And so I think this is great for them.

The question I think then becomes, would it have been better if they worked at Microsoft or at OpenAI?

And I actually think it might be better happening at OpenAI from at least a near term perspective because number one, the pace of innovation doesn't have to slow down.

And number two, all of this is happening at an arm's length transaction.

It's not happening within Microsoft.

So I think OpenAI is free to innovate maybe even faster than they would have had to under Microsoft where there would have been more guardrails.

Not saying this is ever going to happen.

But if OpenAI invents Skynet tomorrow, Microsoft can walk away and say, that's OpenAI that did that.

We didn't do that.

So that's where I think this is actually the best case scenario for Microsoft coming out of this.

- I mean, it felt like everybody's high school crush was back on the market after a long relationship at the end of the day when all of this took place.

All of the companies that were throwing their hat into the ring, trying to get this talent, Salesforce, you had Microsoft trying to woo everybody over there, you had Oracle throwing their hat in the ring.

So all of this considered, I mean, it really sets up for next year even where more of this investment could take place.

To what extent are we looking at some of the potential capital expenditures that could come forward at these arm's length type of deals from mega-cap tech companies next year?

RISHI JALURIA: It's going to be massive.

Remember, the amount of resources required for these generative AI workloads is exponentially higher than a traditional cloud workload.

By my math, it's at least five times higher if not more, especially if you're talking about the training versus the inferencing part of the LLM.

And so there's just a lot of capacity you have build out in terms of actual infrastructure, GPUs.

And the more that companies are trying to figure out what their actual generative AI strategy is experimenting with it and those experiments turn into real use cases, you're going to see, I think, exponential growth in the usage of this underlying platform, which is a great leading indicator for the amount of CapEx that will be required.

I'll also add on, remember, Microsoft is already working on their own competitor too, NVIDIA's GPUs.

I believe AWS is as well.

Google, I've heard is as well.

There's a lot of investment has to become to even narrow the gap.

I mean, what Microsoft has, if you look at the specs, well behind the H100 from NVIDIA.

Not to say that can't close over time, but there's a lot of money that has to be put into capital expenditures.