To The Moon

Nvidia has largely answered the question of whether AI demand remains strong. Investors are now examining how its customers finance GPU purchases—and how much risk Nvidia is assuming to keep the infrastructure boom moving.

Nvidia jumped 8.7% in the first session after earnings, lifting the Philadelphia Semiconductor Index by 2.3%. One day later, NVDA fell 4.6%.

Fed Chair Kevin Warsh’s hawkish remarks pushed Treasury yields and rate-hike expectations higher, but monetary policy was not the whole story.

Friday’s decline was likely driven by three overlapping factors:

• Higher rates pressured technology valuations.

• Investors took profits after Thursday’s 8.7% surge.

• Concerns about circular financing, customer spending and falling margins returned to the spotlight.

The earnings were genuinely strong

Nvidia reported fiscal Q2 revenue of $96.2 billion, up 106% year over year and above the $92.2 billion expected by analysts.

Other highlights included:

• Data-center revenue of $89.0 billion, up 117%.

• Adjusted EPS of $2.22, versus $2.10 expected.

• Adjusted gross margin of 75%.

• Q3 revenue guidance of $108 billion, above the $104.2 billion consensus.

The biggest surprise was management’s preliminary forecast for approximately 70% revenue growth in fiscal 2028. Wall Street had been expecting roughly 44%.

Vera Rubin is already shipping and is expected to contribute around 20% of data-center revenue in the current quarter. Nvidia and AWS also plan to deploy an additional two million GPUs across Amazon’s infrastructure in 2027 and 2028.

These numbers provide strong evidence that the AI infrastructure cycle is not facing an immediate demand collapse. Nvidia earnings | Reuters

What does “circular financing” mean?

Traditionally, a chip company sells processors while its customers arrange their own financing, facilities and end-market demand.

Nvidia is now becoming more involved in customer funding, infrastructure development, capacity leasing and project guarantees.

The potential loop looks like this:

Financial institutions or Nvidia support an AI infrastructure project → cloud providers and AI labs build data centers → they buy Nvidia GPUs → Nvidia records hardware revenue → Nvidia provides additional investment, guarantees or capacity support.

This does not automatically mean the revenue is artificial. However, it makes the source and quality of demand more complicated.

1. More than $500 billion in third-party financing

Nvidia has signed memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to create financing platforms intended to mobilize more than $500 billion for AI infrastructure.

That is not $500 billion of Nvidia’s own capital. The platforms are expected to use third-party money and independent underwriting.

Still, one purpose of the financing is to help Nvidia’s customers build data centers, purchase GPUs and expand capacity. Nvidia is therefore becoming involved in both supplying the equipment and helping create the financing channels that support its sales. Nvidia announcement

2. A guarantee of up to $105 billion for OpenAI infrastructure

Nvidia has also agreed to provide a guarantee of up to $105 billion connected to OpenAI’s Ohio data-center project and invest $1.5 billion in developer SB Energy.

OpenAI remains responsible for paying the rent, and Nvidia is not guaranteeing every obligation associated with the project.

However, if OpenAI defaults, Nvidia could be responsible for part of the gap between the guaranteed minimum value and what the project’s owner could recover through a sale or new lease. Reuters

3. Backstopping unused cloud capacity

Nvidia had also explored a financing model for smaller AI cloud providers.

Under the proposed structure, Nvidia would offer credit support and potentially rent back capacity that customers could not sell, making it easier for those companies to borrow money and purchase Nvidia GPUs.

Nvidia could earn revenue from the initial hardware sale and later receive a share of the cloud revenue generated by that capacity.

The initiative has reportedly been paused amid concerns about circular transactions, control over customers and potential regulatory scrutiny. Reuters

Why are investors concerned?

Circular financing does not prove that Nvidia’s current revenue is fake. It raises four more nuanced questions.

First, is financing pulling future demand forward?

If some customers need guarantees, external debt or capacity backstops to purchase GPUs today, current orders may include demand that would otherwise have arrived later.

Second, can the buyers generate enough cash?

AI cloud providers such as $CRWV and $NBIS are spending heavily on infrastructure. Their ability to continue purchasing GPUs depends on access to financing, capacity utilization and the price they can charge for AI compute.

Third, who absorbs the downside?

When utilization is high, these structures accelerate industry growth. If GPU rental prices fall or customers default, Nvidia could face weaker orders, investment losses, guarantee payments and obligations linked to unused capacity.

Fourth, how durable is GPU collateral value?

AI processors depreciate as new generations arrive. The economics of infrastructure financing depend on Rubin and future systems generating enough revenue before older hardware loses competitiveness.

Gross margin adds another layer of risk

Nvidia reported a 75% gross margin, but management expects:

• Approximately 74% in fiscal Q3.

• A potential bottom of 71%–72% in fiscal Q4.

• A recovery to around 72%–73% during fiscal 2028.

Memory and other component costs are rising faster than previously expected.

Nvidia can therefore continue delivering extraordinary revenue growth while facing pressure from both manufacturing costs and financial commitments across its ecosystem.

Related stocks to watch include:

• $NVIDIA(NVDA)$ — the central AI infrastructure supplier.

• $CoreWeave, Inc.(CRWV)$ and $NEBIUS(NBIS)$ — financing-sensitive AI cloud providers.

• $Micron Technology(MU)$ — a direct memory beneficiary.

• $Broadcom(AVGO)$, $Marvell Technology(MRVL)$ and $Arista Networks(ANET)$ — networking and custom-infrastructure exposure.

Being part of the Nvidia ecosystem does not guarantee that every company benefits equally. Balance sheets, customer concentration and margins still matter.

Tiger View

The AI demand story remains intact. Hyperscalers such as Microsoft, Amazon, Alphabet and Meta have real cash flows and continue to spend heavily on infrastructure.

The $500 billion financing initiative also relies primarily on third-party capital and independent underwriting. It is not simply Nvidia handing customers money to buy its own chips.

The debate is now about the quality and funding structure of incremental demand.

Nvidia is evolving from a chip supplier into a financial coordinator for the broader AI ecosystem. That may expand its addressable market, but it also creates new exposure to customer credit, project economics and residual asset values.

Friday’s 4.6% decline should therefore be viewed as a combination of higher rates, profit-taking and renewed scrutiny of circular financing—not just a reaction to Warsh’s speech. Reuters

What is Nvidia’s biggest risk from here?

A. Financing pulls future GPU demand forward
B. Customer credit and guarantee exposure
C. Falling gross margins
D. Interest rates remain the main driver

This post is for discussion purposes only and does not constitute investment advice.