Jul 8, 2026 · 24 min · 10 segments
In this episode, Evercore’s Marc Harris, Amit Daryanani and Mark Lipacis sit down to discuss the AI trade: where we are in the investment cycle, how it compares to previous cycles, what has changed…
Mark LipacisGuest
Mark HarrisHost
Amit DaryananiGuest
And maybe, you know, let's start with you, Mark, because you're at the center of it with obviously the ups and downs.

Give us a little bit of a feel for what you think has changed, evolved, and where are we now?

Yeah, I think if you look at the history of computing eras, be it mainframes or minicomputers or PCs or cell phones, you see Each era typically lasts 15 to 20 years.

And, you know, if you think chat GPT really kind of hit the market about three years ago, you could argue that we have another 10 to 15 years to go in this AI investment cycle.

Now, that said, these 15-year cycles are not monotonically increasing functions, right, particularly on the stocks.

And I think one important development we have observed is that the biggest spenders have increased their capex so much that their own free cash flow has started to decline and some are forecasted to go negative.

And I think now at the same time that's happening, token growth, demand for AI tokens has gone from exponential to parabolic.

And then the hyperscalers are now starting to come market to raise cash through debt and equity offerings.

And so I think we're entering a different part of the cycle where it's no longer getting funded purely out of free cash flow.

But also you've got the biggest spenders coming into the having to come to the public markets.

And I think that what that means is we're going to see a decelerating investment curve in the group.

And maybe, you know, let's start with you, Mark, because you're at the center of it with obviously the ups and downs.

Give us a little bit of a feel for what you think has changed, evolved, and where are we now?

Yeah, I think if you look at the history of computing eras, be it mainframes or minicomputers or PCs or cell phones, you see Each era typically lasts 15 to 20 years.

And, you know, if you think chat GPT really kind of hit the market about three years ago, you could argue that we have another 10 to 15 years to go in this AI investment cycle.

Now, that said, these 15-year cycles are not monotonically increasing functions, right, particularly on the stocks.

And I think one important development we have observed is that the biggest spenders have increased their capex so much that their own free cash flow has started to decline and some are forecasted to go negative.

And I think now at the same time that's happening, token growth, demand for AI tokens has gone from exponential to parabolic.

And then the hyperscalers are now starting to come market to raise cash through debt and equity offerings.

And so I think we're entering a different part of the cycle where it's no longer getting funded purely out of free cash flow.

But also you've got the biggest spenders coming into the having to come to the public markets.

And I think that what that means is we're going to see a decelerating investment curve in the group.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.