“So when you combine all of these pieces together, the overall margin profile of OCI continues to strengthen and grows rapidly.”

In context

“ing to understand is that in these AI data centers, whether it be for inferencing or for training work, the only thing being procured is not AI accelerators. There is a lot of general-purpose compute. There is a lot of, whether it be high performance or large-scale blob storage. There is load balancing. There are identity and security products, etc. That is typically on the order of 10% to 20% of the total spend that ends up going to adjacent services. And when you factor that in, which has higher margins depending on the mix of services, overall profitability continues to improve. And that is without taking into account, as I mentioned earlier about our multicloud database business, that that is a much higher-margin business—more in the 60% to 80% range. It is growing very, very rapidly. So when you combine all of these pieces together, the overall margin profile of OCI continues to strengthen and grows rapidly. The thing I would say—the question that I think underlies this that maybe people do not understand—is the limitation on the profitability is not on the capacity we have delivered. So let us say that I am building a data center and it has four data halls, and I deliver the first data hall. That one is profitable. The reason we are not even more profitable right now, despite the fact that we are continuing to grow EPS, etc., is because we have so much under construction at one time, and we have some expenses for those things. Now we are really good at that. We are very, very good at minimizing the time under which that construction is happening. We are very, very good at reducing those costs during that time period. But they are not zero. And so as our business is going through this hypergr”