure: Growth is currently below revenue growth, but bases must remain consistent

Latest figures and change: GAAP operating expenses were US$84.08 hundred million, compared with US$54.13 hundred million in the same period last year, up 55%; non-GAAP operating expenses were US$82.32 hundred million, compared with US$53.61 hundred million, up 54%. Both were below the earnings release’s 106% revenue growth rate. Among the main GAAP items: cost of revenue was 240.79 / 128.90 hundred million dollars, up 86.8%; depreciation and amortization was 11.27 / 6.68 hundred million dollars, up 68.7%; research and development was 70.54 / 42.91 hundred million dollars, up 64.4%. The growth rates for the latter three items are calculated from the “three months, 2026 vs. 2025” columns.
Why it matters: Expense growth below revenue growth supports operating leverage; the expense items identified above have not currently shown growth faster than revenue.
What to look for in the next earnings report: If expense growth continues to lag revenue growth and gross margin remains stable, margins remain supported; if the relationship reverses, R&D, costs, and product mix should be examined.
Source and label: Expense table in the 8-K earnings release, disclosed; third-quarter GAAP / non-GAAP operating expenses of US$92 / 90 hundred million are company guidance.

④ Remaining performance obligations

Latest figures and change: As of July 26, 2026, remaining performance obligations were US$3.2 billion, up 68.42% year over year; year-over-year growth rose from 44.44% in the previous quarter to 68.42%.
Why it matters: This can help track signed but unfulfilled obligations, but is not equivalent to all hardware orders or overall data center demand.
What to look for in the next earnings report: If growth in remaining performance obligations is consistent with revenue realization, it can provide supporting evidence; if they diverge, the fulfillment cycle and disclosure scope should be examined.
Source and label: SEC financial data obtained by the platform, disclosed; implications for demand are an inference.

⑤ Cash conversion and free cash flow

Latest figures and change: Quarterly free cash flow was US$214 hundred million, compared with US$485.87 hundred million in the previous quarter, down 55.96% quarter over quarter; TTM operating cash flow / net income fell from 0.890 in the same period last year to 0.697.
Why it matters: The review should verify whether rapid growth ultimately turns into cash available for R&D, investment, and capital returns.
What to look for in the next earnings report: If cash conversion recovers and receivables and inventory turnover improve, the current constraint will ease; if cash continues to lag while profit grows, working capital should be examined further.
Source and label: SEC quarterly and TTM cash flow data, disclosed; the cause has not been fully confirmed.

10. Decision reference

10.1 Four-dimension assessment

• Fundamentals: +1 (model state strong, score 84.8)
• Technical: 0 (model state neutral, score 65)
• Valuation: 0 (model state fair_value, score 51.7)
• Market environment: +1 (model state Bullish, score 88, benchmark nasdaq100)

10.2 Base tier

1 + 0 + 0 + 1 = +2 → Moderately Positive

10.3 Risk adjustment

3 risk signals identified:

• A valuation model dimension is overvalued / deeply_overvalued: enterprise value valuation dimension state is overvalued (score 30)
• Fundamental growth or cash flow has deteriorated / cash conversion metrics have deteriorated: fundamentals growth dimension trend = deteriorating; fundamentals cash_flow dimension trend = deteriorating; cash conversion ocf_to_net_income: 0.89 → 0.697 (trend = deteriorating, TTM year-over-year basis); growth sub-indicator operating_income_yoy: year-over-year operating profit growth: 147.4% → 124.1% (trend = deteriorating, compared with the previous quarter’s year-over-year growth rate)
• Broad-market breadth is clearly weak or trading volume is contracting: volume state = contracting volume

Moderately Positive → lowered one tier to Neutral

10.4 Final model interpretation

Strong fundamentals and a bullish broad market provide support, but both individual-stock technicals and overall valuation are neutral, so the base tier does not reach Positive. The 3 risk signals lower Moderately Positive to Neutral; the main constraints are cash conversion, profit growth trends, enterprise value valuation, and market trading volume.

Decision reference summarizes the current multidimensional state; it is not a buy, sell, or hold recommendation and does not include a price target or position size. Specific decisions should take account of personal holdings, time horizon, and risk tolerance.

11. Data notes

• Sources and dates: The reference price is from Finnhub as of October 9, 2026; technical indicators are from Massive daily data for the same date; the market model is from the Alpaca market data series; fundamentals are from the SEC through July 26, 2026. The valuation TTM statements are marked through July 31, 2026, and the balance sheet uses SEC data as of July 26.
• Valuation date difference: Valuation metadata marks October 10 as the calculation time; it should not be treated as a market trading day. The reference price of $229.28 is consistent with October 9 market data.
• Time-span warning: Modules combine the financial reporting period end, daily close, news publication time, market data time, and valuation date, spanning 76 days. The data can be analyzed together, but they are not a same-day operating snapshot.
• Units and basis: Unless otherwise noted, amounts are in U.S. dollars; quarterly, TTM, and balance sheet snapshots are cited separately. GAAP and non-GAAP expenses are shown separately. P/E is based on price / diluted EPS, while earnings yield is based on net income / market capitalization; one is not replaced with the reciprocal of the other.
• Debt and cash: Interest-bearing borrowings exclude lease liabilities; only cash was obtained for cash-like assets in this review, and short-term investments were not obtained, so net debt should not be interpreted as deducting all current investments.
• Data anomaly: The raw EPS field for the quarter ended January 26, 2025 was anomalous, and the platform reconstructed it using net income and comparable shares. This anomaly is not evidence of company losses; the anomalous raw value is not cited in this report.
• Completeness: Data were obtained for all core models in this review; no dimension failed to retrieve data. SEC financial reports, insider transactions, company announcements, individual-stock news, analysts, and market news were available; no new 8-K or periodic report was obtained in the recent SEC window. Earnings call Q&A was not obtained, so no management Q&A views are presented.
• Limitations: Peer-relative valuation was not included; the ratings summary is lagged; price targets and consensus estimates have no statistical date; the next earnings report date remains a forecast. News sources include aggregated republications and opinion articles, which are not treated as independently confirmed facts.
• Sixth risk signal (major negative news): Not triggered. No verified major negative item directly affecting revenue, profit, or operating qualifications was obtained in this review; the Firmus financing dispute lacks sufficient sourcing and has not been verified against a filing, and this week’s search did not provide a qualifying item. This assessment does not mean that no real-world risks exist.
• Decision reference: decision_id=14c0b84323; rules_version=decision-rules-v1.1.

Data quality validation notes:

• [Explained basis difference] Valuation module EBITDA and fundamentals module EBITDA differ by +16.2%: different bases (valuation module yfinance_ebitda; fundamentals module SEC operating profit + depreciation and amortization)
• [Explained basis difference] D/E uses different units in the two modules: the valuation module is a ratio of 0.1457, while the fundamentals module is 14.57 percent; they are consistent after conversion