Nvidia released its Q3 earnings report on November 19th. While not exactly stellar, the results could be described as exceeding expectations. The problem is that despite such a scorecard, the market wasn't buying it — after an initial 5% rise, the stock began to plunge sharply. Many folks in the crypto world were left completely baffled. This article attempts to compile, interpret, and analyze, from the perspective of the bears, what unspeakable truths lie beneath this seemingly "too good to be true" earnings report.
Additionally, there are already far too many bullish articles out there — I won't bother rehashing those here.
This article is approximately 4,000 words. If you can't be bothered to read the full thing, here are the core bearish points — take them and go 🤣:
Circular financing manufacturing revenue: Nvidia has constructed a capital reflux closed loop by investing in customers like xAI, converting investment funds into its own reported income — lacking substantive cash delivery.
Abnormal surge in accounts receivable: The accounts receivable balance reached $33.4 billion, growing far faster than revenue, with suspicions of obfuscation in the days-outstanding calculation — suggesting serious "channel stuffing" and back-end loading phenomena.
Inventory narrative divergence: Under the narrative of "demand exceeding supply," finished goods inventory unexpectedly doubled — signaling potential risks of customers deferring pickup or product stagnation.
Cash flow inversion: Operating cash flow is significantly below net income, proving that the company's profits are primarily sitting on paper and have not been converted into real cash.
Disclaimer: This article does not constitute any investment advice. This is purely a compilation of perspectives.
1. Circular Revenue and the Vendor Financing Model
1.1. The Closed-Loop Mechanism of Capital Flow
Background: In November 2025, Elon Musk's xAI completed a $20 billion funding round, in which Nvidia directly participated with approximately $2 billion in equity investment. But this is not a simple "investment action." Follow the logic step by step:
Capital Outflow (Investment Side): Nvidia allocates cash (approximately $2 billion) from its balance sheet, recorded under "Purchases of non-marketable equity securities," as an equity injection into xAI or related SPVs. This capital outflow appears under the "investment activities" section of the cash flow statement.
Capital Conversion (Customer Side): xAI receives these funds and uses them as a down payment or capital expenditure budget for purchasing GPU clusters (the Colossus 2 project, involving 100,000 H100/H200 and Blackwell chips).
Capital Reflux (Revenue Side): xAI then issues a purchase order to Nvidia. Nvidia ships the goods and recognizes "Data Center Revenue."
Financial Result: Nvidia has effectively converted the "cash" asset on its own balance sheet into "revenue" and "net income" on its income statement — using xAI as the intermediary.
While this type of operation is generally permissible under GAAP accounting standards (as long as properly assessed), it constitutes a form of "Low-Quality Revenue." (IFRS would like to have a word here 🤣)
This is also what short sellers like Michael Burry criticize — because the pattern of "almost all customers being funded by their own supplier" is a classic hallmark of the late stages of a bubble. When a company's revenue growth depends on its own balance sheet expansion, the moment it stops investing externally, its revenue growth will dry up alongside it. (Does this feel a little like the recursive token structures in crypto?)
1.2. The Leverage Effect and Risk Isolation of SPVs
If the circular revenue model sounds impressive, the Special Purpose Vehicle (SPV) structure involved in the transaction may open your eyes even wider.
According to news reports, xAI's financing included both equity and debt, with the debt portion structured through an SPV whose primary purpose was to purchase Nvidia processors and lease them to xAI.
SPV Operating Logic: The SPV, as a legally independent entity, holds the GPU assets. Nvidia is not only the GPU seller but also the equity investor in the SPV (first-loss capital provider). This means Nvidia plays a dual role in the transaction: supplier and underwriter.
Circular Arbitrage in Revenue Recognition: By selling hardware to the SPV, Nvidia can immediately recognize the full hardware sales revenue. However, from the end user xAI's perspective, this is essentially a long-term operating lease, with cash outflows made in installments (for example, over a 5-year term).
Hidden Risk: This structure converts long-term credit risk (whether xAI can pay rent in the future) into immediate revenue recognition. If AI compute prices collapse in the future, or if xAI fails to generate sufficient cash flow to service the lease, the SPV will face default — and Nvidia, as an SPV equity holder, will face asset write-down risk. But during the current earnings season, all of this manifests as dazzling "genesis-level revenue."
1.3. The Shadow of Vendor Financing from the Internet Bubble Era
The current business model bears some resemblance to the 2000 internet bubble. At that time, Lucent lent billions of dollars to customers to purchase its own equipment. When internet traffic growth fell short of expectations, those startups defaulted, Lucent was forced to write off massive bad debts, and its stock price collapsed 99%.
Nvidia's current risk exposure (direct investment + SPV debt support) is estimated to have already exceeded $110 billion — a significant proportion of its annual revenue. While Nvidia does not currently directly list this as "customer loans" on its balance sheet, through holding customer equity and SPV interests, the substantive risk exposure is effectively identical.
2. The Accounts Receivable Mystery
2.1. Accounts Receivable Growth "Velocity"
According to the Q3 FY2026 earnings report, Nvidia's accounts receivable balance reached $33.4 billion.
The year-over-year growth rate of accounts receivable (224%) is 3.6 times the revenue growth rate (62%). Under normal business logic, accounts receivable should grow in sync with revenue — especially for a company as "dominant" as Nvidia. When accounts receivable grows far faster than revenue, this typically points to two possibilities:
a. Declining revenue quality: The company has relaxed credit terms, allowing customers to defer payment in order to stimulate sales.
b. Channel stuffing: The company rushes shipments to channel partners at the end of the quarter to recognize revenue, but this inventory has not actually been absorbed by the end market. (More on this below.)
2.2. The DSO (Days Sales Outstanding) Calculation
The DSO for this quarter is reported as 53 days, a slight decline from 54 days last quarter. So what does the actual picture look like?
First, the standard DSO formula: DSO = (Accounts Receivable / Total Credit Sales) × Period Days
Beginning AR (end of Q2): $23.065 billion
Ending AR (end of Q3): $33.391 billion
Average AR: $28.228 billion ((Q2 + Q3) / 2)
Quarterly revenue: $57.006 billion
Days: 90
Standard DSO ≈ 28.228 / 57.006 × 90 = 44.57 days
Yet the reported DSO is 53 days. Logically speaking, from a "window dressing" perspective, one would typically report a more "aggressive" (lower) number — so why is this one conservative? This suggests Nvidia may be using ending accounts receivable as the numerator, or that its calculation logic leans toward reflecting end-of-period capital occupation.
Using the ending balance for calculation: 33.391 / 57.006 × 90 = 52.72 days — which aligns with the reported figure.
But what does this mean? It means the end-of-quarter accounts receivable balance is extremely high relative to the entire quarter's sales. This implies a Back-End Loading phenomenon — a large proportion of sales activity is concentrated in the final month or even the final week of the quarter.
If sales were evenly distributed, the ending accounts receivable should only contain approximately the last month's sales (roughly $19 billion). But the current balance is $33.4 billion — meaning nearly 58% of the quarter's revenue has not yet been collected in cash.
In the so-called "seller's market" and "demand exceeding supply" narrative, Nvidia should theoretically possess extremely strong bargaining power — potentially even requiring prepayments. Yet the reality is that Nvidia has not only failed to collect prepayments but has instead extended customers nearly two months of credit terms?! This seems rather inconsistent with the "fighting to get their hands on chips" narrative.
3. The Inventory Puzzle: The Paradox of Demand Exceeding Supply While Inventory Accumulates
While Jensen Huang proclaims "Blackwell demand is off the charts," Nvidia's inventory data appears to be telling a rather different story.
3.1. Why Inventory Has Doubled
Total inventory for Q3 FY2026 reached $19.8 billion — nearly double the $10.0 billion at the beginning of the year, and up 32% from $15.0 billion last quarter.
Even more telling is the inventory composition:
Raw Materials: $4.2 billion
Work in Process (WIP): $8.7 billion
Finished Goods: $6.8 billion
In a word: finished goods inventory has surged dramatically. At the beginning of 2025, finished goods inventory was only $3.2 billion. It has now surged to $6.8 billion. Particularly when Jensen is shouting about explosively off-the-charts demand, under the assumption of chip shortages and customers lining up waiting for orders, finished goods should be "shipped the moment they're produced" — with inventory levels maintained at extremely low levels.
Why, then, is this happening? Waiting until after the New Year to collect payment?
3.2. The $50 Billion Purchase Commitment
Beyond the on-balance-sheet inventory, Nvidia has also disclosed a staggering $50.3 billion in supply-related Purchase Commitments — the amount Nvidia has committed to pay to suppliers like TSMC and Micron for future purchases.
This is a massive hidden liability. If AI demand experiences any slowdown or weakening in coming quarters, Nvidia will face a double blow:
Inventory impairment: The existing $19.8 billion in inventory may depreciate in value.
Default or forced purchasing: The $50 billion in purchase contracts could force even more inventory accumulation, or require enormous penalty payments.
The emergence of this "heavy asset" characteristic signals that Nvidia is no longer the light-asset chip designer it once was — it increasingly resembles a hardware manufacturer burdened with a heavy supply chain.
Production capacity running ahead, inventory chasing from behind — has the soul left the body?
4. Profits Rising, But Cash Flow Declining?
4.1. The Inversion of Operating Cash Flow (OCF) and Net Income
Under normal circumstances, a healthy tech company's operating cash flow should exceed net income (because depreciation, amortization, and stock-based compensation are non-cash expenses that get added back). However, Nvidia's data shows the opposite trend.
Q3 Net Income: $31.9 billion
Working Capital Changes:
Increase in accounts receivable (cash outflow): −$5.58 billion
Increase in inventory (cash outflow): −$4.82 billion
Q3 Operating Cash Flow (OCF): approximately $23.75 billion
Conclusion: Q3 operating cash flow is significantly below net income. For every dollar of profit, only approximately $0.74 has actually converted into cash inflow — the rest has become chips in a warehouse (inventory) and customer IOUs (accounts receivable).
Of course, this OCF < Net Income phenomenon is open to interpretation. It can mean the company's profits are being recognized by accounting standards rather than supported by real cash in bank accounts — or it can mean the company is growing at high speed.
4.2. The Cash Hemorrhage in Investment Activities
Purchases of non-marketable equity securities — colloquially, "investments": $3.7 billion flowed out this quarter.
This $3.7 billion is precisely what flowed to "ecosystem partners" like xAI, CoreWeave, and Hugging Face. By comparison, the same figure a year ago was only $473 million. Nvidia is increasing its buyout of ecosystem participants at SpaceX-like velocity.
Based on the xAI model, the investment flow may look like this:
Nvidia accumulates cash through bond issuance or previous profits.
Cash is invested into startups (cash outflow).
Startups use the money to buy chips (recognized as Nvidia revenue).
Nvidia's paper profits increase, stock price rises, talent is attracted via stock-based compensation, further financing via bond issuance or share offerings.
(Does this feel a bit like the recursive token structures in crypto?)
If this is truly the model, it has a bit of a musical chairs quality to it. As long as the music doesn't stop, the game can continue indefinitely. But once the financing environment tightens (such as rising interest rates or an AI bubble bursting), this game could freeze instantaneously.
5. Nvidia's Dominance Is Not Sacred and Inviolable
In its 10-Q filing, Nvidia disclosed an extremely high customer concentration, with "Customer A" accounting for 22% of revenue. While unnamed, there are only so many companies on this planet with the financial firepower to qualify — it is almost certainly, definitively, 100% Microsoft.
There is another layer of related-party transaction risk hidden here. Microsoft is OpenAI's largest backer, and Nvidia has also invested in OpenAI. A significant portion of Microsoft's GPU purchases from Nvidia are presumably provided for OpenAI's use. Both are shareholders of OpenAI — the revenue-sharing arrangements within that structure are unknown. Who knows whether there are any special clauses involved?
And what if Microsoft were to walk away from purchases tomorrow?
Furthermore, the existence of Customer B (15%), Customer C (13%), and Customer D (11%) means that the top four customers collectively hold Nvidia's lifeline. This level of concentration means Nvidia does not possess the absolute dominant pricing power that the outside world imagines. On the contrary, these giants are leveraging their enormous purchase volumes to force concessions from Nvidia on supply chain allocation and custom chip design — and are accelerating development of in-house chips (such as Google TPU, AWS Trainium, and Meta MTIA) to reduce their dependence on Nvidia. This is also faintly visible in the rising accounts receivable figures.
The diagram below gives you a visual sense of the complex structure of the OpenAI cluster — just ask yourself: can you even make sense of all these accounts?
Afterword
Why write this article? First, it's been a long time since I wrote this type of long-form financial statement analysis, and I wanted to revisit the feeling of poring over a balance sheet.
Second, there is currently a voice — or a logic — in the market: the crypto market watches the US stock market, the US stock market watches the AI revolution, and AI watches Nvidia's performance. Although Nvidia's earnings exceeded expectations, there are still quite a few bearish views circulating.
Compared to other vague and ambiguous bearish arguments, honestly digging through the financial statements for clues is the only approach with real evidentiary discussion value.
It's been a long time since I wrote a similar financial statement analysis. Consider this an offering of an alternative perspective for viewing the current market environment.
Source: https://x.com/agintender