What is the five-pillar scorecard?
Three pass, two partial, no fails. The pillars grade the business, not the price. The Q3 print strengthened the three passing pillars and split the two partials, with cash conversion doubling while receivables stretched again. The grades hold, and the score rises on the print and the lower price rather than on a pillar change.
What would change the score. The score rises if Vera Rubin holds the fiscal 2028 growth guide near 70% with gross margin settling at or above 72% and receivables growth back in line with revenue, and it falls if gross margin breaks below 70%, if a top hyperscaler shifts a training generation to its own silicon at scale, or if a guaranteed customer campus fails to reach service.
| 76/100 | DalalBytes score |
| 3 pass | 2 partial, 0 fail |
| $155.46 | Reference price. November 17 2026 close (Finnhub) |
| $145 / $440 / $630 | Fiscal 2029 bear / base / bull |
| Pillar | Grade | The one-line reason |
|---|---|---|
| 1. Sunrise | PASS | Revenue beat the guide by 7.3% and fiscal 2028 growth was raised to about 75% |
| 2. Leadership | PASS | Fifth straight revenue beat and a first call under Kress that held guidance specificity |
| 3. Moat | PASS | CUDA plus a fabric attach that reached a fifth of Data Center and a Rubin ramp at nearly $20B |
| 4. Iron Fortress | PARTIAL | Net cash and record returns, against $390B of purchase commitments and $118B of guarantee caps |
| 5. Free Cash Flow | PARTIAL | Operating cash flow doubled to $48.3B, with receivables up 28% to $81B |
Each pillar below carries its own evidence, with filed numbers carrying the grades and management claims labeled as claims. This scorecard is struck on the fiscal Q3 2027 print of November 17, 2026 and the call that followed. The prior scorecard of November 2 graded the same three pass and two partial at 72/100, and the score now stands at 76/100.
Pillar 1, Sunrise: why does it pass?
The sunrise test asks one question. Is the industry itself rising?
- Total revenue grew 20% sequentially in Q3 fiscal 2027 to $115.8B and beat the $108.0B guide by 7.3%.
- Data Center revenue reached $107.6B, and Vera Rubin contributed nearly $20B of it in its first ramp quarter.
- Management raised the fiscal 2028 revenue growth frame to about 75% and still describes demand as outpacing supply.
- The constraint set is memory and packaging, not orders. Purchase commitments of $390B are the tide measured as purchase orders.
- Every large buyer is expanding capacity into 2027, and sovereign programs in Korea and Japan are building national AI infrastructure on NVIDIA platforms.
- The tide extends below the data center. Edge Computing revenue was a derived $8.2B in Q3, up 14% sequentially.
- A sunrise this strong forgives many execution sins. NVIDIA's sins, the margin path and the ecosystem pledges, are graded where they belong, under Iron Fortress and Free Cash Flow.
- The grade is PASS.
Pillar 2, Leadership: why does it pass?
- Jensen Huang founded the company in 1993 and led it for 33 years. The platform strategy being executed now was set years before the demand arrived, and the institution he built is the asset this pillar now grades.
- Colette Kress leads the company as interim chief executive while keeping the CFO role. Her first call carried the full print, and she answered all eight analyst questions as primary answerer.
- Q3 fiscal 2027 revenue was guided at $108.0B and delivered at $115.8B, the fifth straight revenue guide beat and the largest of the five at 7.3% over the midpoint.
- Gross margin beat its guide for the first time in the five, printing 74.8% GAAP against 74.0% guided, on better product mix.
- The Q4 guide stayed specific at $130.0B of revenue and a 71% to 72% gross margin bottom. The fiscal 2028 growth frame was raised to about 75% on the same call.
- The annual platform cadence has held since Blackwell, and Vera Rubin moved from unveiling in February 2026 to nearly $20B of revenue in the October 2026 quarter.
- Q3 buybacks reached $22B and total capital returns hit a record $28.1B. Management committed to returning all excess free cash flow net of strategic uses.
- Stock-based compensation was brought inside non-GAAP results from Q1 fiscal 2027, an accountability upgrade few companies volunteer.
- The deduction is now succession itself. The leadership is interim, no permanent chief executive has been named, and one strong call is first evidence rather than a settled record.
- The grade is PASS on the delivery record, with the first call under new leadership counted as evidence for the institution rather than against it.
Pillar 3, Moat: why does it pass?
- CUDA is the base layer, two decades of libraries and developer habit that make NVIDIA the default target for new architectures.
- Networking revenue rose from $13.0B in fiscal 2025 to $31.4B in fiscal 2026, and networking reached 19.7% of Data Center revenue in Q1 fiscal 2027.
- A customer who buys the NVLink fabric has bought the architecture. The fabric is the switching cost, and it is now a fifth of the platform sale.
- Revenue opportunity per gigawatt rises from about $18B on Hopper to about $40B on Vera Rubin, per management, because the platform sale keeps absorbing more of the building.
- Vera Rubin reached nearly $20B of Data Center revenue in its first ramp quarter, just under a fifth of the line. The apex product is ramping at every major cloud at once.

- NVIDIA holds about 75% of AI accelerator revenue on 2026 estimates and about 86% of the three-vendor data center revenue set, per the DalalBytes compute research.
- The moat is narrowest at the base of the workload pyramid, where inference and fine-tuning run on TPU, Trainium and AMD at an adaptation cost.
- The apex, frontier training where time-to-model is the product, remains NVIDIA's by merit, with Vera Rubin racks live at every major cloud.
- A moat that holds the apex and taxes the base passes. The networking attach rate is the line that would show widening or narrowing first.
- The grade is PASS.
Pillar 4, Iron Fortress: why only partial?
- Cash and marketable debt securities were $56.6B at July 26, 2026, against total debt of $33.4B. The company holds about $23B of net cash on liquid securities alone.
- Shareholders' equity was $229.0B, and first half free cash flow was $69.9B. Survival in a bad year is not in question.
- The July filing carried $366B of total future commitments, led by $279B of supply and capacity commitments phased across fiscal 2027 to 2029.
- The Q3 update lifted purchase commitments to $390B with named new supply agreements.
- A further $36B of AI cloud agreements and $20B of leases signed for third parties sit below the commitment table.
- Guarantees were stated at $118B of maximum exposure as a total only in Q3, up from $108.5B in August, including $105B behind the SB Energy campus in Ohio that will host NVIDIA infrastructure under 20-year leases to OpenAI, phasing in from fiscal 2029 as sites reach service.
- New commentary language makes neocloud commitments take effect only if the neoclouds secure their own financing. That condition cuts the other way and is recorded in the company's favor.
- Q3 capital returns hit a record $28.1B and remaining repurchase authorization stood at $227B at quarter end. The fortress is funding returns and pledges in the same quarter.
- The equity book stands at $94.0B, concentrated in the ecosystem the company sells to, including roughly $21B in SpaceX and stakes in CoreWeave, Nebius and Intel.
- The fortress test is survival, and this fortress survives. The grade turns on direction. Each quarter a larger share of the strongest balance sheet in the industry is pledged to the buildout it supplies, and the pledges run for twenty years.
- The grade is PARTIAL, the only partial in the library that is caused by strength being spent rather than by weakness.
Pillar 5, Free Cash Flow: why only partial?
- Free cash flow was $27.0B in fiscal 2024, $60.9B in fiscal 2025 and $96.7B in fiscal 2026, a conversion record few companies in history match.
- The first half of fiscal 2027 produced $69.9B more. The engine is not in question.
- Q3 operating cash flow doubled from the Q2 figure to $48.3B and reached 74% of GAAP net income of $65.3B. The conversion split repaired on the cash line.
- The receivables line did not repair. Accounts receivable rose 28% in Q3 to $81.0B against revenue growth of 20%, with days sales outstanding up from 60 to 64 days, on extended terms for large multi-quarter agreements with investment-grade customers.
- Management attributed the rise to Rubin shipment timing and to those extended terms. Two direct customers account for 18% and 17% of the balance.
- Inventory rose to $31.6B from $21.4B at fiscal year end ahead of the Vera Rubin introduction, and the Q3 ramp validated the build.
- Cash conversion and receivables now point in opposite directions in the same quarter. Both things are true, and the grade follows the line that failed its preset gate.
- A cash pillar that finances its customers' payment schedules while its own guarantees phase in does not pass on the strength of its trailing total.
- The upgrade path is stated in the scorecard. Two straight quarters of receivables growth in line with revenue restores the pass, and the count restarts from Q4.
- The grade is PARTIAL.
What do the pillars miss?
The pillars grade the business, and on that test NVIDIA is the strongest operating company in this library. What they cannot grade is the calendar. The sunrise, the moat and the leadership record are all measured at the top of a pricing cycle the company's own guidance says is turning, and the two partials are both forward-looking instruments, guarantees that phase in from fiscal 2029 and receivables that grade the demand's credit quality rather than its size.
That is why the investment report, not this scorecard, carries the final number. The business earns three passes against the strongest field it has ever faced. The price no longer capitalizes a 75% gross margin as a steady state, because it fell 24% into the print. The print and the price together earn the 76/100 Positive that goes with them, up from 72/100 on November 2.
Frequently asked questions
Why does Iron Fortress grade only partial at a company with $56.6B of cash and securities?
Because the fortress is being spent as ecosystem ammunition. Cash and marketable debt securities of $56.6B at July 2026 exceeded $33.4B of debt. The Q3 update then lifted purchase commitments to $390B and guarantee caps to $118B. Survival is not in question. The direction of the obligations is.
Why does Free Cash Flow grade partial when conversion just doubled?
The quarter itself converted well and the balance still stretched. Operating cash flow doubled to $48.3B in Q3 fiscal 2027 and reached 74% of net income. Receivables still rose 28% in the quarter to $81B against revenue growth of 20% and days sales outstanding reached 64 days.
What would change the pillar grades?
Iron Fortress upgrades if guarantee exposure phases in against campuses that reach service and begin paying. Free Cash Flow upgrades when receivables growth returns in line with revenue for two straight quarters. Both partials fall if gross margin breaks below 70% or a guaranteed campus stalls.
Sources and disclosures
- NVIDIA press release, Financial Results for Second Quarter Fiscal 2027, August 26, 2026, for the statements, balance sheet and cash flows behind the pillar evidence.
- NVIDIA CFO Commentary on Second Quarter Fiscal 2027 Results, Form 8-K exhibit, August 26, 2026, for platform revenue, commitments, guarantees and working capital commentary.
- NVIDIA press release, Financial Results for Fourth Quarter and Fiscal 2026, February 25, 2026, for the fiscal 2025 and 2026 history.
- NVIDIA Q2 fiscal 2027 earnings call reporting of August 26, 2026, for the fiscal 2028 growth guide and the margin path.
- NVIDIA Financial Results for Third Quarter Fiscal 2027, November 17, 2026, and the earnings call that followed, for the Q3 print, the raised fiscal 2028 frame, the receivables line and the commitments update, as graded in the DalalBytes Q3 preview update of November 17, 2026.
- DalalBytes NVDA fiscal Q3 2027 print gate sheet, November 13, 2026, for the preset gates applied to the print.
- Finnhub, for the Tuesday November 17, 2026 closing price of $155.46 used as the reference price.
- The Compute Share Wars and The Compute Price Reset, DalalBytes, November 2026, for accelerator share and the custom silicon curve.
- The companion investment report carries the full financial statements, the margin analysis, the roadmap audit and the fiscal 2029 scenarios behind this scorecard.
- Disclosures. DalalBytes holds no position in NVIDIA (NVDA) at publication. This analysis is research and opinion for educational purposes only and is not investment advice, a recommendation, or a solicitation. Prices and estimates are as of November 18, 2026.