Nvidia just released its latest earnings report and, of course, it beat Wall Street analyst expectations once again, triggering algorithms’ FOMO buying in after-hours trading. Fun fact: Have you noticed that for the last 9 quarters, Nvidia has ALWAYS beaten analyst expectations for revenues and revenue guidance by $2 billion? That’s a bit of a long streak to be considered a coincidence, right?
Take a look at what just happened and why it is becoming very obvious how Nvidia is crafting its earnings to continue pleasing Wall Street. As you can see here, without the amount of revenue growth made possible by financing clients’ purchases, as represented by the growth in accounts receivable, Nvidia would have missed revenue expectations by $3.49 billion.
What would have happened if Nvidia’s revenues had missed expectations? Well, you know…
Moving on, before we move into the juicy parts of today’s analysis, I cannot avoid bringing attention to the skyrocketing of Nvidia’s accounts receivable and inventories.
Once upon a time, such a pattern, combined with a slowdown in revenue growth, was considered a pretty big red flag about a company inflating the size of its business. I guess today, it isn’t the case anymore.
What if I tell you that, on top of what I just highlighted, Nvidia’s prepaid supply and capacity agreements have been DECREASING for several quarters now?
Honestly, doesn’t all of this clash with the generally accepted narrative that demand is overwhelming and supply is tight? Shouldn’t Nvidia GPUs be flying off the shelves rather than piling up in its warehouses, and shouldn’t the company be paying up its suppliers to book as much production capacity as possible?
The rabbit hole gets deeper, though. While at this point nobody questions Nvidia’s circular financing schemes to increase its revenues, why nobody raises questions about Nvidia’s revenue concentration is a mystery to me.
In the last 9 months, two clients accounted for 34% of Nvidia’s Data Center and computing revenues. In the same period last year, three clients accounted for 36% of that.
What about geographical distribution? This quarter, Nvidia changed its revenue recognition method from considering the billing country to considering the clients’ headquarters location, restating all previous figures. The result? Considering that it also disclosed that 86% of the revenues accounted for geographically as sold in Taiwan are for products delivered between Europe and the US, Nvidia’s sales to the US are going to be between $124 and $98 billion, equivalent to 84% and 66% of the total.
Fair question: why is no one else in the world outside the US rushing to buy so many GPUs? You can find the answer in my previous report: “THE DATA CENTERS FRENZY WILL BE REMEMBERED AS THE LARGEST WASTE OF CAPITAL IN HISTORY“.
Aren’t these numbers incredibly odd for a company almost worth $5 trillion? Yes, and this has been the case for a long time, although clearly, the majority of people did not care.
To get a sense of how big Nvidia’s circular financing scheme is growing, on top of the incredible growth in accounts receivable and inventories, we should take a look at the growth in Nvidia’s multi-year cloud service agreements that, as a matter of fact, represent Nvidia’s commitments to buy computing capacity from customers who buy its own GPUs. This amount literally exploded to $25 billion.
If you still struggle with understanding what’s going on here, I suggest reading my guide “HOW TO FABRICATE REVENUES FOR DUMMIES“.
Now let’s get to the most juicy part. Do you remember the big announcement about the $100 billion partnership between Nvidia and OpenAI back in September that triggered several hundred billion-dollar gains in Nvidia’s market capitalization that day (“Nvidia to invest up to $100 billion in OpenAI“)? Apparently, according to Nvidia’s fresh disclosures, not only has the letter of intent not been formalized into a legally binding agreement after two months, but there is a chance this might never happen!
What the hell is going on here? Is all that’s happening a giant game of smoke and mirrors? You tell me.
This isn’t even the best part yet. It is public now how CoreWeave, one of Nvidia’s key clients, investments, and partners, is struggling; however, what isn’t public is that, after Nvidia agreed to buy from CoreWeave any unsold computing capacity up to $6 billion until 2032 (“CoreWeave, Nvidia sign $6.3 billion cloud computing capacity order“), Nvidia now also agreed to directly finance CoreWeave’s build-up of data center infrastructure. How can I say this? This is another fresh disclosure from Nvidia’s earnings report:
“In the third quarter of fiscal year 2026, we entered into an agreement to guarantee a partner’s facility lease obligations in the event of their default. The agreement allows our partner to secure a limited-availability facility lease backed by our credit profile, in exchange for issuing us warrants. The maximum gross exposure is $860 million, which is reduced as the partner makes payments to the lessor over five years. The partner has placed $470 million in escrow and executed an agreement to sell the data center cloud capacity, mitigating our default risk. If the escrow and cloud capacity agreement are not sufficient to cover an event of default, we have the option to assume the lease for internal use or sublease. The guarantee, classified as a credit derivative with changes in fair value recognized in Other income and expense, has an insignificant fair value.”
It would be a hell of a coincidence if CoreWeave’s total amount of “restricted cash” in its non-current assets is exactly $470 million, right? Well, take a look at what the amount is that they just reported in the last earnings.
How big a deal are these commitments and guarantees becoming for Nvidia’s business? The company answers it in another fresh disclosure:
“We have entered into, and may in the future enter into, commercial arrangements, including long-term capacity purchase obligations and financial guarantees supporting our customers’ and partners’ buildout of datacenter infrastructure. These arrangements expose us to counterparty risk, including customers’ or partners’ inability to secure necessary financing or infrastructure, significant project delays, or financial distress or insolvency. Despite our efforts to mitigate these exposures, if triggered, these obligations could require payments that may negatively impact our business, financial condition, or results of operations.”
If this doesn’t give you an idea of how fundamental circular financing is for the whole business of Nvidia, and to justify its totally nonsensical valuation, I do not know what else you need at this point.
To conclude, if the company sought to reassure investors that there is no overspending for AI and its whole business and growth are sustainable, it surely did not do a great job, leaving all these questions unanswered. However, nowadays, people neither read financial earnings nor ask questions anymore, so I guess the whole charade will carry on for a little longer.
SYNNAX PROMO
Get exclusive intelligence on stocks you can’t find anywhere else – insights that could give you a serious edge. Claim your free Synnax trial month now: https://synnax.app/PROMO/JUSTDARIO