Oracle's sudden AI stardom gives the energy of 1999

This version of the story appeared in CNN Business's NightCap Newsletter. Sign up for free to get it in your inbox here.

new york
–

Oracle, a large but generally sleepy cloud computing company, has an absolutely odd day on Wall Street.

Stocks (ORCL) shot more than 40% on Wednesday morning. This is the biggest day jump ever. It was a huge leap, with Oracle co-founder Larry Ellison building $100 billion for under $10 billion, making him the wealthiest person in the world, and hitting Elon Musk for second place.

Catalyst wasn't a flashy product development or a surprising revenue beat. In fact, Oracle's quarterly revenue and profit Below Tuesday evening, Wall Street expectations.

Instead, the fires came from Oracle's outlook for the next few years. According to the Wall Street Journal, it's a big “if”, especially considering that the majority of Oracle's rosy outlook depends on revenue from unemployed Openai, one major customer.

Oracle's outlook “is so enthusiastic that if they got this kind of prediction from a less established company, they might have shrugged it off as either a lie or a false digit,” said Steve Sosnick, Chief Strategist of Interactive Brokers.

Here's what's important to power Oracle's inventory with clips that have not been experienced since the dot-com bubble era in the late 90s, when it rose nearly 600% in nearly a year before returning to Earth by 2002.

Oracle, which sells database software, has infiltrated investors a bit quietly this year by securing AI Gold Rush investors to deal with AI companies that boast computing power. (If the semiconductor giant Nvidia (NVDA) is the current frenzy of “Pick and Shovel” play, think of Oracle as a Levi Strauss play. It doesn't just provide durable pants, rather than mining gold.)

If Oracle's head spinning projection seems to be better than true, then it's all part of the fanhouse mirror effect of the generated AI bubble (yes, I said “bubble”). AI customers, including Openai, need to make a lot of money, as any of Oracle's future predictions make sense. That ChatGptMaker doesn't show a clear path to what you do right away. (The information last week reported Openai's forecast cash burning to reach $115 billion by this year 2029 is about $80 billion more than the previous company.)

Like other big technology names, Oracle bets much of the future on its promise that demand for computing power will continue to grow as a generative AI adoption of some kind, yet undecided revolution. Thus, tech companies are spending hundreds of billions of dollars to build huge, energy-sucking building data centers full of computer servers, ensuring the technological infrastructure that the US provides all its AI magic.

According to Renaissance Macro Research, gambling in infrastructure was extremely large, overturning consumer spending this year as a major factor in GDP growth.

“This data center buildout continues to be a huge support for the US economy… so I hope Larry Ellison is right and that this massive buildout is sustainable.”

But Boockvar rang out a note of caution. “Oracle just knocked the cover off the ball, but seeing the market capitalization of such epic proportions one day increases, I can't imagine what we saw in 1999.”

(ahemthe 1999.com crash began. )

Oracle's capital expenditure is “really extraordinary” with $35 billion this fiscal year, roughly 52% of its revenue. In 2024, that was 13% of the company's revenue. “We have never seen such capital strength from these previously large, fast-growing businesses.”

In other words, Oracle is a huge company and has never spent this kind of money before.

Of course, the risk here is that Openai, a large customer of Oracle, doesn't offer it.

Generation AI, the engine of ChatGpt, is one of the rare technologies that can degrade marketability over time. The more ordinary people encounter AI in their lives, the more they associate it with “slops” in their Facebook feed. Chatbots cannot reliably respond to human questions, and they have a nasty tendency to drag humans into delusional, sometimes fatal, spiritual spirals.

Certainly it's not completely useless, but AI advocates have had a very difficult time building applications that cater to their hype (and certainly endured the noble ratings that favor American tech companies).

Without a game-changing technology update that will dramatically reduce costs or dramatically increase profits, Openai could be a toast. And it's not just Oracle, The technology sector is wider.

If Oracle can post a landing, Sosnick said:

“Even so, you're right to point out the risks inherent in a complete reassessment of Oracle's markets. Oracle's shareholders are not only crucially dependent on companies that meet that guidance, but also the wider market.”