GS Privorotsky De Grossing; Tech, Momentum Interesting. Fed

Asia remains under pressure. Korea was down double digits at one point before recovering. SK Hynix delivered record absolute numbers that nonetheless missed lofty consensus expectations, sending the stock down more than 10% and accelerating the ongoing deleveraging. The bar had simply become too high. It’s been remarkably tough sledding for hedge funds, multi-strats and quants over the past few sessions. Everyone owned some version of the AI momentum trade and many were slow to react to the rollover in AI winners and hardware. The last few days have seen multiple-sigma moves across long/short books without much movement in the broader indices. This is much more a deleveraging and de-grossing event than a macro one.

Prime: Positioning continues to unwind aggressively. Prime estimates Fundamental L/S down 1.3%, Systematic L/S down 1.0% and Multi-Strats down 1.7% yesterday (have to go back to Covid to find events like this), with momentum the overwhelming driver across every strategy and Asia, particularly Korea, accounting for much of the pain. After adding a record amount of gross leverage through the first five months of the year, hedge funds have now unwound roughly 45% of that increase since the end of May. At the same time, global technology has seen one of the largest waves of long selling in the past decade… the biggest since January 2021 on a two day basis. Japan, Korea and Taiwan have all been heavy sources of selling, with both Korea and Taiwan now net sold year-to-date. The positioning reset is well underway.

So what’s changed? People point to concerns around supply and CXMT’s listing, but I think the bigger shift is that we’ve gone from effectively a two-player frontier market to one with five or six credible competitors. That doesn’t invalidate scaling laws, but it does raise questions about the return on ever-larger training runs versus post-training and reinforcement learning (still bullish for inference). More importantly, I don’t think investors are questioning the hyperscalers’ willingness to spend. I expect that to be reinforced again by earnings this week. The question is capacity and returns on capital. That’s where the repricing has been most acute… debt-funded expansion, leverage and the highest-multiple parts of the ecosystem. Once sustainability comes into question, multiples compress (not just spenders but the receivers as well).

Momentum/Tech: Last few sessions have been an extremely aggressive deleveraging event. We’re approaching levels that are becoming genuinely interesting. If you have duration capital, whether in momentum, memory or hardware, I think opportunities are emerging. My preference remains semiconductor capex, where competitive moats are more durable. I'm less interested in memory and I think there should be differentiation but even there tactically things feel extreme.

Oil pushed higher overnight after Iran launched another surprise attack on U.S. bases. Mid $80s increasingly feels like the new equilibrium until the conflict is resolved. The equity market still shows remarkably little anxiety. The pattern remains the same… the administration talks tough but becomes more conciliatory whenever markets begin to price in economic pain. In the low $80s they’re willing to prosecute the conflict; in the high $90s they’re looking for negotiations. The broader strategic picture remains unsustainable, but the Strait will likely remain contested and the practical outcome is simply slower shipping rather than closure.

Fed: Key event in many ways, has prevented investors from adding risk. We simply need to get through it. The objective case for a July hike isn’t particularly compelling. Officials have kept the door open, but recent data hasn’t justified moving early. We’ve had softer labor data, softer inflation and while inflation remains above target, there isn’t an obvious need to rush. The argument for hiking is really about credibility… getting ahead of the curve and reinforcing the Fed’s inflation-fighting credentials. There is a scenario where they surprise with a hike but soften the surprise with an emphasis of a meeting by meeting rather than the start of an aggressive tightening cycle.…