Recommended Free Tools
Assess an AI company’s financing and revenue as one connected network: trace who supplies capital, who buys the product or compute, who ultimately uses it, and whether reported sales turn into cash from an independently funded customer. Measure concentration separately across revenue, receivables, contracts, funding, suppliers, and capacity. A connected investor, supplier, or customer is a diligence signal—not proof of improper financing.
What counts as circular financing—and what does not?
The useful question is not simply whether an investor is also a supplier or customer. It is whether capital and commercial obligations link up in a way that leaves the company reliant on the same party or funding ecosystem to finance demand, fulfill a sale, or absorb unused capacity.
As an Amazon Associate I earn from qualifying purchases.
For each material relationship, map the flows in both directions. Capital may take the form of equity, debt, convertible instruments, loans, warrants, guarantees, cloud credits, prepayments, or vendor financing. Commercial flows may include hardware purchases, cloud services, capacity reservations, purchase commitments, offtake, and revenue sharing. Note whether each party is an investor, lender, supplier, direct customer, reseller, cloud operator, or end user—and whether supposedly separate entities share ownership, control, financing, or strategic direction.
Keep the conclusion specific. “Linked financing and demand exposure” may describe a documented arrangement more accurately than “circular financing.” Explain whether capital is expected to return through purchases, revenue, repayment, or another contractual route. A relationship alone does not establish that a sale is fictitious, uncollectible, or improper.
Who is actually paying for the AI product or compute?
A direct buyer is not necessarily the independent end customer. A cloud provider may buy hardware or capacity, then sell cloud services to an AI company; the provider’s purchase may therefore depend on demand from a separate end user. Ask who funds that end user, who is obligated to pay the direct seller, and who bears the loss if downstream demand or funding fails.
NVIDIA says its estimates of indirect customer exposure draw on purchase-order information, product specifications, internal sales data, and other sources. In its fiscal 2026 annual filing, the company also described an AI research and deployment company as contributing a “meaningful amount” of revenue through cloud services purchased from NVIDIA customers, without naming that end customer or quantifying the contribution. That disclosure illustrates both the value and the limits of public customer attribution: a named direct customer may not reveal the ultimate source of demand.
Request the contracts, ownership and beneficial-owner information, payment path, and end-customer or utilization evidence needed to resolve the chain. If some links cannot be verified, record them as unknown rather than treating the direct buyer as proof of independent end demand.
What do public disclosures show about concentration and linked exposure?
These issuer filings provide examples of different exposure channels; they are not findings about every AI company or industry-wide benchmarks. The periods and denominators differ, so do not combine the figures as though they measured the same risk.
| Company and period | Reported disclosure | What it indicates |
|---|---|---|
| NVIDIA, second quarter of fiscal 2027, quarter ended July 26, 2026 | One direct customer accounted for 16% of second-quarter revenue. | Direct-customer revenue concentration for that quarter; it does not identify all indirect end-customer exposure. |
| NVIDIA, first half of fiscal 2027, through July 26, 2026 | Three direct customers accounted for 16%, 15%, and 13% of first-half revenue. | First-half shares, which should not be compared as if they were annual or quarterly shares. |
| NVIDIA, as of July 26, 2026 | $36 billion in AI-cloud commitments, typically six years in duration. The agreements can include revenue sharing and purchases of capacity an AI cloud has not sold to third parties. | Contractual demand and utilization exposure specific to NVIDIA, not a sector benchmark. The filing says commitments can decrease as third-party customers or NVIDIA research and development use capacity. |
| NVIDIA, fiscal year ended January 25, 2026 | Direct customers accounted for 22% and 14% of annual revenue. | Annual direct-customer shares from a different reporting period than the later quarterly and first-half disclosures. |
| Cerebras Systems, 2025 revenue, as reported in its 2026 prospectus | G42 accounted for 24.0% of revenue and MBZUAI for 62.0%. The prospectus identifies G42 and MBZUAI as related parties with respect to each other under ASC 850. | Revenue concentration and a disclosed relationship between two major customers; these are not the same measure as receivables concentration. |
| Cerebras Systems, accounts receivable as of December 31, 2025 | One customer accounted for 77.9% of accounts receivable. | Collection exposure at a point in time, not that customer’s share of annual revenue. |
| Cerebras Systems, OpenAI collaboration; loan funded January 2026 | The prospectus describes an approximately $1.0 billion secured working-capital loan funded by OpenAI to support infrastructure and capabilities needed to provide compute services OpenAI had contracted to purchase; it also describes a warrant. | A disclosed lender/customer link to examine for contractual terms, repayment, and dependence. The arrangement by itself does not establish improper financing. |
Sources: NVIDIA Form 10-Q for the quarter ended July 26, 2026; NVIDIA Form 10-K for the fiscal year ended January 25, 2026; and Cerebras Systems 2026 prospectus. These filings report what the issuers disclose; they do not independently establish every commercial assertion or determine the overall risk of an arrangement.
How should you measure customer concentration?
Use a consistent period and denominator for each calculation. For revenue, calculate the largest-customer share and the combined shares of the top three and top five customers. Then repeat the analysis separately for other exposures instead of using one revenue percentage as a proxy for everything.
- Revenue: identify the share attributable to each customer or customer group, and note whether the issuer names the customer or uses an anonymized label.
- Receivables and cash collections: measure outstanding balances and collected cash separately. A large receivable balance can create collection dependence even when revenue is spread across more buyers.
- Bookings, backlog, and remaining performance obligations: examine how much future business depends on a small number of contracts, and account for termination rights and conditions.
- Funding: identify dependence on particular equity investors, lenders, or other sources of capital, including any parties that also buy, supply, or finance capacity.
- Suppliers and committed capacity: measure dependence on critical hardware, cloud providers, facilities, and minimum purchase or take-or-pay obligations.
Keep quarterly and annual figures separate, and distinguish customer groups from individual legal entities where the disclosure allows. A direct buyer may be only a channel to an end user, while several named buyers may depend on the same source of funding. Do not apply a universal “safe” concentration threshold: the appropriate concern depends on customer credit, contract terms, cash collection, alternatives, and the company’s own covenants or financing constraints.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Does booked revenue turn into independent cash collection?
For each significant arrangement, build a contract-to-cash record. Capture the payer and payee; amount or range; timing; recourse; termination rights; payment conditions; performance obligations; and any noncash consideration. Reconcile recognized revenue to invoices, cash receipts, receivables aging, credit losses, deferred revenue, customer advances, and customer-related warrants or other instruments.
Then test the path to payment. Did the company deliver the contracted service? Is the buyer paying from its own operating cash, or does its ability to pay depend on financing supplied by the seller, an investor, a related party, or another counterparty in the same network? If the seller has advanced funds or guaranteed obligations, establish whether those terms change the company’s exposure to delayed payment, default, or cancellation.
Revenue recognition and collectability depend on the specific contracts and applicable accounting guidance. The examples here do not establish a universal accounting rule; review the company’s terms and accounting treatment rather than inferring a conclusion from a headline revenue figure.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What happens if the largest customer or end user falters?
Model at least five cases for the largest customer or material end user: loss, delayed payment, contract dispute, failed renewal, and lower-than-planned utilization. Include correlated stress: the same investor, lender, or funding source that supports the company may also be financing the customer’s ability to pay.
Free tools Windows power users keep installed
One-click scans. No signup required.
Trace the secondary effects, not just lost sales. Could the company still service debt or leases, meet minimum purchase obligations, or pay for reserved cloud capacity? Would it have to buy back capacity, absorb a guarantee, or fund working capital while invoices remain unpaid? Consider GPU depreciation or obsolescence, the residual value of equipment, and whether the company can realistically redeploy capacity to another customer. Identify which counterparty bears each loss and whether contractual protections are enforceable in practice.
How should you compare two financing or customer arrangements?
Compare arrangements on the same terms rather than calling one “diversified” based only on the number of customer names. A concentrated but creditworthy customer that pays in cash under short, cancellable commitments can present a different risk from several buyers that rely on the same connected funding ecosystem.
- How independent is end demand from the company’s investors, suppliers, and financing sources?
- What are the customer’s and funder’s credit quality, and what share of revenue and receivables is exposed to each?
- How much recognized revenue has been collected in cash, and how much remains in receivables or future performance obligations?
- How long do commitments last, and what termination flexibility, recourse, guarantees, or purchase-back exposure apply?
- Who bears utilization and residual-value risk if capacity is not rented or equipment cannot be redeployed?
- Are there related-party links, and how clearly do disclosures identify the end customer, payment chain, and contract terms?
How to state a defensible conclusion
Separate what the documents establish from what remains uncertain. Use precise descriptions such as “concentrated revenue,” “high receivables dependence,” “linked financing and demand exposure,” or “limited end-customer transparency.” Identify the reporting period, denominator, named parties, contractual mechanisms, and missing evidence behind the conclusion. Do not label a structure circular merely because a supplier invested in a customer or a strategic buyer is also a lender; explain the actual flows and who bears the risk if payment or demand stops.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
