Nvidia’s latest financial disclosures reveal that its role in the artificial intelligence infrastructure boom extends beyond supplying advanced processors. As of Oct. 25, 2026, the company reported maximum gross exposure of US$118 billion across several guarantee arrangements, while separately listing US$32 billion in future commitments tied to neocloud operators.
The figures, disclosed alongside Nvidia’s third-quarter fiscal 2027 results on Nov. 17, show how the chipmaker is becoming increasingly connected to the financing and construction of AI infrastructure.
But the numbers require careful interpretation. The US$118 billion represents maximum gross exposure under specified guarantees, not money Nvidia has already paid. The US$32 billion represents conditional future commitments, not an unconditional loan facility.
The distinction matters because Nvidia’s financial exposure depends on project completion, financing conditions, partner performance and the terms of individual agreements.
The US$118 billion figure is a guarantee ceiling, not a cash payment
Nvidia’s chief financial officer commentary disclosed three categories of maximum gross guarantee exposure as of Oct. 25.
The largest was US$105 billion in credit-support guarantees connected to SB Energy’s PORTS-Pike Technology Campus in Ohio. Another US$9.5 billion covered power and systems credit-support guarantees, while US$3.5 billion related to land, power and shell guarantees for neoclouds.
Together, those categories produced the US$118 billion maximum gross exposure figure.
A guarantee is different from a conventional loan or direct investment. It generally creates an obligation to make payments if specified conditions occur, such as a partner defaulting on its obligations.
Nvidia explicitly disclosed that it had not made any payments related to its guarantees as of the reporting date.
That does not eliminate the risk. It means the disclosed maximum should be understood as potential exposure under the arrangements rather than a record of cash already spent.
Why the Ohio data-centre project matters
The US$105 billion guarantee category is connected to an infrastructure development involving SB Energy’s PORTS-Pike Technology Campus.
Nvidia said its August 2026 guarantees provide credit support for land, power and shell construction covering approximately 4.25 gigawatts across nine buildings. The campus is intended to host Nvidia infrastructure under 20-year leases to OpenAI, subject to limited exceptions.
The guarantees do not all become effective immediately. Nvidia said they take effect in phases, with each guarantee becoming active only when its corresponding building is completed and ready for service.
The first is expected in fiscal 2029. Exposure is expected to build as buildings come online, reach a maximum of US$105 billion and then decline over each 20-year lease term as OpenAI makes lease payments.
Nvidia also has an option to provide credit support in phases for approximately 3.8 additional gigawatts as the site expands.
This arrangement illustrates the scale of capital required to build modern AI infrastructure. Data centres need more than processors: they require suitable land, power connections, buildings and long-term commercial arrangements.
For Nvidia, the guarantees help support infrastructure intended to use its technology. They also create a potential financial obligation if the relevant counterparties fail to meet their commitments.
Neocloud commitments have conditions attached
A separate part of Nvidia’s financing strategy involves neocloud companies, specialised providers that build computing infrastructure and sell access to AI processing capacity.
Nvidia reported US$32 billion in future additional commitments under its neocloud agreements as of Oct. 25. The schedule allocates US$6 billion to fiscal 2028, US$7 billion to 2029, US$6 billion to 2030, US$5 billion to 2031 and US$8 billion to 2032 and thereafter.
There is no amount listed for the remainder of fiscal 2027 in that schedule.
The commitments only take effect if the neocloud partners secure financing to purchase Nvidia infrastructure and the associated computing capacity is deployed, made operational and accepted by Nvidia.
If those conditions are not met on time, the amount and timing of commitments may change, and certain commitments could be reduced or eliminated.
Under the arrangements, Nvidia earns revenue from the initial sale of infrastructure. It may also participate in revenue generated by the neoclouds from third-party customers if specified criteria are satisfied.
This structure creates an additional potential revenue stream, but it also ties the commercial outcome to the ability of infrastructure partners to attract customers and operate their facilities.
Why the financing model is attracting attention
Nvidia’s strategy is designed to help customers overcome the enormous capital requirements associated with AI computing.
In August, the company announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent financing platforms intended to mobilise more than US$500 billion in third-party capital over time.
The announcement described memorandums of understanding, with the partnerships still subject to execution of final agreements.
The proposed platforms are intended to help finance AI infrastructure across Nvidia’s ecosystem. The important point is that the US$500 billion figure is a target for capital mobilisation over time, not a completed Nvidia investment or a pool of cash already available for distribution.
Nvidia’s latest disclosures also make clear that financing is expected to be independently underwritten and provided by capital providers. The company may provide limited residual-value support for selected projects, subject to its own evaluation and risk-management processes.
This approach could help expand access to computing capacity without Nvidia directly financing every project. However, the guarantee disclosures show that the company can still retain exposure to specific infrastructure arrangements.
Strong sales provide scale, but do not remove financing risk
Nvidia reported third-quarter fiscal 2027 revenue of US$115.8 billion for the period ended Oct. 25, up 103% from a year earlier.
Data Centre revenue reached US$107.6 billion, an increase of 110%. The company also reported US$48.3 billion in cash from operating activities during the quarter.
Those results demonstrate the strength of demand for Nvidia’s computing platforms. They also provide financial context for its expanding commitments to the infrastructure ecosystem.
However, revenue growth does not automatically remove the risks associated with guarantees and long-term financing arrangements.
The projects still depend on construction, power availability, customer demand and the ability of partners to meet their contractual obligations. If a project encounters serious financial or operational difficulties, guarantees could become more than a disclosure in a financial report.
For investors, the important question is not simply how large Nvidia’s commitments are. It is how much of that exposure eventually becomes active, how partners perform and whether the resulting infrastructure generates enough business to justify the capital involved.
What happens next
Nvidia’s next quarterly disclosures should help investors track changes in its guarantees, future commitments, infrastructure agreements and cash generation.
The Ohio project is particularly important because its guarantees become effective in phases as buildings are completed. Fiscal 2029 is the expected timing for the first guarantee to take effect under the company’s disclosed schedule.
The neocloud agreements also require monitoring. Their value depends on partners obtaining financing and bringing computing capacity into operation.
For now, Nvidia’s records show a company pursuing an ambitious infrastructure strategy while placing conditions around some of its financial support.
The US$118 billion guarantee figure is significant, but it is not a bill already due. The real test will be whether the projects deliver functioning infrastructure, attract sustainable customer demand and allow Nvidia to expand its AI business without turning potential guarantees into substantial cash losses.
