InvestNot yet confirmed elsewhere1 publisher3 min readPublished
At least half of Nvidia's receivables now sit with five customers
Five customers each held at least 10% of Nvidia's receivables in the July quarter, up from three in January, according to a filing CNBC cited. That puts at least half of what Nvidia is owed with five buyers while the number of firms renting out its chips keeps growing.
The Investor · Invest desk

What happened
- SemiAnalysis counted 323 providers offering Nvidia GPUs as of September, up from 209 less than 11 months earlier.
- Amazon CEO Andy Jassy told analysts in July that Amazon will not be able to serve all the demand it foresees this year.
- CoreWeave's near-term capacity is essentially sold out, CEO Mike Intrator said on the company's August earnings call.
- Google and Microsoft have started tapping CoreWeave for capacity even though they compete with it.
Why it matters
- exposure A funding squeeze at any one of five buyers now reaches Nvidia's cash collections directly, because each owes it a tenth or more of its outstanding receivables.
- constraint Buyers needing GPUs on short notice have to commit, and often prepay, months ahead, since neocloud capacity is built only after contracts are signed and financed.
- decision Google and Microsoft are choosing to pay a competitor for capacity, so some hyperscaler spending on Nvidia chips now passes through CoreWeave.
When five customers each hold at least 10% of receivables, their combined share has a floor [16]. CNBC treats the July-quarter count as a sign that a business long led by cloud providers is diversifying [4]. Five names at a tenth or more each means those five hold at least half of what Nvidia is owed. In January, when three cleared the line, the floor was 30% [2][16].
The filing fits more than one reading. If the two new names are neoclouds, Nvidia's collections depend on how those companies are financed. The best-known neoclouds raise money against signed contracts and can ask their own customers to pay upfront, Modal chief executive Erik Bernhardsson said [10]. If the new names are the cloud providers that have topped Nvidia's customer list for years, the exposure sits with its oldest buyers and the change is mostly one of volume [4]. A third possibility is that the largest single customer's share fell while others caught up, so the top of the book spread out even as the floor rose [2]. CNBC's account of the filing does not name the five [2].
SemiAnalysis's count went from 209 providers to 323. That is 114 more, a rise of about 55% in under 11 months, or more than ten new providers a month [1][15]. Modal moved off the hyperscalers and now runs on 25 neoclouds [7]. "You can get a few hundred GPUs or maybe a thousand, but at our scale, we needed way more GPUs," Bernhardsson said [8]. Hyperscalers don't always have as many GPUs as enterprises require, according to Gartner analyst Hardeep Singh [11].
The big clouds are sending some of the demand they cannot meet to rivals. "Some of the hyperscalers have approached us about taking care of customers they're worried about because they don't have the ability to service those customers when they need it," said Marc Boroditsky, chief revenue officer of Nebius [12].
Nvidia's chief executive talks about the newer providers in terms of their order books. "You're going to see a whole new crop of really, really exciting neoclouds with hundreds of billions of dollars backlog together," Jensen Huang said at a Goldman Sachs conference in San Francisco last month [14]. Management's $108 billion forecast for the October quarter, up 89%, puts the year-earlier quarter at about $57 billion [3][17].
We think the receivables count tells you more about Nvidia's risk than the provider count does. In our view the book is more concentrated than it was in January, and its newest buyers are financed against backlogs whose value depends on chips staying scarce. The counter-case is the third reading above: with more names over 10%, Nvidia may rely less on any single one. Our view is wrong if the five turn out to be mostly the hyperscalers, whose own GPU spending already has signed demand behind it. Anthropic and OpenAI have committed more than $500 billion between Amazon and Microsoft, which held 59% of the cloud infrastructure market in 2025, according to Gartner [5].
What to watch
- Nvidia's next quarterly filing: whether the count of customers above 10% of receivables rises past five or falls back toward January's three.
- Nvidia's October-quarter result against management's $108 billion forecast, and whether receivables grow faster than revenue.
- CoreWeave's next earnings call: if near-term capacity is no longer sold out, the contract-backed financing behind newer neoclouds gets its first test.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence55
- Adoption65
- Hype gap+10
- Incentives70
- Confidence55
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
Industry research firm SemiAnalysis counted 323 Nvidia GPU providers as of September, up from 209 less than 11 months earlier.
- [2]
Five clients accounted for at least 10% of Nvidia's accounts receivable in the July quarter, up from three in January, according to a filing.
- [3]
Nvidia management anticipates $108 billion in revenue for the October quarter, which would be an 89% year-over-year jump.
- [4]
Cloud infrastructure providers have ranked at the top of Nvidia's customer list for several years, but the business is diversifying.
- [5]
In the past year Anthropic and OpenAI have committed to spending over $500 billion between Amazon and Microsoft, which controlled 59% of the cloud infrastructure market in 2025, according to Gartner.
- [6]
Amazon CEO Andy Jassy told analysts in July that Amazon won't be able to serve all the demand it foresees this year.
- [7]
Modal went from running on the hyperscalers to signing up with the major neoclouds and now uses 25 of them, said CEO Erik Bernhardsson.
- [8]
"You can get a few hundred GPUs or maybe a thousand, but at our scale, we needed way more GPUs," Bernhardsson said.
- [9]
Google and Microsoft have started tapping CoreWeave, even as they all compete with one another.
- [10]
The most visible neoclouds can require some upfront payment, and chips might not come online for months, because providers raise funding based on contracts and set up the data center equipment, Bernhardsson said.
- [11]
Hyperscalers don't always have as many GPUs as enterprises require, said Gartner analyst Hardeep Singh.
- [12]
"Some of the hyperscalers have approached us about taking care of customers they're worried about because they don't have the ability to service those customers when they need it," said Marc Boroditsky, chief revenue officer of Nebius.
- [13]
CoreWeave's near-term capacity remains essentially sold out, CEO Mike Intrator said on the company's August earnings call.
- [14]
"You're going to see a whole new crop of really, really exciting neoclouds with hundreds of billions of dollars backlog together," Nvidia CEO Jensen Huang said at a Goldman Sachs tech conference in San Francisco last month.
- [15]
SemiAnalysis's provider count rose by 114, about 55%, in under 11 months, or more than ten new providers a month.
- [16]
Five customers each holding at least 10% of receivables means they hold at least 50% combined, against a floor of 30% when three customers cleared the threshold in January.
- [17]
An October-quarter forecast of $108 billion at 89% growth implies about $57 billion of revenue in the same quarter a year earlier.
Sources
1 independent publisher whose own reporting we read for this story.
- cnbc.comNvidia GPUs are everywhere. Here are the ways companies are accessing them
1 article · October 10, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- NeocloudsFollow
- GPU Cloud CapacityFollow
- Customer concentrationFollow