As the beating heart of the artificial intelligence (AI) boom, Nvidia is a kingmaker in the industry. When it takes a stake in a company, the market pays attention -- so too when it exits a position.
And that's exactly what Nvidia has done with AI data center developer Applied Digital (APLD 1.50%). After acquiring a significant stake in the company, Nvidia just sold all of its 7.7 million shares.
So what does that mean for investors? If the most powerful company in AI just sold its stake, should you sell, too?
NASDAQ: APLD
Key Data Points
Nvidia may be concerned about financials
Applied Digital's growth story is real, and the opportunity is enormous, but the company is walking a financial tightrope to chase it. The company's debt has exploded from $44 million in Q1 of fiscal 2024 to $2.6 billion by November 2025, a staggering increase in a little over a year.
And this isn't a profitable company absorbing leverage from a position of strength. Applied Digital is operating in the red, losing $125 million over the last 12 months. That alone would be enough to raise yellow flags, but the risk looks even worse when you consider where the revenue is supposed to come from.
Applied Digital has a revenue concentration problem
Applied Digital's future hinges on roughly $16 billion in contracted lease revenue, spread over 15 years. Just two companies account for the entire pipeline, with one, CoreWeave, responsible for the lion's share -- $11 billion.
And CoreWeave itself is heavily leveraged, taking on massive debt while operating at a loss. If CoreWeave were to fold, Applied Digital's leases would not be at the top of the capital stack. Bondholders, employees, and other creditors would get paid first. What would be left for Applied Digital could be minimal.
On top of that, CoreWeave's lease agreements include a provision allowing it to walk away from any of its Applied Digital data centers -- completely penalty-free -- if certain conditions aren't met, like significant construction delays. It's unlikely CoreWeave would actually walk away in this case, unless, that is, outside circumstances force its hand -- like softening demand or a deterioration of its financial picture -- or the delay is extreme.
But even if CoreWeave doesn't walk, the termination clauses give it leverage -- the good kind. If Applied can't deliver on time or can't refinance its loan, CoreWeave would have an opportunity to renegotiate and could push for even more favorable terms. That could completely change the math, seriously affecting its future profitability.
Image source: Getty Images.
There could be huge upside
There is an enormous opportunity here. At present, AI data center demand is insatiable, and Applied Digital has something many high-growth AI companies do not: tangible, contracted revenue. And critically, the company has, thus far, delivered its buildouts on time. If everything goes according to plan, the upside is significant. But contracted revenue is only as durable as the counterparty standing behind it.
The bottom line
I see why bulls believe in Applied Digital. The growth narrative is definitely attractive. But for my money, its reliance on extreme leverage and its extreme customer concentration leaves no room for error. Any significant deviation from management's projections -- delayed timelines, financing issues -- could be devastating rather than merely disappointing.
And while we can't know exactly why Nvidia sold, it's a company with deeper insight into the AI infrastructure pipeline than perhaps anyone else on Earth, so its decision to exit entirely is a signal worth taking seriously.
For most investors, I wouldn't recommend Applied Digital. And for those who own shares, unless you have a very high appetite for risk, I would consider just how much conviction you have in the company's ability to execute and in AI at large to deliver on the hype.
