The Oracle of the AI trade

The aftermath of tech-earnings season is sort of like Santa’s workshop on December 26: Before any relief at a job well done can set in, it’s already time to look ahead to all the work for next time.

Oracle’s earnings after the bell yesterday gave investors objectively good news: The company’s total quarterly revenue grew 30% to reach $19.3 billion, beating the $19.13 billion forecast by analysts. All of that growth was fueled by—you guessed it—growing AI demand. Oracle’s cloud-infrastructure revenue grew 121% year-over-year to $7.4 billion, while total cloud revenue grew 62% to $11.6 billion. But even more impressive was management’s claim that demand for Oracle’s AI training and inferencing capacity is growing faster than the company’s ability to supply it.

On top of that, Oracle’s remaining performance obligations (RPO) showed that it has customers lined up around the block. “The key growth driver is the RPO conversion to revenues, related to new additions of datacenter capacity,” explained Bank of America analyst Tal Liani in a note today. He also reiterated his Buy rating and $240 price target, nearly 60% higher than where shares trade today.

New quarter, same old fears

On the heels of the earnings announcement, shares popped 7% in after-hours trading yesterday. But the stock sank 1.77% today as investors focused on the negative: While Oracle didn’t up its capex guidance over the quarter, it still has a full-year capex target of between $90 billion and $95 billion, and the company has taken on massive debt to fuel its AI buildout.

That stoked fears among investors that while these AI companies continue to see strong demand, the economics of the infrastructure buildout could still come up short of working out. Even as AI giants boast increasingly bullish quarterly results, the high stakes mean the risks are getting bigger, too.

Now, even green flags look yellow to investors, as they wait to see if gargantuan infrastructure spending will indeed bring in the promised returns.—LB