Nvidia’s role as a financier of the AI build-out faces its first big test as neocloud stocks fall, CNBC reported on 24 October 2026. Nvidia’s own Form 10-Q for the quarter to 26 July 2026 sets out what it has put behind its customers.
The filing lists guarantees capped at $105 billion for an OpenAI data centre campus in Ohio, $36 billion of AI cloud agreements and $20 billion of data centre leases it expects to hand to third parties.
The same filing shows $74.42 billion of operating cash flow in the first six months of the fiscal year and $4.43 billion spent on property, equipment and intangible assets. By our arithmetic that is about $70.0 billion of free cash flow, in line with the $70 billion CNBC cites. The Ohio guarantee cap alone is about 1.5 times that half-year figure. Nvidia declined to comment to CNBC.
Business Pill · FINANCIAL CONDITIONS
A short Business Engineer explainer on financial conditions: how the cost and availability of credit change. It teaches the general idea only and says nothing about Nvidia or any figure in this piece.
The key insight: As we read it, the ledger matters more than any single total: the 10-Q lists caps, purchase commitments and leases of different kinds, and one of them, the $105 billion Ohio guarantee cap, is about 1.5 times Nvidia’s free cash flow for the first six months of the fiscal year, by our arithmetic.
The lines below are different in kind. Some are commitments to buy, some are caps on what Nvidia could have to pay if a tenant defaults, and some are holdings. They run over different years. We list them side by side and do not add them up.
Line
Amount
What it is
Source
SB Energy guarantees (OpenAI, Ohio)
$105.0B cap
Credit support on leases at the PORTS campus; phases in from fiscal 2029
10-Q
Land, power and shell guarantees for AI clouds
$3.5B
Maximum gross exposure if partners default on leases
10-Q
AI cloud agreements
$36B
Nvidia commits to buy cloud services from AI clouds; typically six years
10-Q
Data centre leases for third parties
$20B
Leases of about fifteen years, to be reassigned
10-Q
Residual value guarantees on GPU financings
up to $125B
Promised to six asset managers in August
CNBC
Equity investments held
$99B
Plus $25B of equity investment commitments
10-Q
Supply and capacity commitments
$279B
For Nvidia’s own products, up from $119B a quarter earlier
10-Q
Free cash flow, six months
$70.0B
Operating cash flow less property, equipment and intangibles
10-Q; our arithmetic
Source: Nvidia Form 10-Q for the quarter to 26 July 2026, commitments and guarantees notes and cash flow statement; CNBC, 24 October 2026, for the residual value guarantees. Free cash flow is our arithmetic. The lines are separate scopes and are not additive.
Nvidia’s financing exposures and resources, in billions of dollars. Each bar is a separate scope and the bars are not additive. Source: Nvidia Form 10-Q for the quarter to 26 July 2026; CNBC, 24 October 2026 (residual value guarantees). Free cash flow is our arithmetic.
What Was Posted on X
Nvidia is helping finance demand for its own chips through more than $170 billion in investments and commitments, even as shares of AI cloud providers tumble and several Nvidia-backed firms delay IPO plans.
The Information reports that Nvidia is helping finance demand for its own chips through more than $170 billion in investments and commitments, as shares of AI cloud providers fall and several Nvidia-backed firms delay IPO plans. The linked article is paywalled and was not read here; the figure is The Information’s.
The AI boom is increasingly running on circular money, and itโs reminding some investors of 2008, per Bloomberg.
Chipmakers and cloud giants invest in or lend to AI companies, which then use that money to buy chips and computing power from the very same companies.
Recent examples:
– Broadcom, $AVGO, agreed to lend Anthropic up to $42 billion to buy chips and computing hardware – Nvidia, $NVDA, has put more than $40 billion into AI companies this year, including $30 billion in OpenAI.
Unusual Whales summarises a Bloomberg report that the AI boom increasingly runs on circular money, reminding some investors of 2008: chipmakers and cloud giants invest in or lend to AI companies that then buy their chips and compute. The examples it lists, Broadcom’s up-to-$42 billion facility for Anthropic and Nvidia’s $40 billion-plus of AI investments this year, are as the post states them; the Bloomberg article was not read here.
What the Filing Says About Each Line
The Ohio guarantees were signed in August 2026, the filing says. They cover leases for approximately 4.25 gigawatts of IT load at SB Energy’s PORTS Technology Campus in Pike County, under 20-year leases to OpenAI. CNBC describes the site as a proposed 10 gigawatt data centre; the two figures use different measures, and the 10-Q also notes an option on about 3.8 more gigawatts.
The cap is not money owed today. The filing says each guarantee generally becomes effective as each of nine data centres is placed in service, starting in fiscal 2029. It adds: “Our exposure declines as OpenAI fulfills its lease payments.” The guarantees end if OpenAI reaches a satisfactory credit rating, it says.
The AI cloud agreements are the neocloud line. Under them, AI clouds buy Nvidia’s data centre products and Nvidia commits to buy cloud services back. The filing says the clouds “can unilaterally stop providing to us and sell to third-party customers at more advantageous rates.” The commitments total $36 billion and decrease as capacity is used, it says.
The filing explains the reason in its own words: “We believe these AI cloud partners have strong customer demand and robust sales pipelines but are constrained by the large-scale infrastructure that is required to meet that demand.”
The Market Test
CNBC reports that CoreWeave, Nebius and IREN, all in Nvidia’s portfolio, are down at least 30% this month. By our reading of Yahoo Finance closes from 30 September to 23 October, CoreWeave fell 37.7%, Nebius 32.7% and IREN 39.1%. Nvidia fell 7.6%, the same figure CNBC gives.
CNBC says CoreWeave has roughly $51.6 billion of total debt, that yields on its bonds maturing in 2031 and 2032 topped 15%, and that its credit default swap spread is past 11%. Nvidia’s CoreWeave investment includes a commitment to buy up to $6.3 billion of its compute if it cannot find another taker, CNBC reports.
Huang told CNBC this week that neocloud demand is “off the charts.” He added: “What is missing in the middle is the working capital to fuel that growth.” David Trainer of New Constructs told CNBC: “If the demand for compute doesn’t meet these really super lofty expectations, we’re going to have a problem.”
New Constructs puts the net present value of Nvidia’s off-balance-sheet liabilities in fiscal 2026 at over $153 billion, CNBC reports. Moody’s David Gonzales told CNBC that how much exposure investment grade AI players can handle is “the question we’re evaluating at all times every day.”
Connecting the Coverage
On 21 October we reported The Information’s figure, via SBS News, of up to $164.5 billion in Nvidia guarantees and other commitments for data centres. By our arithmetic, the 10-Q’s $108.5 billion maximum guarantee exposure, $36 billion of AI cloud agreements and $20 billion of third-party leases come to the same $164.5 billion. We have not seen The Information’s method, so that is a match of numbers only, and the three lines are not one exposure.
The neocloud credit story has run all week. We reported CoreWeave’s 15% debt yields on 21 October and its default swap spread at 1,065.84 basis points on 21 October, from CNBC’s analysis of FactSet data. Bloomberg’s account of the 2023 CoreWeave loan described chip-backed debt on track to become a trillion-dollar market.
Our borrowing-costs piece set AWS’s sold-out capacity against wider pricing on AI debt. Our chip ETF piece found the PHLX Semiconductor index down 14.6% over the same October window. And Meta’s free cash flow went to negative $2.12 billion in its third quarter. CNBC sets Nvidia’s $70 billion against AMD’s just over $4 billion and notes that Alphabet went cash flow negative in the second quarter.
The Structural Read
As we read it, the largest line is also the most conditional: the 10-Q says the Ohio guarantees become effective as each of nine data centres is placed in service, starting in fiscal 2029, and that the exposure declines as OpenAI pays.
As we read it, the neocloud line is the one being tested now: CNBC describes the $36 billion of AI cloud agreements as backstops for neoclouds, and CoreWeave, Nebius and IREN, neoclouds in Nvidia’s portfolio, fell more than 30% in October by our reading of Yahoo Finance closes.
As we read it, the cash side is the counterweight: about $70.0 billion of free cash flow in six months, by our arithmetic from the 10-Q, against the just over $4 billion CNBC gives for AMD.
Jensen Huang, Nvidia chief executive, to CNBC, October 2026
“What is missing in the middle is the working capital to fuel that growth”
Three Implications
FOR INVESTORS The 10-Q gives the six-month operating cash flow as $74.42 billion and the Ohio guarantee cap as $105 billion; the guarantee phases in from fiscal 2029, the filing says.
FOR LENDERS CNBC reports CoreWeave’s 2031 and 2032 bond yields topped 15% and its credit default swap spread is past 11%.
FOR NEOCLOUDS The 10-Q says the AI clouds can stop providing services to Nvidia and sell to third-party customers at more advantageous rates, and that Nvidia’s commitments decrease as capacity is used.
The framework describes its subject as “How the money flows, where it cracks, what it buys, who can absorb it”.
As we read it, Nvidia’s ledger sits on two of the framework’s questions, where it cracks and who can absorb it: the neocloud partners behind the AI cloud agreements are where the market is testing, and Nvidia’s own free cash flow is the absorber. The framework was published in May 2026 and does not discuss Nvidia’s 10-Q.
What Is Not Established
How much of the $36 billion of AI cloud agreements has been used is not given in the passages of the 10-Q we read, so the live exposure on that line is not known from the filing alone. Whether the $6.3 billion CoreWeave commitment sits inside the $36 billion is not stated in the CNBC report or in the passages we read.
The $125 billion residual value figure comes from CNBC. A text search of the 10-Q found a Form of Residual Value Guaranty filed as Exhibit 10.1 but no $125 billion amount. CNBC’s description of over $500 billion in off-balance-sheet commitments also spans several tables of different kinds, which we do not combine here.
Business Engineer Framework
The AI Capex Map & The State of AI Hyperscalers
A Business Engineer map of how the AI build-out is financed and who can absorb it.
Nvidia’s filing puts a $105 billion cap on one guarantee, $36 billion behind neocloud partners and $20 billion in leases for others, against about $70 billion of free cash flow in six months. The guarantee phases in from fiscal 2029, and the filing says it declines as OpenAI pays. Shares of CoreWeave, Nebius and IREN, all in Nvidia’s portfolio, fell more than 30% this month by our reading of Yahoo Finance closes, which is the test CNBC describes.
A note on sourcing. Nvidia figures come from its Form 10-Q for the quarter to 26 July 2026: the commitments and guarantees notes, management discussion and the cash flow statement. The residual value guarantee figure, the neocloud share and credit figures, and all quotations from Huang, Trainer and Gonzales come from CNBC’s report of 24 October 2026. Share-price moves are our reading of Yahoo Finance closes.
Free cash flow, ratios and the reconciliation with the $164.5 billion figure are our arithmetic; the lines in the ledger are separate scopes and are not additive. Context figures come from our earlier pieces, linked in the text.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
