Saturday, July 20, 2024

Oracle’s AI infrastructure arm is booming, with billions of dollars of business yet to come online

Must read

(Oracle )

Oracle’s riding the generative AI wave as demand for its Oracle Cloud Infrastructure (OCI) continues to grow. Or as CTO Larry Ellison puts it: 

Oracle’s Gen 2 AI infrastructure business is booming. That’s become pretty clear to everybody.

For Q3 2024, Oracle turned in total revenue of $13.3 billion, up seven percent year-on-year. Infrastructure-as-a-Service (IaaS) was $1.8 billion of that, up 49% year-on-year. Net income was $2.4 billion. Other stats of note: 

  • Total cloud revenue, including Cerner, was $5.1 billion, the first time that total cloud revenue has topped total license support revenue. 
  • Cloud application revenue was up 14% year-on-year to hit $3.3 billion. 
  • Fusion Cloud ERP and NetSuite Cloud ERP each reported $800 million, up 18% and 20% respectively. 

OCI demand

But it was OCI that was picked out for attention by CEO Safra Catz: 

OCI has emerged as the largest driver of our overall revenue acceleration, growing much, much faster than our cloud competitors. Customers have figured out that by moving to OCI they can really get more while paying less, but it’s not just the cost that matters to our customers

There are several drivers for customer adoption, she suggested: 

First, we know better than anyone what it takes to run the full stack of technology that goes into mission-critical workloads….Secondly, our AI capabilities are unique as they’re built-in to help customers drive business outcomes. This is more than integrating generative AI across our Fusion and Industry Cloud applications and autonomous database, which we have done. It’s also about enabling and refining these AI models with the customer’s own data to better understand and serve their operations without them losing control of their own data.

Third, we provide deployment flexibility for customers based on how they want to run in the cloud. In addition to offering public cloud services, we remain the only vendor which also offers a dedicated and complete cloud of customer, dedicated regions, sovereign clouds, and Alloy, our partner cloud so customers don’t have to compromise the services they receive while meeting their deployment needs.

And finally, we provide multi-cloud offerings so customers can consume our cloud services in the public cloud of their choice. We offer Oracle database at Azure with Microsoft as well as MySQL HeatWave through multiple clouds and you can expect more multi-cloud services to come.

Oracle continues to focus on getting more capacity online to meet demand, she noted: 

I think there are many, many customers who have come on and that haven’t gotten capacity yet. We’ve got at least 40 new AI bookings that are over $1 billion that hasn’t come online yet.

The company is building some of the largest data centers in the world, added Ellison: 

We’re building an AI data center in the United States where you could park eight Boeing 747s nose-to-tail in that one data center…So, we’re bringing on enormous amounts of capacity over the next 24 months because the demand is so high, we need to do that to satisfy our existing set of customers.

It’s inevitably taking time to build out these data centers, but Ellison argued: 

We are getting very good at building them quickly and getting the power and the communication links in. We’re doing it faster than [has] ever happened in the past. And the thing is once we deliver the hardware, the hardware comes up very, very quickly because the process of bringing up the hardware is now automated. It’s very different than it used to be. So, we’re able to bring additional capacity online very quickly if we have that the electric power and the communication lines. So, the long pole in the tent is actually building the structure, connecting the electricity, connecting the communication lines.


In addition to selling infrastructure for training AI Large Language Models, Oracle is also “completely re-engineering” its industry-specific applications to take advantage of generative AI, according to Ellison: 

The best example of this is in Healthcare where Oracle did not just add a bit of AI around the edges of the existing applications. Instead, we developed completely new applications using our Apex Application Generator and our Autonomous Database. These all new applications use generative AI throughout the application. 

He cited a new Ambulatory Clinic System which is being delivered to customers in the current quarter: 

This completely new application features a voice interface called the Clinical Digital Assistant. The Clinical Digital Assistant listens to a doctor’s consultations with a patient and automatically generates prescriptions, Doctor’s orders, doctor’s notes, then automatically updates the patient’s electronic health records. The Clinical Digital Assistant’s voice interface makes our new healthcare systems dramatically easier to use and saves hours of doctors precious time every day which now can be spent with patients rather than typing into a computer. 

The kind of next generation applications that are emerging are “very attractive”, he said, citing one generative AI example to speed up the process of joining the European Union (EU): 

It took Serbia eight years to harmonize their laws to be able to join the EU. Albania is facing the same thing. But with generative AI, we can read the entire corpus of the Albanian laws and actually harmonize their laws with the EU [in] probably more like 18-months to two years rather than the eight years it took Serbia. So, there are all sorts of interesting new AI applications out there that people you’ve probably never heard of before or at least I hadn’t heard of before, until this last 12 months. Now that we’ve worked on and we’re now in the process of delivering.

My take

A very strong quarter that made Wall Street very happy, with a hint of more to come, not least from Catz who teased: 

We expect to have some very nice joint announcements with NVIDIA next week.


Latest article