PodBrowser
1000x

MARKET UPDATE: How Situational Awareness Blew Up, Will Rates Nuke Us, What’s Next?

Friday, 31 July 2026 · 4 min read · Listen to the episode ↗

Leopold Aschenbrenner, whose Situational Awareness paper gave him credibility to raise billions, blew up his fund by running roughly 4x leverage across high-volatility AI and memory stocks while owning approximately 25 percent of NBIS, far exceeding the 10 percent of average daily volume guideline.

Leopold Aschenbrenner, whose Situational Awareness paper accurately mapped AI scaling and gave him credibility to raise billions, blew up his fund by taking on roughly 4x leverage across high-volatility AI and memory stocks. He held approximately 25 billion dollars of market exposure against around 30 billion dollars in underlying positions and owned roughly 25 percent of NBIS, far exceeding the guideline of not exceeding 10 percent of average daily volume. At that concentration he effectively became the market in those names and could not exit without moving prices against himself.

The blowup followed a predictable pattern. Aschenbrenner went to the street to raise emergency capital to meet margin calls, broadcasting his distress to competitors. Pod shops identified his book and shorted his positions to accelerate the pressure, turning a bad situation into a forced liquidation. Long Term Capital Management used offsetting positions across different brokers specifically to prevent this kind of exposure, and Aschenbrenner failed to obfuscate his book in the same way. Jane Street, Millennium, and Citadel were the three bidders for the portfolio. Citadel was the most natural buyer because Ken Griffin has publicly stated his next form of alpha is three-to-five year alpha, Millennium's risk limits are far tighter, and Jane Street operates on short time horizons better suited to structured products than straightforward long equities like Bloom Energy, Micron, or SMIC. One speaker estimated Citadel was up three to five billion dollars on the acquired portfolio following the market recovery.

One speaker raised the possibility, while explicitly acknowledging they cannot confirm Griffin's motivations, that Griffin's public comments on national television suggesting rate hikes could happen were timed to stress an already fragile market and accelerate the liquidation, positioning Citadel to acquire the portfolio at better prices. The same speaker suggested Griffin may have shorted Micron before bidding, allowing Citadel to offer the best price on the block. The fund was not shut down and continues to operate as a hybrid public-private fund, managing its public book on a fully paid-for basis going forward. Aschenbrenner was reported in an investor letter to be up 80 percent year to date despite a 67 percent down month and a margin call, though one speaker treats that figure skeptically because gains may be attributable to locked illiquid positions such as Anthropic and 11 Labs rather than liquid market trading.

The lesson drawn was that leverage makes any long-term thesis entirely path dependent. One speaker described running a portfolio at Block Tower from 40 million dollars to approximately one billion dollars, reaching roughly 130 percent net long including options delta during Bitcoin's move from 25,000 to 50,000 dollars, and feeling invincible in a way that mirrors the psychology behind Aschenbrenner's downfall. The prescription is to express a long-term thesis with no more than 2x leverage and to remain structured so the investor can stay solvent if the market stays irrational longer than expected. Charlie Munger's formulation was cited: liquor, ladies, and leverage are the three things that will always sink a person.

Following the forced liquidation, SMIC was up almost 40 percent from its lows and some assets bounced 25 to 30 percent driven by short covering and new buyers. One speaker predicted the market would then experience churn rather than a clean directional move as Citadel and others take risk off after the bounce. One speaker said physical memory supply is under-supplied by at least 40 to 50 percent for the next 12 months and that memory module prices could triple in that period. DRAM traded down to 44 dollars after they suggested nibbling around 57 dollars with a full-in target near 47 to 48 dollars, and was at 51 dollars at the time of recording. They are reallocating from cash back into memory stocks including DRAM, Intel, Micron, and Sandus while explicitly avoiding SK Hynix and South Korean equities after SK Hynix missed earnings. Micron is preferred over Samsung and SK Hynix specifically because of its US domicile and national security implications.

Apple was down approximately 9 to 10 percent on the day discussed, attributed to lower revenue guidance for September and Q3. The speakers argue Apple is underperforming because the company lacks vision, has not meaningfully innovated since the iPhone, and has failed to compete in AI, with Tim Cook having built the company around supply chain optimization rather than product innovation. Google printed its first quarter of negative free cash flow in decades, which is cited as the direct reason Google stock declined sharply, with analysts re-rating it from a software margin multiple toward a manufacturing multiple because the company is buying large amounts of hardware to produce AI tokens not yet visibly returning value to the bottom line. Amazon is described as well positioned as atoms matter more than bits, and Jeff Bezos seeded a new company called Prometheus aimed at lowering the cost of production of physical goods, targeting American parts where the US has fallen behind, modeled on Elon Musk's idiot index concept measuring the ratio of raw material cost to purchase price of a finished part.

This summary was generated from the episode transcript and can contain mistakes.