Stay informed with free updates
Simply sign up for our Tech Department’s myFT Digest and get it delivered straight to your inbox.
It’s too early to talk about the curse of OpenAI, but I’m going to talk about it anyway.
Since September 10, when Oracle announced a $300 billion deal with the chatbot maker, its stock’s market capitalization has fallen by $374 billion.
OK, yes, market caps don’t work that way and post headlines are clickbait. But neither is completely wrong, as Oracle’s comparable stock prices have changed little over the same period (Nasdaq Composite Index, Microsoft Index, Dow Jones U.S. Software Index). Oracle’s “amazing quarter” actually cost as much as Dell or eBay.
As MainFT reported last week, investor concerns stem from Big Red’s bet on OpenAI with its debt-financed data farm. There is nothing to add to this report other than the graph below that shows how much Oracle has effectively become OpenAI’s US public market agent.
OpenAI is rushing to define Discover AGI, the theory being that Oracle can independently scale the computing power it needs. Because Oracle is a data center tenant rather than a landlord, it promises the lowest upfront costs and fastest path to revenue among hyperscalers.
Alternatively, Oracle doesn’t have as much operating profit as its competitors, so it is committed to supporting one large customer in exchange for an IOU.
At its analyst day in Las Vegas last month, Oracle said it aims to generate $166 billion in cloud computing revenue by 2030.
© Oracle Presentation
To get there, Oracle’s capital spending budget for the fiscal year ending in May is $35 billion. The consensus is that annual capital spending will level off at around $80 billion per year in 2029, with revenues continuing to increase thereafter.
And starting in 2027, the majority of revenue will come from OpenAI.
However, Oracle’s net debt is already 2.5 times ebitda, has more than doubled since 2021, and is expected to nearly double again by 2030. Cash flow is expected to remain negative for the fifth consecutive year.
So while the OpenAI contract more than amortizes capital, expansion risks from underfunding remain, and the cost of hedging Oracle’s debt is at a three-year high.
I should add the usual caveats. Credit default swaps are not very liquid. The increase in demand for Oracle CDS comes after an $18 billion bond sale in September. CDS premiums in the low 100 basis points range are less attractive. And some of the companies on the other side of the deal aren’t mugs. Even so, it’s sharp.
Beyond the charts, the broader question revolves around whether the deal with OpenAI is still worth announcing.
A few months ago, any deal with OpenAI could send the stock price higher. OpenAI has flexed its muscle to reflect glory, most notably in October when it acquired AMD warrants as part of a chip deal, boosting its stock price by 24 percent.
But Oracle isn’t the only laggard. While Broadcom and Amazon have both fallen on news of the OpenAI deal, Nvidia has remained largely unchanged since its September investment deal. It’s meaningless if stock prices don’t rise. While $1 trillion in capital investment in AI may seem like a commitment, investment trends are fickle.
Read more:
— Oracle’s Amazing Jam Tomorrow OpenAI Trade (FTAV)
