Oracle’s struggle with capacity meant they made the difficult but responsible decisions

During an earnings call late on Wednesday, Oracle executives boasted/complained about their demand far exceeding capacity. That is a nice problem to have, but it is absolutely still a big problem.

“We recently got an order that said ‘We’ll take all the capacity you have, wherever it is. It could be in Europe, could be in Asia, we’ll just take everything,’” said Oracle chairman Larry Ellison. “We never got an order like that before. The demand is astronomical.”

Oracle CEO Safra Katz echoed Ellison’s sentiment: “I am still in a position where our supply is not meeting our demands. We are still waving off customers or scheduling them out into the future so that we have enough supply to meet demand. This is a situation that we have not seen in our history, and the numbers themselves are so enormous.”

Such claims might normally be seen as sales-motivated hyperbole, but analysts think that Oracle execs may be playing it straight this time.

Scott Bickley, advisory fellow at the Info-Tech Research Group, said Oracle management faced a no-win scenario and they chose the more fiscally conservative route. 

“Oracle’s build-as-you-go approach to AI data centers may have put them behind the curve at a critical moment. While hyperscalers were investing aggressively in site acquisition, power provisioning, and manufacturing capacity well ahead of the demand surge, Oracle appears to have waited for confirmed orders before scaling,” Bickley said. “That strategy might reflect financial discipline or be influenced by the debt burden from the Cerner acquisition, but the result is clear: they’re now facing significant delivery delays. With GPU availability no longer the issue, the bottleneck has shifted to physical infrastructure. Enterprises relying on Oracle for AI workloads may need to reconsider timelines or diversify providers while this capacity gap persists.”

IDC President Crawford Del Prete agreed, and said that Oracle senior management made the right move, despite how difficult the situation is today.

“Oracle is being incredibly responsible here. They don’t want to have a lot of idle capacity. That capacity does have a shelf life,” Del Prete said. CEO Katz “is trying to be extremely precise about how much capacity she puts on.”

Del Prete said that, for the moment, Oracle’s capacity situation is unique to the company, and has not been a factor with key rivals AWS, Microsoft, and Google.

During the investor call, Katz said that her team “made engineering decisions that were much different from the other hyperscalers and that were better suited to the needs of enterprise customers, resulting in lower costs to them and giving them deployment flexibility.”

Oracle management certainly anticipated a flurry of orders, but Katz said that she chose to not pay for expanded capacity until she saw finalized “contracted noncancelable bookings.”

She pointed to a huge capex line of $9.1 billion and said, “the vast majority of our capex investments are for revenue generating equipment that is going into data centers and not for land or buildings.”

The lack of capacity issue is being addressed, as Oracle has recently started making those building purchases. 

“We have building partners who charge us rent once they’ve finished constructing things. And when we all of a sudden have a higher capex, it means we are filling out data centers, and we are buying components to build our computers, which are different than other people’s, and we are putting them on the floor,” Katz said. “We had an opportunity to buy up for deployment, and so we did. And we are putting out as much capacity as we possibly can, as quickly as we can. We don’t build unless we’ve got orders for our capacity to be built out.”

Del Prete said that Oracle truly does approach manufacturing differently than others, which is why this shortfall is likely unique to the company.

Oracle uses “a complicated process that involves custom components, manufacturing, and off the shelf components. It also requires people to bring the systems online and to onboard the customers. So it’s understandable that this is not simply a GPU issue,” Del Prete said. “Keep in mind that Oracle is a bit unique in the infrastructure space. Oracle systems use many ASICS and FPGAs that are custom to their designs. These chips are used for things like [hardware] acceleration for things like offloading SQL and analytics workloads. Yes, using custom parts is true for other manufacturers, but for Oracle, it’s to a greater degree. This also contributes to the long lead time.”

Del Prete said that one background factor here is that everyone in Silicon Valley knows that sales management tends to inflate how likely it is that a specific sale will materialize and the dollar level where it will materialize. 

“Everything is always unicorns and rainbows from the salespeople,” Del Prete said, explaining why CEOs have to take those predictions with a few tons of salt.

“[Katz] did the right thing, but that right thing ended up being a champagne problem. The bigger problem would have been building out that capacity and then not being able to fill it,” Del Prete said. 

That said, Oracle still has a big problem. They have enterprise customers who want to buy — who need to buy — but they can’t accept their money. Those enterprises are being asked to wait, but if their CIOs feel like they can’t wait, this situation is sending them into the arms of AWS, Microsoft, or Google. And once they make the move, those enterprises might choose to stay there.

Oracle “will look for clever ways to keep customers engaged,” Del Prete said, perhaps by offering them “favorable terms on the backend of the contract or other services like SQL offload, advanced analytics.”