NVIDIA credit default swap spikes, is AI debt about to blow up?

TL;DR

• Nvidia's 5-year CDS spiked intraday on July 27, as credit investors began reassessing potential obligations related to AI infrastructure.
• Oracle's 5-year CDS was around 1.25% over the same period, higher than Nvidia, indicating the credit market has already priced AI cloud infrastructure expansion into risk premiums.
• Guarantee and partnership frameworks can lock in forward demand but may also pass client financing risks back to Nvidia.
• Related securities: Nvidia (NVDA), Oracle (ORCL), SK Hynix, Broadcom (AVGO), TSMC (TSM), Microsoft (MSFT), Amazon (AMZN).

According to Bloomberg citing ICE Data Services, the cost of protecting Nvidia's debt against default for five years peaked at around 0.82% intraday on July 27, rising approximately 14 basis points in a single day, the largest single-day increase since the contract began active trading in November 2025.

A CDS can be understood as 'default insurance' for corporate debt. A price increase doesn't mean the market believes Nvidia is in immediate trouble, but it signifies that credit investors are demanding higher risk compensation. For a company recently upgraded to AA by S&P Global Ratings in June with still-strong cash flow, the absolute level is not high, but the signal is worth noting.

On the same front, Oracle has long been a reference point for the credit market to gauge AI infrastructure financing pressure. According to Bloomberg citing ICE Data Services, Oracle's 5-year CDS was around 1.25% over the same period, significantly higher than Nvidia. While the two companies differ in ratings, balance sheets, and business models, they both point to the same issue: the larger the AI data center orders, the more the credit market cares about who bears the risk for building and leasing this capacity.

What triggered this repricing of Nvidia were two pieces of AI infrastructure news. One is that Nvidia is reportedly discussing financing guarantees of up to $250 billion for projects related to OpenAI and SoftBank. The other is the announcement of AI plans and collaborations exceeding $500 billion between SK Group and Nvidia.

Equity investors tend to first see extended demand. The credit market asks another question first: if client financing falls short, who will provide the credit for this round of AI infrastructure expansion?

The Credit Market Tests Risk Allocation

Over the past two years, the market has been accustomed to explaining Nvidia through demand. Cloud providers, AI companies, and sovereign wealth funds are all buying GPUs, leading to tight supply, high margins, and strong cash flow. In this framework, the bigger the client's capital expenditure, the higher the order visibility for Nvidia.

But as AI data centers reach a larger scale, the question shifts from 'is there demand?' to 'who can finance the power, land, servers, and chips first?' If end-users haven't generated strong cash flows yet, suppliers helping clients finance their purchases could become a way to accelerate orders.

This is why Oracle's CDS level is a relevant reference. It's not a GPU supplier but a cloud infrastructure and database company. However, the market's focus on its AI cloud expansion, data center construction, and client concentration has already been reflected in its credit pricing. Oracle's 5-year CDS hovering around 1.25% means credit investors demand higher risk compensation for this type of AI infrastructure exposure than Nvidia's current level.

This doesn't imply an immediate default risk for Oracle or Nvidia. CDS trades on marginal risk repricing. For Oracle, the market looks at the pressure of cloud infrastructure expansion on liabilities, capital expenditure, and cash flow. For Nvidia, the market examines whether the asset-light, high-cash-flow chip supplier will partially bear client-side financing risks due to guarantees or partnership arrangements.

Credit investors aren't focused on the growth story; they look at the order of repayment in extreme scenarios. Potential guarantees, lease support, repurchase commitments, or other off-balance-sheet arrangements might not immediately appear as debt items, but they could transform into real obligations if project cash flows fall short.

Therefore, this CDS spike isn't trading on a 'surge in Nvidia's default probability'. More accurately, the market has started to re-price a portion of orders previously considered high-quality demand within tail risk.

Oracle Provides the Benchmark: AI Cloud Orders Also Have a Credit Price

Including Oracle in this discussion isn't because it bears the same type of obligation as Nvidia, but because it shows the market that AI infrastructure expansion is not just a revenue story for the equity market, but also a balance sheet story for the credit market.

Oracle's core narrative is growth in AI cloud infrastructure demand. Data centers, compute leasing, cloud service contracts, and large client demand enhance revenue visibility. However, the flip side is that such expansion typically requires upfront construction, involving facilities, power, servers, networking, and long-term procurement arrangements. There's a time lag between revenue realization and capital outlay, and the credit market prices this lag first.

Thus, Oracle's 5-year CDS level of around 1.25% serves as a coordinate for observing the credit cost of AI infrastructure. It's higher than Nvidia's ~0.82%, reflecting not which company has a stronger business, but the credit market's differentiated pricing based on different balance sheet structures, debt burdens, and capital expenditure intensity.

This also explains why Nvidia's CDS spike, even at a low absolute level, draws attention. Nvidia was previously seen as closest to the 'pick-and-shovel seller' in the AI cycle: getting paid upfront, high margins, strong cash flow, relatively low balance sheet pressure. But if more and more AI infrastructure projects require suppliers, cloud providers, energy companies, and financial institutions to co-design financing structures, Nvidia's credit risk boundary will be re-evaluated.

In other words, Oracle's CDS is already telling the market: AI cloud infrastructure growth isn't free. Nvidia's CDS jump asks another question: will the chip supplier also have to bear part of the credit cost for its clients' infrastructure expansion?

$250 Billion Guarantee Remains a Negotiating Variable

The easiest misinterpretation is to simply add the $250 billion guarantee and the $500 billion cooperation plan and conclude that Nvidia has taken on $750 billion in debt. The nature of these two items is different.

According to media reports, Nvidia is discussing providing financing guarantees of up to $250 billion for OpenAI's leasing of data centers related to SoftBank. The project involves SB Energy's participation in the Pike County Portsmouth Site in Ohio. A Department of Energy announcement in March mentioned plans to build 10GW of new generation capacity to serve 10GW of data centers.

This is still in early-stage negotiations. The terms, triggers, guarantee caps, and accounting treatment are not finalized, and it cannot be reported that Nvidia has signed or taken on debt.

In plain language, a guarantee means the guarantor might have to assume some responsibility if the lessee fails to pay in the future. It doesn't mean $250 billion is being lent out today, but it prompts investors to ask: is Nvidia simply selling chips, or is it starting to use its own credit to help clients buy chips?

This question affects valuation methods. As long as Nvidia remains a high-margin chip supplier, the market awards it a 'pick-and-shovel seller' premium. If it begins to shoulder project financing risks for clients, order quality depends on the terms, not just the size.

SK Cooperation More Like an Industry Expansion Framework

The SK line requires closer scrutiny. A more cautious public description is that SK Group and Nvidia announced AI plans and collaborations exceeding $500 billion. These include next-generation memory cooperation with SK Hynix, a 2GW data center with SK Telecom, and other AI infrastructure directions.

Such collaboration can strengthen Nvidia's position in the Korean AI infrastructure and memory supply chain. High Bandwidth Memory (HBM) is a critical component for scaling AI chips, and SK Hynix is a key supplier. For equity investors, this represents a deepening of Nvidia's ecosystem lock-in.

However, it cannot be automatically interpreted as a formal procurement contract, let alone that Nvidia is guaranteeing SK's $500 billion plan. Collaboration plans, supply directions, data center construction, and financing responsibilities are commitments at different levels of intensity.

The credit market needs to look precisely at this legal intensity. Which are mere cooperation frameworks, which become procurement commitments, which become financing guarantees, and which need to be disclosed as contingent liabilities in financial reports. The larger the numbers, the more critical the terms become.

Vendor Financing Analogy Enters Pricing

In this discussion, a frequently cited analogy is the vendor financing cycle of the 1990s telecom equipment industry. Equipment makers helped operators finance purchases of their own gear to sell more. In the boom phase, it seemed like a win-win: clients built faster, and equipment makers saw revenue growth.

The risk was that if client business revenue lagged expectations, the supplier wouldn't just lose orders but could also suffer financing losses. Revenue growth and credit risk became linked on the same chain.

Nvidia cannot be mechanically forced into this historical template. It's not a second-tier equipment maker from back then. GPUs remain at the core bottleneck of AI training and inference, and Nvidia has stronger pricing power, cash flow, and ecosystem control. The current absolute CDS level is still within a range understandable for an investment-grade company.

Oracle's situation also cannot be simplistically mapped to the vendor financing template. It's more like one of the bearers of AI cloud infrastructure expansion: it needs to build and lock in compute capacity first, then wait for client demand to translate into sustainable cash flow. The credit market assigning a higher CDS price to Oracle essentially demands a risk premium for its heavier infrastructure role.

Nevertheless, this analogy highlights a change: when an industry moves from supply shortages to a phase of massive capital expenditure, investors cannot just ask who can sell more chips. They must also ask who is providing the credit for this expansion.

Term Disclosure Determines Valuation Anchor

For NVDA holders, this event is not yet a balance sheet crisis, but a shift in the valuation anchor. The market is moving from 'how large are the orders?' to 'what conditions are needed for these orders to materialize?'

Oracle's CDS price provides a real-world reference: when the AI infrastructure story is tied to higher capital outlays, longer payback cycles, and more complex client structures, the credit market demands compensation ahead of the equity narrative. Nvidia still holds a stronger cash flow position and a lower absolute CDS level, but the market is beginning to test whether it transitions from 'selling chips' to 'providing credit support for chip demand.'

If subsequent disclosures show limited guarantee size, strict trigger conditions, clear project cash flows, and the SK collaboration gradually converting into real orders, then this CDS reaction might just be an upfront pricing for opaque terms. The equity market could still interpret it as a cost of locking in demand.

Conversely, if similar arrangements keep appearing, guarantee sizes expand, project payback relies on more distant AI commercialization, and financial reporting remains vague, the market will continue to discount the 'risk-free pick-and-shovel seller' valuation hypothesis. At that point, Nvidia's problem is no longer whether demand exists, but how much credit support is needed for that demand to materialize.

The variables that can change the judgment lie in the specific terms: Is the guarantee signed? Is there a cap? When does it trigger? How is it disclosed? Do rating agencies comment? And can the data centers coming online in 2027-2028 generate sufficient cash flow? Only after the information is confirmed can the market determine whether this CDS anomaly is short-term noise or an early warning for the AI infrastructure financing cycle.