The hype surrounding artificial intelligence (AI) has reached a fever pitch. While it's possible that the technology will live up to its promise, I think that enthusiasm has outpaced fundamentals.
If that's true and the hype fades, here are two stocks that I would not want in my portfolio.
Image source: Getty Images.
1. CoreWeave is highly leveraged
CoreWeave's (CRWV 2.27%) central role in the AI data center buildout makes it a favorite among those betting that the spending spree will continue. While the "neocloud" operator is on track this year to nearly triple its revenue from 2024, the uncomfortable truth is that this growth is being fueled by enormous amounts of debt.
NASDAQ: CRWV
Key Data Points
The company's model requires it to scale at lightning speed, far quicker than it could with its own cash. CoreWeave's capital expenditures (capex) have far exceeded not just its cash from operations, but its entire revenue for each quarter on record. Over the last 12 months, the company's capex was more than double its sales and 6 times its operating cash flow.
This has led to a balance sheet with more than $10 billion in long-term debt, and this is not cheap debt. CoreWeave is forced to borrow at high interest rates, and it now pays 6 times its gross profit just to service its debt.
The company has an agreement with Nvidia that guarantees it up to $6.3 billion if demand cools enough that CoreWeave has more supply than it can sell. But despite that seemingly massive figure, it's nowhere near enough if the hype fades considerably.
2. Oracle's reliance on OpenAI is not an asset
Much like CoreWeave, Oracle (ORCL 2.05%) is relying on leverage to fund its growth, making it especially risky if the market goes belly up. As a mature company with a substantial business prior to the AI boom, Oracle's numbers aren't quite as dramatic as CoreWeave's, but it's close. The company's all-in AI strategy means that it is expending enormous resources to meet the demands of its AI customers, namely OpenAI, and growing its credit obligations.
NYSE: ORCL
Key Data Points
Oracle has flooded debt markets with tens of billions of dollars worth of bonds this year, with the rate of issuance rapidly increasing in recent months. Oracle's credit default swaps -- essentially insurance policies against the company failing to pay its debts and a gauge for how risky the market believes its debt to be -- have reached their highest point since the 2007-08 great financial crisis.
That is, in large part, because of the company's reliance on one customer. Oracle's massive debt raise is funding the construction of data centers intended to primarily serve OpenAI, which is on the hook to pay Oracle at least $300 billion over the next five years. This is just one of OpenAI's many commitments that I am unsure how it will fulfill, even if the AI spree continues.
OpenAI's total annual sales are only in the low double-digit billions. And that is its top line; the company is losing money hand over fist, according to the limited financial information available from the private company.
If the hype fades, Oracle could find itself having taken on massive amounts of expensive debt with little to show for it.
Oracle and CoreWeave are risky bets
Even the most confident bulls can't deny that the scale of spending is, more or less, unprecedented, and that AI hype today feels a lot like dot-com hype in the 1990s.
If the hype fades, just like in past extreme hype cycles, there will be companies that continue to thrive in the aftermath. They won't include CoreWeave or Oracle.
Both companies are "betting the farm," so to speak, rapidly accumulating expensive debt under the assumption that AI demand will not only continue, but accelerate. If AI adoption slows and the hype fades, these are not the stocks you want to own.
