Odd Lots · · 1h 0m

The Perfect Storm That Clobbered Hedge Funds

With Michael Ball, Nishant Kumar

Summary

Nishant Kumar explains how October's sell-off hit multi-strat hedge funds because crowded trades in AI stocks, French bonds, high-yield credit and the Brazil election all moved against them at once, with Millennium, Balyasny, ExodusPoint, Vara, Vail Rock and Kirkoswald posting heavy losses. He argues that the top multi-strats' roughly $2.2 trillion of gross exposure and their loss-limit systems forced correlated de-risking that deepened the drawdown, and he doubts a new multi-strat can attract and retain talent despite strong investor demand for capacity. Michael Ball attributes the broader equity break to options gamma, dealer delta-hedging, leveraged ETFs and systematic selling that pushed the S&P 500 through its 7,600-7,700 range, and says the AI capex narrative flipped as financing, free cash flow, local opposition and power constraints came to the fore. He is constructive on Nvidia and Iren, sees neoclouds as oversold with washed-out positioning, flags Anthropic's roughly $80 billion IPO as a supply event, and expects energy-driven inflation and an oil surplus to shape rate expectations.

Key takeaways

  • Nishant Kumar on Asset managers, brokers & exchanges: Hedge funds' crowded positions in AI and other trades made October a synchronized loss across strategies.
  • Nishant Kumar on Asset managers, brokers & exchanges: Hedge funds face a very difficult next two months after a bleeding November and PMs sidelined.
  • Michael Ball (bullish on IREN): Iren's customer prepayments are strong, improving its free cash flow and likely reducing its borrowing needs.
  • Michael Ball (bullish on NVIDIA): Nvidia remains the AI trade's cornerstone, and a CEO change would not undermine that.
  • Michael Ball (neutral on Anthropic): The Anthropic IPO's roughly $80B raise can be absorbed by the market if the story is compelling.

What was said, by company and theme

21 views from 2 speakers. Every quote is checked against the episode's transcript (how).

IREN

Across the shows: 2 speakers in other episodes, 1 bullish and 1 bearish calls, past 30 days

  • Michael Ball · BloombergBullish

    Iren's customer prepayments are strong, improving its free cash flow and likely reducing its borrowing needs.

    Iren reported last week that customer prepayments were quite good. Ball sees this as improving the company's free cash flow relative to the rest of the neocloud group, and expects its borrowing needs may fall as customers pay upfront. He cites this as a sign of differentiation within neoclouds, where others such as CoreWeave carry more leverage.

    “I wrote about them yesterday because Iren came out last week and said basically the prepayments from customers was quite good. So the free cash flow is looking better for that particular name, and their borrowing needs may need to go down.”at 42:39

NVIDIA

Across the shows: 100 speakers in other episodes, 81 bullish and 54 bearish calls, past 30 days

  • Michael Ball · BloombergBullish

    Nvidia remains the AI trade's cornerstone, and a CEO change would not undermine that.

    Ball accepts that Nvidia is central to AI financing, but pushes back on the idea that a CEO change would alter that. He says a leadership change could affect the growth path of chips and M&A, but he sees nothing that makes Nvidia look shaky. He adds that the AI trade is now bigger than any single company, and that recent moves were mechanical short covering rather than a shift in the AI narrative.

    “I get that Nvidia's seen as central for the financing video, but I also push back that's going to change if the CEO changes.”at 47:32

Anthropic

Across the shows: 111 speakers in other episodes, 152 bullish and 90 bearish calls, past 30 days

  • Michael Ball · BloombergNeutral

    The Anthropic IPO's roughly $80B raise can be absorbed by the market if the story is compelling.

    Ball expects the new equity supply to crowd out other AI stocks as investors rebalance. He argues the market is large enough to absorb about $80 billion if the story is compelling, which he thinks it is. He adds that the lower valuation (about $1.4T, down from around $2T) reduces the risk, so the IPO is an event to get through before the market moves on.

    “these are big numbers in theory, but the market's pretty big, and it can absorb this as long as the story behind it is compelling, which I think it is. And valuations have come off now.”at 49:29

Alphabet

Across the shows: 66 speakers in other episodes, 61 bullish and 45 bearish calls, past 30 days

  • Michael Ball · BloombergBearish

    AI chatbots like Claude are eroding Google's search business, so its search weakness makes sense.

    Ball says Google had issues with search, which he thinks makes sense given the shift in behaviour. He uses himself as the example: he now uses Claude more than Google, which he admits hurts token usage. He expects this kind of shift to weigh on traditional search businesses while people adjust to the new technology.

    “I'm using Claude more than I'm using Google, probably to the detriment of token usage, and I shouldn't be. But we're all learning how to kind of adjust to this new tech. That's probably gonna be a little bit of a jitters for traditional businesses like search.”at 51:16

AI infrastructure & data centers

  • Michael Ball · Bloomberg

    Blue-wave AI regulation would curb hyperscaler capex, the engine of their earnings growth.

    On Wednesday after the midterms, hyperscalers rallied because the market thought the blue wave and AI-related ballot measures would force them to pull back on capex. That would have meant better free cash flow and stronger core businesses. By Thursday that logic reversed. Capex had been the driver of their earnings growth, so curbing it became a negative. Tangential benefits no longer offset the hit to the main growth engine, and the stock moves on Thursday reflected that repricing.

    “from this blue wave, and we saw all these things on the ballot for the AI initiatives, would force basically them to pull back on CapEx. It would be better free cash flow, and hence their core businesses would do better.”at 34:56
  • Michael Ball · Bloomberg

    October marked a shift: AI capex stopped being a growth engine and became a financing, cash-flow and power-constraint story.

    Ball argues that in October the AI capex narrative flipped. Capex was no longer seen as a growth engine but as a problem of financing, free cash flow, local opposition (NIMBYism) and electricity constraints in actually executing buildouts. He links this to bond markets: yields reversed, the 10-year and the curve flattened, and investors grew more worried about growth because they expected AI capex to fall next year, removing the main growth driver.

    “October really saw that shift in the CapEx narrative, where all of a sudden it wasn't a growth engine anymore, and it was about financing, about free cash flow, and again, about this nimbyism and electricity problems”at 39:21
  • Michael Ball · Bloomberg

    Neocloud stocks are overdone after a 70% drawdown, and positioning has now been cleaned up.

    The neocloud index is down about 70% and sentiment is very negative. Ball argues the sell-off went too far relative to the fundamentals he sees in backlogs, prepayments and earnings. He notes that positioning is now washed out, so the reversal risk is real, and that the story looks overdone at these drawdown levels.

    “But overall, like, I still feel like the story was overdone when you see these 70% drawdowns. ...the positioning's cleaned up.”at 44:55
  • Michael Ball · Bloomberg

    After the blue wave, risk is in new data center projects hitting 2028 earnings, not projects already under construction.

    Ball says the 2028 outlook changed after the midterms because projects already under way are unlikely to be cancelled, citing Texas as an example. He notes that other states such as Ohio, Georgia, New Mexico and Arizona may vary. The real risk is new projects that have not yet broken ground, which would hit 2028 earnings and therefore the neoclouds whose costs depend on those projects.

    “the whole story on '28 after the blue wave changed because projects in kind of under already being done aren't really at risk of being canceled in Texas... The worry is new projects, and that means 28 earnings.”at 45:40
  • Michael Ball · Bloomberg

    Chips are being upgraded faster than neocloud projects can be built, creating operational risk.

    Ball says a neocloud was hit after Microsoft asked it to upgrade its chips, which set the project back. Because chips are now moving faster than the projects can be completed, he argues the risk is operational rather than purely narrative.

    “Because now the chips are moving faster than the projects can get done.”at 46:35
  • Michael Ball · Bloomberg

    Compute demand is still growing impressively, so the cloud business remains the hyperscalers' cornerstone.

    Ball argues the cloud business and its monetization are the cornerstone of the hyperscalers' story. The key metric investors should watch is compute demand, which he says is still growing impressively. Debates over token prices, financing, competition and NIMBYism distract from that, and as long as compute keeps growing, the companies can work through those headwinds.

    “the cloud business and the growth and the monetization of the cloud business is the cornerstone... the number one... compute pie has to keep growing, and it's actually growing quite impressively still.”at 51:16

Asset managers, brokers & exchanges

  • Nishant Kumar · Bloomberg News

    Hedge funds' crowded positions in AI and other trades made October a synchronized loss across strategies.

    October was a perfect storm because strategies that usually diversify all lost money together. Millennium fell 2.3% in a month when it normally annualizes around 14%, and Balyasny and ExodusPoint also lost. Vara fell 33% despite being up for the year, and Vail Rock was down 12%. Kirkoswald, a macro fund, lost 6.4%, its worst month since it started trading. Pivotal Path found the 10 largest multi-strats control about $2.2 trillion including borrowed money, with roughly $300-350bn of capital backing the trades, so crowding plus leverage amplified the damage. Goldman's work shows correlations between multi-strat funds stayed around 0.2-0.3, which points to crowding in the same names rather than funds simply moving together. Nishant argues the industry was effectively long AI through stocks, credit exposure and possibly rates. Crowded trades in AI stocks, French bonds, the high-yield selloff and the Brazil election all hit at once.

    “But this was like a perfect storm. Everyone failed together.”at 6:31
  • Nishant Kumar · Bloomberg News

    Multi-strat stop-loss rules force synchronized deleveraging, which amplifies drawdowns across the platform.

    Multi-strats allocate capital to PMs and set loss limits. A PM who loses 5% of allocated capital often has it halved, and at 7.5% total loss is closed out and can be fired immediately. This ruthless system keeps any one PM from endangering the whole platform. The same rules mean that when many pods lose money on correlated trades, they all cut risk at once. That forced selling deepens the drawdown, so the risk machine that protects the firm also magnifies the market moves.

    “if you have 100 to manage and if you lose 5% of that, so $5 at many of the firms, your capital will be cut into half. If you lose another 2.5%, so total loss becoming 7.5, you are closed out.”at 19:50
  • Nishant Kumar · Bloomberg News

    Hedge funds face a very difficult next two months after a bleeding November and PMs sidelined.

    About ten days into November, many portfolios are still bleeding. A rebound this week looks like mean reversion and dip-buying, but it is not enough to repair the losses. Anecdotally, dozens of PMs who took big equity, macro and credit risk have been fired, poached, or are sitting out their non-competes. With about 40 trading days left in the year, the question is whether the remaining managers have the nerve to take risk. If they do and are wrong, weak performance can cost them their jobs anyway, so the next two months look very hard.

    “So next two months is going to be very, very, very difficult.”at 25:46
  • Nishant Kumar · Bloomberg News

    Demand for new multi-strats is strong, but few new launches succeed amid talent competition.

    Investor demand for multi-strats is intense. Millennium targeted about $20bn of new capital this year, got demand above $30bn, and took in $22bn. Top players are close to new money, and Point72, Citadel and D.E. Shaw have returned roughly $35-40bn to investors, which needs a new home. So capacity is a real opportunity for a manager who can build a multi-strat with decent capacity and reasonable returns. But many attempts have failed. Nishant doubts a new entrant can hire and retain top talent and poach other traders. Jain Global, Walleye and Isler have already cut back or shut down, so he is skeptical a new multi-strat can get off the ground.

    “If someone can genuinely build a multistrat and create decent capacity with reasonable returns, there is demand for another multistrat.”at 29:32

Market sentiment, positioning & volatility

  • Michael Ball · Bloomberg

    Thursday's single-stock selling was mostly retail, adding to hedge fund degrossing.

    Thursday saw the biggest selling of single stocks, especially in growth and tech. With no major macro catalyst, holders dumped these names after people had been rebuilding positions on Wednesday. Michael reads this as mostly retail, since the hedge fund community was already degrossing and de-netting. Holders of both retail and institutional money were involved, so the selling was broad, but retail made up most of the single-stock flow.

    “it was the biggest selling of single stock. So again, that is gonna be mostly retail from the way we viewed it because we know there's been a decent amount of degrossing, de netting from the hedge fund community.”at 33:48
  • Michael Ball · Bloomberg

    Thursday's selloff was self-feeding as dealers short puts and vol-control funds sold into broken levels.

    The S&P 500 had held a 7,600-7,700 range for months, cushioning weakness in AI names through a rotation into equal weight and value. Once those levels broke, the selling fed on itself. Dealers were short puts, so as the market fell they had to sell to delta-hedge, which accelerated the move. Leveraged ETFs, which have negative convexity, added to the pressure. Vol-control and systematic funds had to degross as volatility rose, and CTAs sold trends. The Nasdaq was already in correction, and the combined mechanics pushed the market through the ranges to new lows. A July clearing event in legacy AI auto-callables and an August rebuild had already suppressed realized volatility, which made the later break more violent.

    “The dealer is short that put. So to Delta hedge that as the market goes down, they are selling into the market, which exacerbates that move, which is exactly what happened Thursday.”at 37:13
  • Michael Ball · Bloomberg

    Investors are almost flat to short AI, leaving room for new buyers to step in.

    Ball says the AI story needs to stabilise, and that positioning needs to rebuild first. Most investors are close to flat or short, and options positioning is costly to hold if volatility collapses or skew shifts. The AI story may need months to stabilise because of regulatory and NIMBY worries, but with positioning so light there is ample room for new buyers.

    “the actual positioning needs to stabilize. We need to see some rebuilding there because quite frankly, everyone's almost flat to short. And there's a lot of room now for the kind of new guys to step in.”at 43:58

US stock market outlook

  • Michael Ball · Bloomberg

    Investors rotated into healthcare, quality and staples, holding up the index as tech was sold.

    Ball describes a bifurcated market in which investors moved into healthcare, quality and some staples, which held the index up while tech was sold and de-grossed. He is describing how positioning and flows were working through the market, not forecasting a particular outcome.

    “everyone was comfortable going into health care and quality and to sort of even staples and then basically hiding out in there, which was holding up the index while just selling and degrossing in sort of tech.”at 41:09
  • Michael Ball · Bloomberg

    Broader equity supply has stopped shrinking, and hyperscaler buybacks for 2027 are now in doubt.

    Ball says 2026 is among the first years in which net equity supply has not shrunk. Buybacks had been driven mainly by hyperscalers, but their free cash flow has fallen enough that the buyback support is in question for 2027. Once employee compensation is included, the picture is weaker still. He contrasts this with the 2024 to 2025 period, when shrinking supply and passive flows supported the market.

    “this year has been one of the first years where actual broader equity supply is no longer shrinking. Buybacks, as we know, were primarily being driven by hyperscalers, which free cash flow now has shrunk enough that that buyback pulse, at least for '27, is totally under question.”at 48:47

Oil & fuels

  • Michael Ball · Bloomberg

    Oil's backdrop is positive if the geopolitical risk premium fades, with supply surplus and falling cracks.

    Ball says oil's outlook is positive if the overhang of geopolitical risk premium linked to Iran and the Gulf ends. Barrels are moving out of storage, JPMorgan raised its 2026 surplus forecast to 1.5 million barrels a day, and product cracks and European gas prices have fallen. He expects a bumpy path because the market is oversold, with inventories still to rebuild and Russia and Ukraine still risks.

    “But as far as oil, the backdrop is actually positive if we can end this kind of overhang of geopolitical risk premium.”at 52:48

US inflation

  • Michael Ball · Bloomberg

    Energy prices have started feeding into core inflation, changing the Fed's calculation.

    Ball points to the CPI report released that day. He argues the energy price pulse had been materially starting to come through into core inflation, which changes the Fed's calculation. This is his reason for thinking energy developments matter for rate expectations and not just for the headline number.

    “So the energy pulse, as we just saw in CPI today, had been materially starting to come into core, changing our Fed calculation.”at 52:57

Across the shows

What other episodes Chorus covered said about the same companies in the 30 days up to this one.

  • Anthropic111 speakers · 152 bullish, 90 bearish calls
  • NVIDIA100 speakers · 81 bullish, 54 bearish calls
  • Alphabet66 speakers · 61 bullish, 45 bearish calls
  • IREN2 speakers · 1 bullish, 1 bearish calls

Topics

  • AI chip leadership risk
  • AI positioning
  • Anthropic IPO
  • compute demand growth
  • data center buildout pace
  • data center capex
  • data center siting & permits
  • energy prices feeding core inflation
  • equity supply & buybacks
  • hedge fund crowding
  • hedge fund risk-taking outlook
  • hyperscaler capex and AI regulation
  • multi-strat capacity and launches
  • multi-strat risk limits
  • neocloud financing
  • neocloud positioning
  • oil supply surplus & geopolitical premium
  • options gamma and dealer hedging
  • retail selling of growth and tech stocks
  • search competition from AI chatbots
  • sector rotation

More from Odd Lots

All Odd Lots episodes

Ask your AI about this episode

Chorus gives Claude, ChatGPT and other assistants every view from this episode and every other show Chorus follows. Ask who agrees, who changed their mind, and what else was said about the same companies.

Setup guide

A weekly email with the calls that mattered from the most important voices in investing. Unsubscribe any time. Privacy Policy

Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus