BullRush

Deep Tech & Growth Investor | Exploring small caps and asymmetric setups.
Sharing my deep dives & fundamental research

[BullRush @BullRush__]

Yesterday’s hedge fund rumors are confirmed by WSJ and Bloomberg, and the details hurt.

Per Morgan Stanley, October saw the largest active reduction of positions by stock pickers in over a decade. More than 40% of them were flat or down YTD, while the indexes were still positive. And of the 46 major funds tracked by Bloomberg, more than half finished October in the red.

The damage:

- Lone Pine: +40% in H1, now -7.5% YTD. Caught on both sides, on its AI longs and on its shorts against companies threatened by AI, which went up.
- VARA: -33% in October
- Arini: -26% YTD, forced to dump over $2B of assets in a week under margin call pressure. Rivals are now shorting its biggest positions.
- Tiger Global: -8.5% in October, back to flat YTD
- Kirkoswald: worst month in its history
- Millennium: -2.3% in October, with several internal pods reportedly shutdown
- Engineers Gate: shutting most of its human-plus-quant strategy, keeping only 5 PMs. The head of the unit is leaving.
- Samlyn: dropping everything except financials. Other sector heads and senior analysts are out.
- Walleye: head of equities and a large part of the long/short teams let go, and capital going back to clients

Thursday made it worse. Per Goldman Sachs, tech funds lost more than 4% in one session, wiping out YTD gains that had reached 24% before the summer. The Nasdaq lost $2.5T, to the point that Dan Loeb (Third Point) joked on X: “Hug your hedge fund manager.”

The problem is that everyone holds the same names: funds, retail and leveraged ETFs. Hedges stop working, and every sale drags down everyone else’s positions. $SOXL lost 39% in one session, and volatility-control funds could still liquidate up to $135B of futures.

But forced selling always stops at some point, and the rebounds that follow can be violent. On Wall Street, an old rule says to wait three sessions after a shock before deciding anything.

$QQQ $SOXX

[BullRush @BullRush__]

From what’s circulating, here are the hedge funds cracking this week, between losses, big redemptions and PMs let go:

- Axebrook
- Walleye
- Engineers Gate
- Lone Pine
- Millennium (internal pods)
- Verition (pods)
- Samlyn

Plus 4 more in the UK, apparently.

Bloomberg is starting to confirm:

- Walleye returning capital
- Engineers Gate cutting most of its equity team
- VARA down 33% in October on AI
- Kirkoswald (Greg Coffey) posts its worst month ever

Lone Pine had already lost 31% in Q3, with a portfolio packed with AI and semis: $NBIS, $ASML, $STX, $AMAT, $TER, $HUT, $WULF.

From the rumors I’m hearing, margin calls have already started. If tomorrow looks like current futures, they’ll keep coming.

source: https://t.co/6xXWc07g11

$QQQ $SOXX

[BullRush @BullRush__]

@BsVirazed @Tirelirelab @tournonskiwi @JeanFoutre22809 c’est non négociable

[BullRush @BullRush__]

Notre sentence est irrévocable avec @Tirelirelab. Nous devons relancer ce bull market.

Merci @tournonskiwi.

[Dring Dring @tournonskiwi]

@BullRush__ @Tirelirelab Le groupe vit bien 🤣

[BullRush @BullRush__]

@OnclePennybags Les acheteurs défendent bien la mm200 mais ils ont pas réussi à garder la main. Il y a encore beaucoup de travail mais c’est déjà un bon début.

[BullRush @BullRush__]

@Tirelirelab @tournonskiwi Il faut le brûler vif pour faire repartir le marché

[BullRush @BullRush__]

@Tirelirelab J’en suis aussi pour l’histoire

[BullRush @BullRush__]

@tournonskiwi Comme tu veux

[BullRush @BullRush__]

@tournonskiwi Imo, la dilution arrivera plus tard mais il faut pas l’oublier

[BullRush @BullRush__]

@Tirelirelab @Manu__C__ @tournonskiwi 🙂‍↕️

[BullRush @BullRush__]

@tournonskiwi $NBIS doit pas clôturer sous 130 ça semaine sinon c’est finito

[BullRush @BullRush__]

@SirMaximusIX Sad, I liked Jensen

[BullRush @BullRush__]

$NBIS has never sold its capacity on better terms, and the market has never doubted more its ability to deliver and fund it.

There are already plenty of earnings recaps on X, so I'd rather focus on the market's concerns, because most of them are legitimate.

The first concern is execution. Data centers run by partners are falling behind, New Jersey above all, where part of the capacity meant for Microsoft slips into 2027 after a mid-course change in the power solution. 2026 guidance is lowered and the ARR target is pushed back. On the same day, Nebius announces it will no longer guide on ARR. The reasoning holds up, but the timing inevitably raises doubts.

The second concern is capex and how it gets funded. Nebius raised its 2026 capex above the guided range, at the very moment it lowered its revenue outlook. The extra spend mostly goes toward preparing the owned sites for 2027, but the effort is only getting started: next year's investment plan is massive, and Nebius will remain dependent on capital markets at least until the end of 2027. Asked about external funding needs, management talked about flexibility without giving a number. The at-the-market program is nearly used up, a short-term bank facility will need to be refinanced next year, and with the stock this low, every equity raise is more dilutive. In a sector where CoreWeave's credit is raising alarms, the market is applying the same discount to every neocloud.

The third concern is the quality of the quarter. Part of the revenue comes from a one-off tied to a new partnership model, which inflates both growth and margin. Management itself acknowledges that margin will step back in Q4. Operating cash flow is still driven by prepayments, which are for services yet to be delivered.

Finally, the fourth concern is the customer base. The names in the spotlight are prestigious, but most of the new contracts come from AI-native companies, often funded by their own fundraising. And within the existing backlog, Microsoft and Meta carry a lot of weight.

All of this deserves close attention. But the quarter also has real strengths.

Customers are signing longer deals, at higher prices, and paying more upfront. On this quarter's contracts, prepayments cover the full capex required, which means growth is increasingly funded by customers. Management even uses this as a lever: if credit markets shut, it can favor long-term contracts with prepayments.

Owned sites are on schedule and will carry most of the capacity from 2027 onward, with better margins and tighter control over execution. Chips for 2027 are already secured.

The 2027 outlook looks conservatively built: it includes buffers for delays and excludes revenue from the new partnership model. And the platform keeps widening its scope, from inference to agents, with its first big-name enterprise customers.

I remain bullish on $NBIS. NFA

[BullRush @BullRush__]

@klutamat Yes, exactly.

[BullRush @BullRush__]

Durable rebound or dead cat bounce? The method to tell them apart.

Rally size proves nothing. The most violent rallies happen in downtrends: the S&P 500's biggest daily gains came in the middle of crises, in the 1930s, in 2008 and in March 2020. On 10/13/2008 the S&P 500 jumped 11.6%, its biggest gain since the 1930s. It then lost another 25% in five weeks. What matters is confirmation.

The benchmark is the IBD method (William O'Neil). Read it on the Nasdaq, which often signals first, and on the S&P 500, which shows whether the rally reaches beyond tech. IBD uses the Nasdaq Composite, but the Nasdaq 100 tracks the same path and gives the same count.

The stages of a real bottom:

- 1. Day 1 is the first up close after a new low. Break that low and the count resets.

- 2. The follow-through day (FTD): from Day 4 on, a gain of roughly 1% to 1.7% on volume above the prior session. It usually lands between Day 4 and Day 7, sometimes much later.

- 3. Breadth: the number of stocks rising and the share above their 50-day have to expand. A rally carried by a few megacaps stays fragile.

- 4. Leaders stick: the big names don't give back their gains, and the market weathers its first dip without returning to its lows.

Why Day 4? Because each stage has its own players.

• Capitulation: forced sellers. Multistrategy hedge funds cut exposure when risk limits trip. Vol-targeting funds sell mechanically as volatility rises. Trend-following CTAs flip short. Margin calls add to it. Nobody sells out of conviction: they have to.
• Days 1 to 3, the reflex rally. Once forced sellers are out, supply dries up. Hedge funds cover shorts, mostly on indexes and ETFs. The market rises fast, but that flow fades within a few sessions.
• The FTD, real buyers return. Mutual funds, pensions, insurers, big asset managers: they have billions to deploy and can't buy it all in one session. They accumulate over weeks, and their first wave leaves a footprint in volume. They're joined by vol-targeting funds buying back as volatility falls, CTAs if the trend turns, and corporate buybacks after earnings.
• After the FTD, the test. If funds keep buying, pullbacks come on light volume and the low holds. If they sell into strength, distribution days pile up and the FTD fails.

Signs of a dead cat: no FTD, gains without volume, then heavy-volume declines (distribution days) and a return to the low.

March 2025 is the textbook case. On the S&P 500, after a 10% drop, the only big gain (+1.8% on 3/24, Day 7) came on lower volume than the prior session: the big managers never came back. 3/28 fell 2% on rising volume, then the low broke on 4/4, right after Liberation Day.

Two real bottoms for comparison, also on the S&P 500. In April 2025, the FTD came on Day 11, 4/22 (+2.6% on rising volume), after a retest of the low. In April 2026 it came on Day 6, 4/9, and the S&P was back above its January high a week later.

But in a bear market, an FTD isn't enough. In 2022, inflation and rate hikes pulled down the whole market: -25% for the S&P 500, -36% for the Nasdaq. Several FTDs failed, four before June alone, and two rallies of +11% (March) and +17% (June-August) looked like bottoms before giving way. A macro shock that hits every stock, persistently high volatility and tightening credit multiply false signals. Across 35 corrections of 9% or more that broke the 200-day since 1970, the first FTD worked 19 times, the second 8 times, and only 8 turned into bear markets (study published in Forbes).

Today:
• 11/5, capitulation: S&P 500 -4.4%, Nasdaq 100 -6.4%, 3x average volume. JPMorgan calls it the biggest momentum unwind in ~40 years, with deleveraging back to Liberation Day levels.
• 11/6, Day 1: S&P 500 +0.3%, low held. Hedge funds turned net buyers, mostly covering index and ETF shorts (Goldman): the reflex rally.
• 11/9, Day 2: S&P 500 +2.3%, Nasdaq 100 +3.4%, one of the greenest days of the year. It's the S&P 500's 3rd-best day of 2026 and the equal-weight S&P's 2nd-best. All 11 sectors green.
• Rare breadth: 423 S&P 500 names up vs 78 down, or 84%. The share above the 50-day jumped from 61% to 69% in a single session.
• Leaders followed: $META +5.1%, $TSLA +5.5%, $AMZN +2.9%, $MSFT +2.8%, $AAPL +2.4%, $AVGO +8.8%, $AMD +13.3%. The one exception: $NVDA (-4.3%), weighed down by the internal probe involving Jensen Huang, ahead of earnings on 11/17.
• Room to buy back: fund exposure is low after the deleveraging, and with the VIX back at 20, vol-targeting funds can start re-adding exposure.
• The backdrop differs from 2022: the selloff is concentrated in AI (S&P -6% at the low, equal-weight S&P up since the peak) and high yield is stable.
• Confirmation window: Wednesday 11/11 to Monday 11/16. The question is whether the big managers come back.
• Key signal: an FTD on the Nasdaq 100 or the S&P 500. The first is the most telling, since the selloff started in AI and semis. Confirmation is stronger if both follow.
• Red line: the 11/5 low has to hold.

$spy $QQQ $SOXX

[BullRush @BullRush__]

@Tirelirelab Je pars quelques heures. Je vois que nos amies les bears se sont fait ramasser. Ils feraient mieux d’aller hiberner !

[BullRush @BullRush__]

@Gonuts0x AMD en profite bien

[Kevin Chen @Defiantclient2]

$QS: Pretty sure this means @QuantumScapeCo is already working with @ToyotaMotorCorp on this.

[BullRush @BullRush__]

@CharlesBoultel C’est un peu ça le problème, on a besoin d’avoir des prévisions plus clair sur comment ils voient les étapes vers la production . Et je pense c’est aussi pour ça qu’ils ont visé de nouveau marché avec QSE5

[BullRush @BullRush__]

$QS Osprey: the charging time is insane.

10 → 80% in ~5 minutes, with "commercially relevant" cathode loading and cell area.
QSE-5 charges in under 15 minutes.

Presented in Kyoto by Mohit Singh, Chief Development Officer and head of the Frontier R&D team, Osprey is QS's next-gen all-solid-state platform.

It is built on three components:
• The QS ceramic separator, made in continuous flow with the Cobra process
• A solid NMC cathode with a sulfide catholyte
• The process to assemble both into an all-solid-state cell

The cell is anode-free as manufactured: energy is stored as pure lithium metal.
QSE-5 uses the same base, but with a high-nickel NMC cathode and a liquid catholyte.

Why a sulfide catholyte:
• It's a single-ion conductor: only Li+ moves, hopping through the lattice
• In a liquid, anions migrate too and create polarization near the cathode, which caps charging speed
• Without polarization, ~5-min charging becomes possible, and the cathode can be much thicker
• Sulfides have good ionic conductivity, process at low temperature, and use relatively inexpensive raw materials

Why QS doesn't use it as a separator:
• Sulfide reacts with lithium metal, which hurts Coulombic efficiency and cycle life
• To work around this, the industry adds silicon to the anode, which reduces or eliminates the energy density gain
• It loses contact with active materials during cycling and must be held under high pressure
• It performs poorly at room temperature and below due to solid-solid interfacial resistance
• Its low critical current density blocks fast charging with lithium metal without dendrites

The QS separator solves these: it is stable against lithium metal, mechanically strong enough to resist dendrites, and operates at low pressure (below 3.4 atm on QSE). In commercial cell designs it has withstood over 300 °C without thermal runaway.

Osprey results announced:
• 1,000 cycles at ~80% retention, 1C/1C, 100% depth of discharge, room temperature, 1 atm, with over 5 mAh/cm² loading
• 10 → 80% in ~5 min on a power-optimized design
• Accessible capacity down to -20 °C
• Thermal stability at high temperature
• According to QS, no other sulfide-based cell had demonstrated 1,000 cycles at 1 atm

On energy:
• Cathode loading up to 10 mAh/cm² vs 3–5 for conventional Li-ion
• On the bench, these high-loading cathodes cycle comparably to lower loadings
• More capacity per layer means fewer layers, fewer separators and fewer current collectors
• Projected energy density up to ~1,100 Wh/L in large format (energy-optimized design)
• The cathode is made with a proprietary dry process: free-standing, handled in roll form, then laminated onto the current collector

The business model:
• Osprey is an open platform
• Customers use the QS sulfide cathode or bring their own, with their own materials and supply chain
• QS has already tested it with cathodes developed by its automotive customers
• QS says it is partnering with world-leading automotive OEMs (may be Honda ???), building on the investment already made to scale up sulfide cathode production👀
• QSE remains the platform being scaled toward mass production, and Osprey is QS's long-term all-solid-state solution

What QS acknowledges:
• Results come from 2-layer unit cells with the same area as QSE-5
• OEMs prefer larger prismatic cells, which will need more layers and larger layers, co-developed with each customer
• The process is demonstrated at lab scale, and higher-throughput processes are in development

[BullRush @BullRush__]

$QS unveils Osprey in Kyoto, and the headline number is ~1,100 Wh/L.

That’s the projected energy density of its new all-solid-state platform.
For context:
QSE-5, the current QS cell: 844 Wh/L

Samsung SDI, announced target for its sulfide all-solid-state cell: ~900 Wh/L
Best-in-class Li-ion cells today: around 700 Wh/L
With Osprey, QS is pulling ahead of the rest of the market, including other all-solid-state players.

The architecture keeps the QS ceramic separator and anode-free lithium-metal design, now paired with a fully solid, sulfide-based cathode.
The cell becomes fully all-solid-state.

First published results:
1,000 full cycles, ~80% retention

10 → 80% charge in 5 minutes

Only 1 atm of pressure
Operational below -20 °C
The industry has poured capital into sulfide electrolytes but keeps hitting two walls: sulfide reacting with lithium metal, and high stack pressure requirements.
QS claims the first public data of a sulfide cathode cycling 1,000 times at 1 atm.
Osprey is an open platform, compatible with multiple cathode chemistries, already tested with cathodes developed by automotive customers.
QS is positioning itself as the missing piece for sulfide all-solid-state developers: the separator that protects the lithium-metal anode.

This fits the Q3 letter: “bring all our automotive customers onto a common technology platform”, plus first Honda billings to integrate the QS platform into Honda’s roadmap.
Announcing it in Japan, the heartland of sulfide development, is no coincidence to me. That’s my hypothesis.
Still to prove: ~1,100 Wh/L is a projection, results come from early testing, and test conditions are in the white paper. QSE-5 remains the product shipping today.

[BullRush @BullRush__]

@Redzfan49 Yes, that is the problem

[BullRush @BullRush__]

@NeedleStrat Yes, this is absolutely insane with high density.

[BullRush @BullRush__]

Is the bottom in?

The Nasdaq 100 rallied 2.2% and is attempting an island reversal. The S&P 500 gained 1.3%. Semis led: $SOXX +6.0%, $AMD +7.3%, $NVDA +3.0%. The VIX dropped 10% to 17.5.

The signal is encouraging: the correction started in tech, and tech is bouncing first.

For me, it's not an FTD yet. Three things are missing:

→ Volume. It fell vs yesterday, on Veterans Day with the bond market closed (ES and MNQ futures about −22%). The big managers aren't back yet.

→ Depth within tech. Semis ripped, software followed ($IGV +1.9%), but the Magnificent 7 lagged ($MAGS +0.5%). $MSFT, $META and $AMZN closed red, even though AI megacaps were at the heart of the selloff.

→ New positions. Open interest on ES and MNQ futures has been falling since the bounce began: mostly short covering. A real bottom gets built by new buyers.

Keep the context in mind too: the most violent rallies happen in corrections. $SOXX , $NVDA and $CRWV are still in a bear market. And the indexes are approaching zones sellers will defend.

A nice bounce, but still noise.

$SPY $QQQ

[BullRush @BullRush__]

Durable rebound or dead cat bounce? The method to tell them apart.

Rally size proves nothing. The most violent rallies happen in downtrends: the S&P 500's biggest daily gains came in the middle of crises, in the 1930s, in 2008 and in March 2020. On 10/13/2008 the S&P 500 jumped 11.6%, its biggest gain since the 1930s. It then lost another 25% in five weeks. What matters is confirmation.

The benchmark is the IBD method (William O'Neil). Read it on the Nasdaq, which often signals first, and on the S&P 500, which shows whether the rally reaches beyond tech. IBD uses the Nasdaq Composite, but the Nasdaq 100 tracks the same path and gives the same count.

The stages of a real bottom:

- 1. Day 1 is the first up close after a new low. Break that low and the count resets.

- 2. The follow-through day (FTD): from Day 4 on, a gain of roughly 1% to 1.7% on volume above the prior session. It usually lands between Day 4 and Day 7, sometimes much later.

- 3. Breadth: the number of stocks rising and the share above their 50-day have to expand. A rally carried by a few megacaps stays fragile.

- 4. Leaders stick: the big names don't give back their gains, and the market weathers its first dip without returning to its lows.

Why Day 4? Because each stage has its own players.

• Capitulation: forced sellers. Multistrategy hedge funds cut exposure when risk limits trip. Vol-targeting funds sell mechanically as volatility rises. Trend-following CTAs flip short. Margin calls add to it. Nobody sells out of conviction: they have to.
• Days 1 to 3, the reflex rally. Once forced sellers are out, supply dries up. Hedge funds cover shorts, mostly on indexes and ETFs. The market rises fast, but that flow fades within a few sessions.
• The FTD, real buyers return. Mutual funds, pensions, insurers, big asset managers: they have billions to deploy and can't buy it all in one session. They accumulate over weeks, and their first wave leaves a footprint in volume. They're joined by vol-targeting funds buying back as volatility falls, CTAs if the trend turns, and corporate buybacks after earnings.
• After the FTD, the test. If funds keep buying, pullbacks come on light volume and the low holds. If they sell into strength, distribution days pile up and the FTD fails.

Signs of a dead cat: no FTD, gains without volume, then heavy-volume declines (distribution days) and a return to the low.

March 2025 is the textbook case. On the S&P 500, after a 10% drop, the only big gain (+1.8% on 3/24, Day 7) came on lower volume than the prior session: the big managers never came back. 3/28 fell 2% on rising volume, then the low broke on 4/4, right after Liberation Day.

Two real bottoms for comparison, also on the S&P 500. In April 2025, the FTD came on Day 11, 4/22 (+2.6% on rising volume), after a retest of the low. In April 2026 it came on Day 6, 4/9, and the S&P was back above its January high a week later.

But in a bear market, an FTD isn't enough. In 2022, inflation and rate hikes pulled down the whole market: -25% for the S&P 500, -36% for the Nasdaq. Several FTDs failed, four before June alone, and two rallies of +11% (March) and +17% (June-August) looked like bottoms before giving way. A macro shock that hits every stock, persistently high volatility and tightening credit multiply false signals. Across 35 corrections of 9% or more that broke the 200-day since 1970, the first FTD worked 19 times, the second 8 times, and only 8 turned into bear markets (study published in Forbes).

Today:
• 11/5, capitulation: S&P 500 -4.4%, Nasdaq 100 -6.4%, 3x average volume. JPMorgan calls it the biggest momentum unwind in ~40 years, with deleveraging back to Liberation Day levels.
• 11/6, Day 1: S&P 500 +0.3%, low held. Hedge funds turned net buyers, mostly covering index and ETF shorts (Goldman): the reflex rally.
• 11/9, Day 2: S&P 500 +2.3%, Nasdaq 100 +3.4%, one of the greenest days of the year. It's the S&P 500's 3rd-best day of 2026 and the equal-weight S&P's 2nd-best. All 11 sectors green.
• Rare breadth: 423 S&P 500 names up vs 78 down, or 84%. The share above the 50-day jumped from 61% to 69% in a single session.
• Leaders followed: $META +5.1%, $TSLA +5.5%, $AMZN +2.9%, $MSFT +2.8%, $AAPL +2.4%, $AVGO +8.8%, $AMD +13.3%. The one exception: $NVDA (-4.3%), weighed down by the internal probe involving Jensen Huang, ahead of earnings on 11/17.
• Room to buy back: fund exposure is low after the deleveraging, and with the VIX back at 20, vol-targeting funds can start re-adding exposure.
• The backdrop differs from 2022: the selloff is concentrated in AI (S&P -6% at the low, equal-weight S&P up since the peak) and high yield is stable.
• Confirmation window: Wednesday 11/11 to Monday 11/16. The question is whether the big managers come back.
• Key signal: an FTD on the Nasdaq 100 or the S&P 500. The first is the most telling, since the selloff started in AI and semis. Confirmation is stronger if both follow.
• Red line: the 11/5 low has to hold.

$spy $QQQ $SOXX