Deep dive · supply chain

Nvidia’s customers and suppliers

Nvidia doesn’t name its biggest customers, and it names only a handful of suppliers. But its filings say more than most readers realize: what share of revenue its largest customers buy, which kinds of companies they are, which foundries, memory makers and assemblers it depends on, and how much it has committed to them. This page puts those disclosures together and connects them to the companies on our map.

of fiscal 2026 revenue came from two unnamed direct customers: 22% + 14%; of fiscal 2026 revenue came from two unnamed direct customers
of data-center revenue from hyperscalers in the quarter to July 26, 2026: $48.7B; of data-center revenue from hyperscalers in the quarter to July 26, 2026
of supply and capacity commitments to its own suppliers: $279B; of supply and capacity commitments to its own suppliers
of revenue over its latest four quarters, +83%: $303B; of revenue over its latest four quarters, +83%

The short answer

Nvidia’s customers come in two kinds. Its direct customers are the companies it invoices: add-in board makers, distributors, original design manufacturers (ODMs) and original equipment manufacturers (OEMs), the category that includes server makers such as Dell, Supermicro and HPE, cloud providers, AI model makers and system integrators. Its indirect customers are the companies that end up using the chips, buying them through those direct customers: cloud providers, “neocloud builders”, AI model makers, enterprises and governments.

Nvidia doesn’t name either group. It discloses that two direct customers accounted for 22% and 14% of its fiscal 2026 revenue, and that some indirect customers individually account for 10% or more. Its largest indirect customers are the hyperscale cloud providers, whose capex is the single biggest driver of Nvidia’s revenue. On the supply side, it names TSMC and Samsung as foundries, SK Hynix, Micron and Samsung as memory suppliers, and Hon Hai, Wistron and Fabrinet as assemblers.

Who buys from Nvidia

Since its first quarter of fiscal 2027, Nvidia splits its data-center revenue between hyperscalers and a group it calls “AI clouds, industrial and enterprise”. In the quarter to July 26, 2026, hyperscale revenue was $48.7B, up from $24.2B a year earlier, and the second group $40.3B, up from $16.9B. Data-center revenue as a whole was $278B over the latest four quarters, +90% on the year before.

Nvidia’s label for the second group points to the GPU clouds, such as CoreWeave and Nebius, that rent Nvidia capacity to AI labs and enterprises. Some of those are also companies Nvidia invests in and buys capacity back from, which is why the split matters: revenue from that group is partly revenue Nvidia has helped finance.

Many of Nvidia’s direct customers are the companies that build servers, although Nvidia doesn’t name them. Dell, Supermicro and HPE each report fast growth in AI servers, the systems Nvidia’s accelerators go into, and Dell and Supermicro each disclose a single customer above 10% of their own revenue. Following the chain from Nvidia’s disclosures to theirs is as close as the filings get to naming the end buyers.

Nvidia: Who Nvidia’s direct customers are
“We refer to customers who purchase products directly from NVIDIA as direct customers, such as AIBs, distributors, ODMs, OEMs, CSPs, AI model makers, and system integrators.”
Nvidia 10-K, filed February 25, 2026
Nvidia: Who Nvidia’s indirect customers are
“We refer to indirect customers as those who purchase products through our direct customers; indirect customers include CSPs, Neocloud builders, AI model makers, enterprises, and public sector entities.”
Nvidia 10-K, filed February 25, 2026
Nvidia: Direct customers, fiscal 2026
“For fiscal year 2026, sales to one direct customer represented 22% of total revenue and sales to another direct customer represented 14% of total revenue, all of which were primarily attributable to the Compute & Networking segment.”
Nvidia 10-K, filed February 25, 2026
Nvidia: Direct customers, latest quarter
“For the second quarter of fiscal year 2027, one direct customer represented 16% of total revenue, which was primarily attributable to the Compute & Networking segment.”
Nvidia 10-Q, filed August 26, 2026
Nvidia: Indirect customers over 10%
“We generate a significant amount of our revenue from a limited number of indirect customers, and we estimate some individually representing 10% or more of our revenue.”
Nvidia 10-K, filed February 25, 2026

Customers Nvidia helps finance

Nvidia has also become a financier of its own customers. Its latest 10-Q reports $99 billion of equity investments in its ecosystem with $25 billion more committed, $36 billion of commitments to buy back cloud capacity from AI cloud partners, guarantees capped at $105 billion for an OpenAI data-center campus, and a $2.0 billion share purchase in CoreWeave. Our money-loop map shows 5 links from Nvidia in all.

None of this is hidden, and vendor financing is a normal tool. But it changes how to read Nvidia’s customer list: some of the demand comes from customers whose ability to pay depends partly on Nvidia’s own support.

Nvidia: $99B invested, $25B more committed
“These include equity investments of $99 billion and equity investment commitments of $25 billion as of July 26, 2026.”
Nvidia 10-Q, filed August 26, 2026
Nvidia: $36B of cloud capacity it has agreed to buy back
“Our commitments, which are typically six years in duration, totaled $36 billion as of July 26, 2026, and decrease as capacity is used by third-party customers or by us for our research and development efforts.”
Nvidia 10-Q, filed August 26, 2026
CoreWeave: $2.0B share purchase, January 2026
“In January 2026, we entered into a securities purchase agreement with NVIDIA Corporation for a private placement of approximately 23 million shares of our Class A common stock at a purchase price of $87.20 per share, for aggregate gross proceeds of $2.0 billion.”
CoreWeave 10-Q, filed August 12, 2026

Who Nvidia buys from

On the supply side, Nvidia relies on a small number of companies for each critical step. TSMC and Samsung make its wafers. SK Hynix, Micron and Samsung supply memory, including the high-bandwidth memory stacked next to each accelerator. Hon Hai (Foxconn), Wistron and Fabrinet handle assembly, testing and packaging. Of those, TSMC, Micron and Fabrinet file with the SEC; the others report in Korea or Taiwan.

To secure that supply, Nvidia commits money years ahead. Its supply and capacity commitments rose from $119 billion to $279B in a single quarter, and it has paid premiums, provided deposits and signed long-term agreements to lock in capacity. Those commitments are what lets it ship at the scale it does, and they are also the risk: when demand for a product falls, Nvidia owns the excess. It took a $4.5 billion charge in early fiscal 2026 when export controls cut demand for its H20 chip.

For suppliers, Nvidia is a concentration of their own. TSMC discloses that its largest customer accounted for 19% of its 2025 revenue and its second-largest 17%, without naming them. Micron disclosed one customer at 17% of fiscal 2025 revenue and none above 10% in fiscal 2026. Those disclosures don’t say who the customers are, and we don’t guess; they show how much of the supply chain depends on a few buyers.

Nvidia: Nvidia’s foundries
“We utilize foundries, such as Taiwan Semiconductor Manufacturing Company Limited, or TSMC, and Samsung Electronics Co., Ltd., or Samsung, to produce our semiconductor wafers.”
Nvidia 10-K, filed February 25, 2026
Nvidia: Nvidia’s memory suppliers
“We purchase memory from SK Hynix Inc., Micron Technology, Inc., and Samsung.”
Nvidia 10-K, filed February 25, 2026
Nvidia: Nvidia’s assembly and test partners
“We engage with independent subcontractors and contract manufacturers such as Hon Hai Precision Industry Co., Ltd., Wistron Corporation, and Fabrinet to perform assembly, testing and packaging of our final products.”
Nvidia 10-K, filed February 25, 2026
Nvidia: $279B of supply and capacity commitments
“We have significantly increased our supply and capacity commitments from $119 billion last quarter to $279 billion as of July 26, 2026 to meet future demand.”
Nvidia 10-Q, filed August 26, 2026
Nvidia: How Nvidia secures supply
“To secure future supply and capacity, we have paid premiums, provided deposits, and entered into long-term supply agreements and capacity commitments, which have increased our product costs and this may continue.”
Nvidia 10-K, filed February 25, 2026
Nvidia: The H20 charge
“As a result of these requirements, we incurred a $4.5 billion charge in the first quarter of fiscal year 2026 associated with H20 for excess inventory and purchase obligations, as the demand for H20 products diminished.”
Nvidia 10-K, filed February 25, 2026
TSMC: Largest customers
“Our largest customer in 2023, 2024 and 2025 accounted for 25%, 22% and 19% of our net revenue in the respective year.”
TSMC 20-F, filed April 16, 2026
TSMC: Second-largest customer
“Our second largest customer in 2023, 2024 and 2025 accounted for 11%, 12%, and 17% of our net revenue in the respective year.”
TSMC 20-F, filed April 16, 2026

How to read Nvidia’s next filing

Three numbers in each Nvidia filing tell you most of what its customer and supplier disclosures can: the share of revenue from its largest direct customers (concentration), the split of data-center revenue between hyperscalers and AI clouds (who is buying), and supply and capacity commitments (how much demand Nvidia is betting on). When commitments grow faster than revenue, Nvidia is building for demand it expects rather than demand it has.

For the companies on the other side of those relationships, see the Nvidia company page, the chips layer and, for the companies that buy Nvidia’s chips and resell its capacity, CoreWeave, Dell and Supermicro.

Free · quarterly · from SEC filings

After each Big Tech earnings season we update the map from the new filings and publish what changed: who spent, who got paid, who kept it and what the filings reveal about the loop. Free, in Rob May’s Investing in AI newsletter.

Form not showing? Subscribe on Substack. No spam; unsubscribe any time.

Questions about Nvidia’s customers and suppliers

Who are Nvidia’s biggest customers?

Nvidia doesn’t name them. Its fiscal 2026 10-K says two direct customers accounted for 22% and 14% of revenue; direct customers are mainly server makers, distributors and cloud providers. Its largest indirect customers, the companies that end up using the chips, are the hyperscale cloud providers, and in its latest quarter hyperscale revenue was the largest part of its data-center business.

What is the difference between Nvidia’s direct and indirect customers?

Direct customers buy from Nvidia: add-in board makers, distributors, ODMs, OEMs such as Dell and Supermicro, cloud providers, AI model makers and system integrators. Indirect customers buy Nvidia products through them: cloud providers, neocloud builders, AI model makers, enterprises and public-sector bodies.

Who are Nvidia’s suppliers?

Its 10-K names TSMC and Samsung as foundries, SK Hynix, Micron and Samsung as memory suppliers, and Hon Hai, Wistron and Fabrinet as assembly, test and packaging partners.

Does TSMC make Nvidia’s chips?

Yes. Nvidia’s 10-K names TSMC, along with Samsung, as the foundries that produce its semiconductor wafers.

How much has Nvidia committed to its suppliers?

Its latest 10-Q reports $279 billion of supply and capacity commitments, up from $119 billion a quarter earlier, to meet future demand. Nvidia has also paid premiums and deposits to secure supply.

Does Nvidia invest in its customers?

Yes. Its latest 10-Q reports $99 billion of equity investments in its ecosystem, $36 billion of commitments to buy back cloud capacity from AI cloud partners and guarantees of up to $105 billion for an OpenAI data-center campus. It also bought $2.0 billion of CoreWeave shares in January 2026.

How concentrated is Nvidia’s revenue?

Highly. Two direct customers accounted for 36% of fiscal 2026 revenue between them, and Nvidia says its revenue is concentrated among a limited number of direct and indirect customers and that this may continue.

Research and education, not investment advice: nothing here is a recommendation to buy or sell any security. Figures come from companies’ SEC filings as reported; companies named were chosen because they show how the stack works, not as a view on their shares.