To The Moon

$NVIDIA(NVDA)$  

Nvidia's receivables rose 54.9% last quarter. Its revenue rose 17.9%.

In our Amazon post on 30 Oct we set a test: if suppliers are extending credit, Nvidia's receivables should grow faster than revenue. The last filed quarter already shows it.

That is the baseline, not the result. The real test is Nvidia's next report on 18 Nov, which is still to come.

When a seller's unpaid bills grow far faster than its sales, the 1999 chart comes back out. Lucent did the same thing in its last good year.

I'm not saying this is 1999. I'm saying it's a pattern worth checking properly, with one formula, before anyone reaches for the comparison. Here is how.

WHAT NVIDIA SAYS

Receivables went from $40.7B to $63.1B in one quarter. Revenue went from $81.6B to $96.2B.

[CHART: mm_recv_growth_E_1nov_tiger_v1.png]

Nvidia reports days sales outstanding (DSO) of 60 days, up from 45. That is how long, on average, it waits to get paid.

The company gives its own reason. It is "due to extended payment terms on large, multi-quarter agreements with certain investment-grade customers."

The 10-Q says those terms run from 90 days up to one year. That is the company's explanation, and it is a fair one. Big buyers on long contracts ask for time to pay.

WHAT LUCENT LOOKED LIKE

In fiscal 1999, Lucent's receivables rose 41.0% while revenue rose 20.4%. Same shape.

As filed, its DSO went from 71 days in FY1998 to 85 days in FY1999. After Lucent restated for the Avaya spin-off, the like-for-like figures were 89 days in FY1999 and 102 days in FY2000.

Then there was the lending. By Sep 2000 Lucent had $8.1B of customer-financing commitments, 24.0% of that year's revenue. Only $2.0B had been drawn.

The bill came in fiscal 2001. Lucent booked $2.25B of provisions for bad debts and customer financing. About $1.3B came from three customers, including Winstar and One.Tel.

Revenue fell to $21.3B. The stock went from over $80 (Dec 1999) to under $1 (2002).

THE METHOD: RUN BOTH ON ONE FORMULA

Comparing 60 days with 102 days is comparing two different rulers. Lucent averaged its opening and closing receivables. Nvidia's 60 uses the closing number only.

So rebuild it. Here are the steps, and they work on any company.

1. Average the receivables at the start and end of the quarter.

2. Divide by the quarter's revenue.

3. Multiply by 91 days.

On Lucent's own method, Nvidia's DSO is 49.1 days. That is well short of Lucent's 102 in FY2000.

[CHART: mm_lucent_nvda_1nov_tiger_v3.png]

One more check. Against a year earlier, receivables rose 126.8% and revenue rose 105.9%. So most of the gap is one quarter's jump, not a long drift.

WHAT NVIDIA IS BACKING

This is the part that rhymes more. All of it is in Nvidia's own 10-Q.

Its private stakes are $47.9B, against $3.8B a year earlier. It also has $36B of cloud-service commitments, typically six years, paying AI clouds for computing capacity.

It guarantees $3.5B of land, power and shell costs for AI clouds, with $712M held in partners' escrow. It also holds $3.3B of stakes in what it calls infrastructure financiers.

Then the big one. Nvidia guarantees 20-year data-centre leases for OpenAI and SB Energy up to a cap of $105B, with the first phase expected in fiscal 2029.

[CHART: mm_backing_F_1nov_tiger_v2.png]

None of this is customer lending in Lucent's sense. It is a seller backing its buyers' capacity in other ways. That is the link worth watching.

WHERE THE WEAK CREDIT SITS NOW

My hypothesis, and I could be wrong: in 1999 the weak borrowers were start-up phone companies. Now they sit one step down, in the neoclouds.

CoreWeave owes $35.6B of debt principal. Of that, $4.4B is due in the rest of 2026 and $6.2B in 2027.

Its net interest was $640M last quarter, 24.9% of revenue. It still made an operating loss of $49M.

Its backlog is $103.7B. Customer A was 36.0% of revenue, B 26.0% and C 10.0%.

Its newest notes pay 9.625% to 9.75%. Nvidia bought $2.0B of its shares in Jan 2026.

WHERE THE ANALOGY BREAKS

The counter-case is strong, and I'd put it first if I were arguing the other side.

Nvidia holds $56.6B of cash and debt securities. It is paying for this out of its own cash, not out of borrowed money.

Its operating margin last quarter was 66.2%. A seller earning that much has room to wait for its money.

Five direct customers owe 70% of its receivables. Nvidia describes them as investment-grade, not start-up carriers.

And the $105B is a contingent cap that starts years out. It is not cash lent today.

What would change my mind toward "this is fine": DSO falling back to about 50 days as thos

e long contracts get paid. What would push me the other way: DSO staying at 60 or above, plus the backing lines growing again.

THE DATED TESTS

First, CoreWeave's Q3 results, with the call on Wed 11 Nov at 06:00 SGT. Watch what it pays to refinance the 2026 and 2027 maturities, against the 9.625% to 9.75% on its notes.

Second, Nvidia's Q3 results, on Wed 18 Nov at about 05:20 SGT. If receivables again grow faster than revenue and DSO stays at 60 or more, it's a trend. If DSO drops toward 50, it was timing.

I'll also check whether the $105B and $36B lines change. I'll score it either way. The pattern rhymes, but the borrower is different.

HOW TO RUN THIS ON ANY SELLER

Take any company that sells to a few large buyers. Compare receivables growth with sales growth for the last quarter and the last year.

Rebuild DSO on one formula before you compare it with anything. Then read the commitments and guarantees notes, and ask who the buyers borrow from.

When a supplier starts giving its biggest customers longer to pay, do you check the customers' debt, or only the supplier's?

Not advice. I may hold positions in the names mentioned.

Stop guessing. Start calculating.

Live to fight another day. 🤙

#options #tradingpsychology #riskmanagement #optionstrading