Chip Stocks Crash, $20B Fund Margin Called, Frontier Labs: SLOW DOWN AI, Mamdani's Grocery Stores
Chamath Palihapitiya, Jason Calacanis, David Sacks, David Friedberg
1 Aug 20266 min read1h 45m
TL;DR
Leopold Aschenbrenner's $20B hedge fund was margin called after running 3.5x leverage on chip stocks during a 20%+ NASDAQ semiconductor crash, with Citadel buying his entire book. The All-In hosts debate whether this is a momentum correction or a signal of deeper macro rot — 30-year Treasury yields hitting 5.2%, persistent inflation, a $2T deficit, and China commoditizing AI models all threaten the thesis that AI productivity gains will bail out US fiscal problems. Sacks and Friedberg argue the AI capex boom is fundamentally real, but leverage is the only way smart people go broke.
Key Moments
David Sacks
“Leverage is the only way that smart people go broke because, you know, if you're not using leverage, your portfolio would just be down 30% this month and then it would already be up 7% today.”
Sacks explaining why Leopold's brilliance didn't save him from a margin call during the chip stock crash.
“Of those 1.2 million levered accounts, somewhere around 350,000 of them were fully liquidated already. And so again, two weeks old, that's so as of today, the number is much bigger, right? So it could be closer to a million accounts fully liquidated today.”
Chamath contextualizing the South Korean retail margin call crisis that amplified the global chip stock crash.
“If China says, you know what, we're actually going to delete that for you and all the value is going to sit with energy, which is what we have a lot of and the stuff that they make, then we're going to end up acrewing a lot of that value.”
Friedberg warning that China's open-source AI and chip manufacturing push could wipe out the US model layer's projected economic contribution.
“There is some incredible efficiencies that I think are about to be demonstrated which effectively cut token consumption by about 50 to 75% for the same task.”
Chamath teasing an unreported AI efficiency breakthrough he believes will dramatically lower the cost of intelligence and reshape market assumptions.
All-In is a weekly podcast hosted by tech investors Chamath Palihapitiya, Jason Calacanis, David Sacks, and David Friedberg. The four 'besties' cover the biggest stories in tech, politics, and markets with unfiltered takes. The show regularly ranks among the top podcasts globally.
Takeaways
1
3.5x leverage turned a 25% drop into ruin Leopold Aschenbrenner grew $225M to $20B riding AI chip stocks, but running 3.5x leverage meant a 25% market move was amplified to ~75% loss, triggering a margin call. Citadel bought his entire public book. The lesson: leverage creates an automatic, one-way ratchet where prime brokers close you out with no recourse.
2
30-year Treasuries at 5.2% kill the case for 50x chip stocks The 30-year Treasury yield crossing 5.2% — a 20-year high — means investors can earn ~8-9% pre-tax equivalent risk-free from the US government. Friedberg argues this structurally destroys the incentive to pay 50-100x earnings for semiconductor stocks, and with a $2T annual deficit and no debt ceiling, yields could keep rising.
3
China's lithography entry crashed ASML 17% in a day Chinese company Aishanga began mass-producing lithography machines that compete directly with ASML, sending ASML stock down 17%. Simultaneously, Chinese memory maker CXMT surged ~500% on its IPO debut, pressuring Micron and Samsung. China is attacking the chip supply chain from both the software (open-source models) and hardware (lithography, memory) sides simultaneously.
4
South Korea's retail margin call dwarfs Leopold's blow-up 1.2 million leveraged retail accounts in South Korea were hit with margin calls during the chip crash, with roughly 350,000 fully liquidated — and that figure was already two weeks old at time of recording. This suggests the real number may approach 1 million fully wiped-out accounts, representing a significant percentage of South Korea's investing population.
5
China's open-source AI could delete the US model layer's value China releasing cheap open-source AI models (90% cheaper tokens) threatens to commoditize the frontier model layer that was supposed to drive US economic growth for the next 30 years. If model value collapses and shifts to energy and compute infrastructure — areas where China has advantages — it removes the key backstop thesis that AI productivity will solve the US fiscal crisis.
6
Hot money dynamics mean late fund investors got wiped out When a fund grows from $200M to $20B, the early investors made 10x — but billions in 'hot money' piled in near the top. Those late arrivals experienced the full crash with no prior gains to cushion it. Sacks notes this is a structural feature of runaway fund success: the majority of AUM often enters at the worst time.
7
AI token efficiency may drop 50-75% — same output, far less compute Chamath teased an unreported breakthrough he expects will cut token consumption by 50-75% for equivalent tasks. If true, this would dramatically lower AI infrastructure costs, reduce the capex needed to scale intelligence, and further compress the revenue model for cloud AI providers — while accelerating adoption across the economy.