Emily Flippen

But my first interpretation is that this has a lot more to do with Oracle in particular than it does with the bond market.

And this could have broader effects on the tolerance for lenders and AI ambitions.

But look, Oracle is laying a ton of debt to fund this AI ambition.

It maybe doesn't have the best track record when it comes with capital allocation in the past.

So I think lenders are just catching on to the level of risk associated with the deals and with Oracle in particular.

I mean, we're talking about tens of billions in new bonds and project finance loans with Oracle.

So I mean, the good thing about being on the equity side of Oracle is that you have theoretically all of the upside of those investments actually pay out.

And yes, of course, all the downside if they don't.

But bond investors really only get that principal and interest back.

So in my opinion, it makes sense that they're demanding just a

a little bit more of a premium for taking on that additional risk.

This is such an interesting question because this calculation has been centered upon what's the return on investment for artificial intelligence in particular.

This equation has always been, we make all these investment dollars, what am I getting out of it?

The equation is actually, what is the return on investment

for the artificial intelligence investments, really data centers, that's what the debt's being used to pay for, versus what is the cost of the financing.

The focus has been on the former part of that equation, not so much the latter.

But maybe we should all be thinking a little bit more about the latter part of that equation as we look at the bond markets here.

Because even if the former gets a decent return, even if we're getting

eight to 10%, let's say, of a return on our AI investments, is that actually more than the cost of our debt financing?

And even if we're not funding with debt, if we're funding with cash, and there's been some question marks about how much of this is being covered with cash,