Where Capital Meets Compute

How compute gets priced, financed and built beyond the hyperscalers. Hosted by Kony K., a conversation with the industry leaders building & financing AI infra.

[Where Capital Meets Compute @WCMC_Podcast]

Introducing Where Capital Meets Compute: a show about how AI infrastructure actually gets financed: the buildout, the offtake, the capital, beyond the hyperscaler headlines.

Episode 1:
Carmen Li (@carmenli), CEO of Silicon Data (@Silicon_Data) and Compute Exchange, on how GPU pricing indices are built, why compute doesn't fit the standard commodity model, and what the forward curve is signaling right now.

[Where Capital Meets Compute @WCMC_Podcast]

"We don't think anything is important unless the data tells us that way."

Carmen Li @carmenli , CEO of @Silicon_Data and Compute Exchange, on how she turns wildly different GPU specs into one clean price index, no assumptions, just what the data actually shows.

Full episode available now.

[Where Capital Meets Compute @WCMC_Podcast]

GPU isn't quite oil, and it isn't quite electricity. It's somewhere in between, with its own logic.

Like oil, it comes in different grades. But unlike electricity, it holds residual value over time. And unlike oil, it's more elastic, because depending on the workload, compute can move to wherever the resources are, in a way other markets can't.

Carmen Li @carmenli @Silicon_Data explained why GPU pricing needed its own logic entirely.

[Where Capital Meets Compute @WCMC_Podcast]

AI companies can grow fast and still bleed margin, not from demand, but from GPU pricing that swings from $3, to $6, to $9, and back to $3, for the same chip, in the same week.

Carmen Li @carmenli, CEO of @Silicon_Data and Compute Exchange, on why that’s not a founder’s job to predict.

Full episode link in bio.

[Where Capital Meets Compute @WCMC_Podcast]

Everyone asks why NeoClouds are 2-3x more expensive than hyperscalers for the same GPU.

It’s not just markup. It’s demand, packaging, and who can actually get the chips, explained by @carmenli @Silicon_Data

Full conversation link in bio.

[Where Capital Meets Compute @WCMC_Podcast]

@carmenli @Silicon_Data The written recap of our conversation with Carmen Li is up, including GPU pricing indices, the freight-market analogy and diverging forward curves, in one read:
https://t.co/vWZDGUx8br

[Where Capital Meets Compute @WCMC_Podcast]

Introducing Where Capital Meets Compute: a show about how AI infrastructure actually gets financed: the buildout, the offtake, the capital, beyond the hyperscaler headlines.

Episode 1:
Carmen Li (@carmenli), CEO of Silicon Data (@Silicon_Data) and Compute Exchange, on how GPU pricing indices are built, why compute doesn't fit the standard commodity model, and what the forward curve is signaling right now.