A look at Nvidia's latest earnings report; Michael Burry says Nvidia is 'ratcheting up the risk' with its AI spending
Every quarter, I like to take a look at chipmaker Nvidia's (NVDA) latest earnings because it's such a bellwether company and stock. With a $4.5 trillion market cap, it's worth $500 billion more than the next most-valuable company in the world, Apple (AAPL).
After the market close yesterday, Nvidia reported another strong quarter...
Revenue soared 20% from the previous quarter and 73% year over year ("YOY"). Adjusted earnings per share jumped 25% from the previous quarter and 82% YOY. They came in at $1.62, handily surpassing estimates of $1.53.
The company also gave strong guidance. Revenue is expected to grow around 15% next quarter and 77% YOY to around $78 billion, well above Wall Street estimates of $73 billion.
Despite these staggering numbers, investors are still worried. The stock was down around 5% this morning on fears of a bubble in AI spending.
Here's a summary from my friend Scott Tashman of Outset Global:
Nvidia is obviously dominating the narrative. The company put up solid numbers (beating in the quarter and guiding ahead of expectations), but this wasn't really a shock, and people are still nervous about the sustainability of the AI capital-expenditures boom. In addition, Nvidia's data-center compute revenue fell (a tiny bit) short of expectations in the quarter, a reflection of the absence of China's H200 AI-chip sales and rising competition.
The competition issue was underscored by this week's Meta Platforms and AMD deal, with the latter company handing over around 10% of its shares to secure the contract, and a report in the Information overnight about how Amazon's $50 billion investment in OpenAI will be contingent on the startup moving a large chunk of its compute to Amazon Web Services, including utilizing Amazon's proprietary AI chips.
Coming into earnings, analysts expected Nvidia to earn around $8.75 per share this year and $10.34 next year.
So with the stock around $185 this morning, it's trading at about 21 times current-year earnings and 18 times next year's.
That appears to be awfully cheap for a company this dominant, profitable, and fast-growing. If Nvidia keeps growing anywhere near the rate it has been the past few years, its stock could double in the next year or two.
But that's a big "if"...
There are plenty of competitors emerging. And even if Nvidia maintains its market share and margins, it's hard to see how AI spending can be maintained at current levels, much less increased.
My old friend Michael Burry of The Big Short fame raised more issues in a post on his Substack today (subscription required). First, he highlights a specific aspect of Nvidia's financials he finds troubling:
The February 2026 Form 10-K shows purchase obligations at $95.2 billion, up from $16.1 billion the same time last year. This is because [Taiwan Semiconductor Manufacturing Company] demanded longer term contracts and cash as it had to build out custom semiconductor fabrication and packaging capacity for NVDA's new technology. That has not been the normal course of business until recently. It is a testament to the extremely complex monsters Nvidia's "chips" have become.
NVDA also warns these obligations are "expected to continue to grow and become a greater portion of supply."
To be clear, NVDA has been forced to place non-cancellable purchase orders well before demand is known. This appears structural to the new trajectory of product development and not temporary.
He then shares this chart, which illustrates how the company's purchase obligations far exceed its cash flows for fiscal 2026, which just ended:
Michael says Nvidia today reminds him of Cisco Systems (CSCO) at the peak of the Internet bubble a quarter of a century ago:
This is not business as usual. This is risk.
Back in 2000-2001, Cisco extended purchase commitments with its suppliers to ensure capacity for that 50% annual growth Cisco expected. When enterprises IT spending and data network spending collapsed almost overnight, Cisco wrote down ~40% of its supply chain obligations and inventory and the stock crashed severely...
The same basic thing happened at Intel, Lucent, JDS Uniphase and other high-flyer "picks & shovels" providers to the 1990s boom.
Nvidia's 70+% gross margins compare to Cisco's ~65% gross margins at the time, and yes that is some greater cushion.
He notes that while companies like OpenAI and Microsoft (MSFT) have peeled back spending plans, Nvidia has been pushing into other business lines. And more spending and obligations are risky:
Investors are not valuing Nvidia as some other stocks today and as some similar stocks were valued in 2000, I believe, because they see that Nvidia's investment proposition is becoming more binary, more cyclical.
Rather than a business that would normally roll comfortably through the hills and valleys of its industry, the weighty supply obligation relative to earnings and cash flow makes a valley more of a potential entity risk for Nvidia. Any downturn, when it comes, will be more severe, perhaps even catastrophic, for Nvidia's earnings and balance sheet.
Michael concludes:
Maybe the cycle is starting to turn, maybe not.
Nevertheless, this is a risky position to be in – one Nvidia would rather not be in, and one that Nvidia's shareholders might assume the largest company in the world would not be in.
I think the biggest risk is that ChatGPT owner OpenAI blows up, which would be a key indicator of the AI bubble bursting, as I've warned in previous e-mails.
My friend Marcelo Lima posted this summary on social platform X of the absurd revenue projections OpenAI is making as it tries to raise $100 billion:
Such a massive capital raise is necessary because OpenAI is losing money like no other company in history, as this X post shows:
I'll repeat what I wrote after Nvidia's earnings report a year ago:
[Nvidia] is the kind of stock I like to pound the table on when it's down at least 50% (if not 75%).
So it's not a stock I would buy today. But in that same e-mail, I also noted that if I owned Nvidia, I wouldn't sell. (On September 5, 2024, I detailed how to handle this sort of "high-class problem" with a stock like Nvidia.)
As I've said many times before, you must let your winners run! And remember to use a fairly tight stop loss to protect your gains.
Best regards,
Whitney
P.S. I welcome your feedback – send me an e-mail by clicking here.
P.P.S. In Monday's e-mail, I asked my readers to share their stories of using GLP-1 drugs to lose weight and improve their health. In part, this was to learn more about these drugs, which I wrote "are some of the most important medical advances in the past century."
It was also an effort to help me persuade a particularly stubborn close friend of mine whom I've been bugging (he might say bullying!) for years to start taking them.
Dozens of readers responded with powerful, insightful, and passionate testimonials, which I shared in Tuesday's and yesterday's e-mails.
Well, I'm delighted to report that it worked – my friend took his first GLP-1 injection this morning!
Here's a picture of us, with the trainer he's going to start working out with to maintain and build muscle mass as he loses weight (which can be one of the downsides to taking these drugs):
(Note that he's holding the tiny syringe that he'll use to inject himself every week going forward – it's a tiny needle, and the shot is less noticeable than a mosquito bite.)
Huge thanks to everyone who helped with this effort!