Elizabeth Holmes and Theranos, AI Data Center Forecast Jumps, Poppi’s Pepsi Deal | Diet TBPN

The discussion centered on why the Elizabeth Holmes response is so polarized after the Theranos movie: the host argued that Holmes still looks like a bad founder even apart from the fraud, because the product execution was sloppy and the story was framed as magical science instead of a normal diagnostics business. He pushed back on the idea of a Theranos redemption arc, saying he would need to see Holmes demonstrate genuine biotech judgment, not just rehabilitation as a celebrity founder.

The host also used Theranos to contrast real medical-device companies with the mythology around them, noting that diagnostics businesses can be hard science without being magic and that FDA authorization is about market access, not whether a device can exist. He repeatedly suggested that Holmes benefited from the TED-talk-era startup culture, where exaggerated claims were often excused as “fake it till you make it.”

From there, the conversation widened to AI infrastructure and monetization. The host agreed that data-center supply looks very real, pointed to a jump in 2027 capacity forecasts to 50 gigawatts, and said demand is strongest in agentic consumer products and enterprise software, though willingness to pay remains uncertain. The episode closed by arguing that Poppi’s Pepsi deal still looks like a huge win despite a much smaller-than-expected earnout.