E125: SpaceX launch, Fox News settlement, "Zombie-corn" exodus to AI, late-stage implosion

Starship’s first integrated test cleared Max Q and reached 39 km before destruction, giving SpaceX crucial flight data from a vehicle Elon Musk had privately described as roughly a 50-50 bet to leave the pad. The next vehicle already had more than 1,000 design changes before engineers incorporated this flight’s data.

If SpaceX achieves full Starship reusability, Gavin Baker estimated it could put more than 100 metric tons in orbit for under $2 million in variable cost, versus Falcon 9’s 17 tons for roughly $15 million. That implies about a 50-fold reduction in cost per unit of payload.

Chamath Palihapitiya said a “bleeding-edge” AI company cut one-third of its workforce after replacing that work with an agent trained in six weeks. He treated the anecdote as evidence that boards may soon demand AI-driven cost reductions before approving more capital.

The economics of AI startups may break the venture model built around large rounds: the hosts argued two or three people can now do work that once required 20–30. That favors hundreds of thousands or low millions in funding, while multibillion-dollar funds may struggle to deploy capital efficiently.

Chamath estimated that 70–80% of Series C-and-later startups could now be worth less than their preferred-investor stack, putting founder and employee common equity at risk. He expects trapped talent, plus investors avoiding recapitalizations, to migrate toward fresh early-stage AI companies.

David Sacks described a portfolio company with roughly $20–30 million ARR and about 50% growth that failed to get a growth-financing term sheet and instead sought a $20 million pay-to-play round. Existing investors had to fund their pro-rata share or face roughly 10-to-1 dilution.

Chamath said his firm marked one portfolio company down by two-thirds after two private-equity firms independently valued it near the same level, then sold its stake. Other investors in the same company left their marks unchanged, illustrating how reported private-market valuations can lag executable prices.

Chamath’s team abandoned plans to raise a new fund despite interest from sovereign-wealth investors because he could not confidently value existing private assets or decide whether future checks should be hundreds of thousands or hundreds of millions. He called it his first such period of uncertainty since moving to Silicon Valley.