$4 billion worth of commitments just for AI: Karan Batta of Oracle

As an early adopter of artificial intelligence (AI), Oracle is accelerating its deployment of AI and machine learning technologies across its apps and cloud infrastructure services. During the Oracle Cloud World 2023 in Las Vegas, the company introduced innovative generative AI solutions in industries ranging from healthcare to helping organisations improve customer service. Karan Batta, Senior Vice President, Oracle Cloud Infrastructure (OCI), in an exclusive interaction with Business Today’s Nidhi Singal talks about Oracle’s Cloud Infrastructure Services, multi-cloud and the company’s AI strategy ahead. Edited Experts: 

BT: Can you provide an overview of Oracle Cloud Infrastructure’s services and offerings? How has Oracle’s business transformed with its focus on cloud? What about your on-premises customers? 

KB: Oracle has always been like a massive enterprise company. If you walk into McDonald’s today, you pay for your burger you, well, that POS system is owned by Oracle. It’s called Symphony. So, the point-of-sale system is owned by us, the software that runs on is on our cloud. You walk into a bank and you’re doing a deposit. The core banking system is called Flexcube. It’s owned by Oracle. Almost every bank in India uses that. You get your paycheck. It’s HR systems that’s run by our HR system. You are a salesperson and you’re putting in a customer lead. Well, it runs on NetSuite or our Fusion. We have so many different applications in so many different industries. Oracle’s always been across the board, a behind the scenes enterprise application company. But also, we’ve had obviously Oracle Database, which is our big sort of bread and butter, that powers all these applications as well. Our databases have been kind of the most popular for a long, long time. 

But I think what a lot of people don’t know is all the surrounding applications that we also sell. Now the reality is for a lot of those applications and database, you need infrastructure, computers to run those things on. And we should build our own cloud for that. And so, it just made sense that we would also invest in cloud infrastructure. 

So, we started building a core fundamental compute network storage, like base layer of cloud infrastructure. And you know, that powers everything that we have now from our Fusion, Financials and applications to our database technology to all the different innovations, we’re doing an application. In some sense, think of it as like cloud infrastructure that competes with like Microsoft, or AWS or Google. 

BT: You have stiff competition in cloud with players like AWS, Azure, Google, but you work with them as well. So, what sets Oracle Cloud Infrastructure apart from others? 

KB: That’s a good question. We can talk about differentiators in two different ways. There are customers that have already picked Microsoft for their primary provider. But they’re also massively integrated into Oracle technology. But there’s no path for them to do that. So, they’re like, “Well, I’ve already spent a lot of money on Microsoft, and have got a credit card sitting on a Microsoft account. Can I please spend my credit card dollars on your services?” Because it’s also very hard to move off the core technology once you’re in it. And so that’s our differentiator—our database tech. We can either be adversarial with our customers and be like, sorry, you can only use Oracle database through Oracle cloud. Or we can say, great, you’re using our technology and we’ll allow you to use it from anywhere you want. And that’s the path that we picked. 

But if you want much better scale infrastructure, or like all these other things that we provide we can do that on OCI or Oracle Cloud as well. So, we really were leaving it to customer choice. Our differentiator really comes from price and performance on the core infrastructure. But you know, when you’re talking about $10-$15 million, like 20-30 per cent may not make much of a difference, but when you’re talking about billions of dollars 20-30 per cent makes a big difference. So, we can work together we can also have differentiators where customers can make their choice. 

BT: Isn’t security a big differentiator for Oracle? 

KB: Security is a massive focus for us. We’re an enterprise company, but we do a lot of business in the government space, in the national security space. Out of those 64 regions, a lot of our regions are also government regions or air gapped security regions. So we invest a lot in public sector and restricted markets. And so yeah, that that drives a lot of a lot of capability. 

BT: The challenge enterprises are grappling with today is the adoption of multi-cloud and managing the same. How do you see this space evolving? 

KB: There’s a lot of FUDS around multi-cloud. Could be anything from security and compliance posture to data sovereignty, data residency, residency requirements, to performance latency, to migration challenges, to just inertia. People just don’t want to do it. And it doesn’t help the fact that like none of the cloud providers want to work with each other. They’re pretty much like I’m the one stop shop, you should come to me for everything. And if you move your data out, I am going to slap you with a big fee. 

That’s where like what we’ve done with Microsoft really like fixes that. Like what it shows is that this is not really a blocker, it just requires focus. So, I think the big challenge is around latency. 

When we have 100 regions plus, eventually, we’re in the same location as everybody else. Like if you think about if you compare like all the clouds with us, where all our locations are, I can probably tell you where like 90 per cent overlapped. So technically, the performance or like the latency between those regions is no different to like, if you were to just pick one cloud. Part of it is a lot of FUD. And part of it is like there's commercial and flexibility. I think that will change eventually. 

BT: What are the new developments happening in the cloud ecosystem? 

KB: Artificial intelligence is where I see it headed. I think time will tell if this is something that will drive kind of the next wave of computing. If you look at our last earnings, and Larry Ellison (Chief Technology Officer of Oracle) talked about the fact that just in the last quarter, we have over $4 billion worth of commitments just for AI. 

Right now, our focus is on AI because that’s where we’re seeing our customers really push us. If I was to give you kind of three major areas, Oracle Cloud is focused on, it’s our multi cloud strategy, it’s our core infrastructure, which is customers like big customers like Uber and others are coming to us for. And then it’s our AI strategy that's driving a lot of growth for us right now. 

BT: Oracle has been using AI and ML for long. But what has changed with generative AI? 

KB: It’s funny, people are calling it generative AI but like machine learning has been around forever. So, I think the difference is the hardware has gotten much better. The microprocessors and accelerators like Nvidia, they’ve just gotten significantly better. But also, the data growth has been huge. The data is just like exploded. And well guess what most of that data sits on Oracle Database. We have these like swarms of data, whether it’s health care, public safety, what kinds of data you have, it’s all sitting on Oracle database. And our customers are like, I would love to make more money, more intelligence out of that. And so that combined with the ability to have like large scale, super computing power to like extract information out of that data, train models have really been the big thing. There was a flash in the pan with Open AI, but I think that’s that was kind of the consumer kick you needed to actually focus efforts on it. But this had been happening in the background for a while. 

BT: What is your strategy going ahead with generative AI? 

KB: Our strategy is really focused towards like enterprise customers, not necessarily like consumer focus. We are focusing on three layers—the infra, the service layer, which is a little bit more sort of curated experience, and then the applications. That’s our three pronged strategy for AI. What I would say is that probably the most critical component of that is going to be the infrastructure. 

All the differentiators will really be dictated by the other semiconductors, it’ll be Intel, NVIDIA, ARM making sort of strides and chip manufacturing. And so, whatever that happens there will directly affect our infrastructure. The service layer will really be dependent upon the base models. And I think in the next couple of years or more, there will be like a core set of base models that are just generic and that are just available for everybody to build upon. 

BT: Do you see the adoption for generative AI amongst the Indian clients? 

KB: I don’t know about adoption but I can certainly think that there’s a lot of experimentation going on. But I think there’s a lot of concern around like sovereignty, data residency. Privacy is a big one. I think there's issues there. And that’s where our focus is very much on security and privacy versus like, you know, some random model that's just can answer every question on the planet. So, I think we’re much more focused in that direction, because most of our customer base is really just enterprise and governments. And so, for them to adopt our technology, we really must be focused on security. And so, I would say right now, I think people are waiting and seeing but my guess is they have a lot of ideas. They just don't know how to execute them safely. 

BT: How do you ensure that generative AI remains unbiased? 

KB: Models are susceptible to hallucinations. You don’t want the model playing the doctor or the provider. The human needs to still be in the loop to be able to check. It’s okay asking a model to go book your flights or book your hotel and build your itinerary. It’s a whole different thing when you’re recommending medicines, and you’re writing prescriptions as there’s actual impact on human life. There’s a clear delineation between the two. We’re definitely much more cautious. During the last earnings, we announced X AI, so we will be working with Elon Musk’s X AI company as well to run their models on top of us. And so, we have a lot of focus and effort on making sure that areas where there needs to be a human in the loop, still happens. 

Also read: Amid AI-led cyber threats, upskilling and investments need of the hour for India: Palo Alto Networks' Anand Oswal

Also read: Oracle integrates generative AI in healthcare, unveils new clinical digital assistant