
Long / short trading with objective of beating $SPY in good times and protecting capital in bad. NOT financial advice. Subscribe for free at https://nitter.gq/t.co/ojgKxYXKt6
California, USA
Joined September 2023
- Tweets9.2K
- Following245
- Followers3.7K
- Likes7.4K
Minerva Capital retweeted
Dan Niles @DanielTNiles on what Jensen’s exit means for the AI trade
“He’s now gone. So that’s awesome for OpenAI and Anthropic and their ability to get regulatory capture. But it’s horrible for the neoclouds. Bank of NVIDIA is going to be closed for a bit as Colette, who I believe is an excellent CFO, she's going to more closely examine all those deals and the circular financing.”
Chapters
00:00 Is the AI selloff really unprecedented?
00:50 The dot-com comparison and three phases of AI
02:22 Falling token costs and the open-source effect
03:19 Jensen’s exit and the “Bank of Nvidia”
04:15 Can prepayments protect the neoclouds?
05:31 SoftBank’s writedown: warning sign or exception?
07:20 Is AI demand different this time?
09:16 Interest rates, financing and the AI trade
11:36 Anthropic’s IPO and the risk of an AI duopoly
12:54 Slowing growth and rising losses at the AI labs
13:25 Why Dan prefers Anthropic to OpenAI
Minerva Capital retweeted
Datasection (3905.T) just cut FY revenue guidance by 89%, from ¥162B to ¥17B, and the stock went limit-down.
This is the "Japanese neocloud" that the FT reported is effectively a Tencent datacenter: Chinese offtake, Nvidia Blackwell GPUs, Japanese flag. (1/8)🧵
Minerva Capital retweeted
Do NOT Trade the Anthropic IPO Without These 3 Rules youtu.be/wwDzEykjLcQ?si=dkPy… via @YouTube
Minerva Capital retweeted
Replying to @MinervaCap @TomorrowXSummit
here you go buddy youtube.com/watch?v=6r3D-CMJ…
Minerva Capital retweeted
Here's some @TomorrowXSummit Day 1 takes for anyone that cares - The variety of conversations across public pensions, credit, long/short funds, power and compute markets has been useful for seeing how differently people underwrite the AI buildout.
• The strongest private skeptical argument was token output improving faster than new uses can absorb it. Rental prices and GPU residual values become difficult to underwrite without another wave of applications.
• Credit investors focused on operator quality and guarantees. Power investors liked specific contracted projects. Conviction in individual financing structures was stronger than conviction in returns across the whole buildout.
• Compute traders saw older hardware becoming more interchangeable. Neocloud participants pushed back on memory, precision and networking requirements. Those differences complicate trading and hedging compute.
• Public-equity investors saw industrial companies benefiting from AI cost savings while still grappling with expectations for select stocks as growth outlook is more mixed.
Main Stage takes - mostly hosted by Gavin and @AntonioGracias
@GavinSBaker / Atreides: His roughly 60-year railroad analogy depends on demand remaining economically useful. Higher rates and power/wafer shortages could restrain overbuilding and extend the investment cycle.
@altcap / Altimeter: Cited Anthropic approaching a $100B revenue run rate with roughly 5,000 employees, versus Google’s 105,000 at that run rate. Estimated inference used roughly 500 MW of its 2 GW. His argument was that revenue can grow rapidly without proportional growth in inference requirements.
@DavidSolomon / Goldman Sachs: Administrative AI is scaling. Goldman expanded from six operating processes under redesign to eight, including onboarding and anti-money-laundering checks. Savings can fund additional client coverage.
@LisaSu / AMD: Bullwhip risk as shortage-driven orders become excess inventory. Different workloads need CPUs, GPUs and custom silicon. Her example of an engineer using Claude to improve an AMD implementation showed how software optimization can make alternative hardware more competitive.
Stan Druckenmiller / Duquesne: “I’m not playing.” Short copper, roughly flat net equities. Infrastructure looks more like a tactical trade after its prior run. His longer-term interest is AI use cases, including labs moving from selling tokens to selling completed tasks.
@MichaelDell / Dell Technologies: No change to his three-year buildout view. Roughly 6,500 AI Factory customers provide demand beyond frontier labs. His September update included 3,300 added in three quarters. Private deployments make API revenue an incomplete adoption measure.
@DinaPowellMcC / Meta: Local tax receipts are the political argument. Her Louisiana example was teacher bonuses rising from $10,000 to over $50,000 through data-center tax revenue. The case for community support gets stronger when residents can see the buildout in their own paychecks.
Jitendra Mohan / Astera Labs: Customers asking for more compute over the preceding two weeks, with Baker later referencing requests for earlier deliveries. Expansion into switches, memory connectivity and optics follows copper’s physical limits.
@rauchg / Vercel: $700M annualized revenue, up 180% year over year. Ten years to its first billion deployments, now 1.25B every four months. Agents expand both the builder population and output per builder.
@lqiao / Fireworks: Still seeking compute. Autonomous workflows remove the human pause between prompts = explosion in future usage. Running software on customer and partner capacity could let growth exceed Fireworks’ ability to finance hardware.
Bret Johnsen / SpaceX: Put early terrestrial builds in the low-to-mid $30s/W versus $50–60/W for others, in his comparison. Power generation, vertical integration and faster construction underpin the advantage. Aims to be among the largest AI infrastructure providers in 2027. His orbital argument is that competitors face rising power, labor and construction costs while reusable launch drives SpaceX down its cost curve. As chips get cheaper, those non-chip costs matter more. Targets meaningful orbital gigawatts in 2029.
Marc Rowan / Apollo: His dollar of credit needs to come back with its contractual return. Equity investors need future compute earnings to make their dollar worth more. He questioned whether investors throughout the stack are being paid adequately for the risks they’re taking. Discussed amortizing chips to zero over five years and financing against cash flows and counterparties he can underwrite today. Disappointing AI revenue can strand assets and impair returns while the hyperscalers themselves remain solvent.
@alighodsi / Databricks: Agents underappreciated, recursive self-improvement overhyped. Adoption slows as industries become more regulated, leaving existing capability underused. Even inside Databricks, he had told employees the prior week to ask the model when it knew more than they did. The opportunity is getting organizations to use intelligence already available and making it less complicated.
@tbpn interviews w/ @jordihays @johncoogan
@DavidSolomon: Thousands to tens of thousands of AI-assisted vulnerability scans daily for seven or eight months. Faster patching also creates operational risk.
@GavinSBaker: A handful of dominant labs could recreate concentrated cloud-buyer economics for semiconductor suppliers. Open models broaden the customer base.
@andrewdfeldman / Cerebras: Uses 5 nm and avoids HBM to compete with Nvidia for fewer scarce inputs. Sees opportunities for 50–200 MW inference sites and expects roughly 75% of Cerebras’s 2027 data centers outside the U.S.
@ChaseLochmiller / Crusoe: Power availability determines where compute goes. Crusoe Spark puts prefabricated compute beside that power. Factory construction reduces dependence on scarce local construction crews, and Crusoe makes its own electrical switchgear. The strategy is to control more of what determines how quickly available power becomes operating compute.
Minerva Capital retweeted
It is great that we are finally getting oversold in short term breadth.
But do not forget that the worst declines come from oversold.
Bulls have a chance to squeeze, monitor closely if they fumble it, that’s information.
Also take note of intermediate term and longterm breadth not that bad yet, while short term breadth is bad.
This IMO favors a dead cat to improve short term breadth to allow the longer term breadth to bleed lower before the real turn comes.
As always open minded, but vigilant.
Minerva Capital retweeted
The GOAT, @standuquesne, at @TomorrowXSummit:
• Thanked @GavinSBaker for identifying the AI opportunity early.
• Nine months ago, he was long copper and equities, short bonds, and thought healthcare would break out. Today, he owns almost none of those positions.
• At the time, he had conviction: the Fed funds rate was around 3.5%, monetary policy was loose, and commodities were hitting new highs. A recession within 6–12 months seemed unlikely.
• Claude Code had people excited about accelerating AI progress, adding to his bullish view on equities and the economy.
• Then war started. AI capex remains strong, but much of the broader picture has flipped.
• He sees a change in the Fed regime, with the Fed hiking and Warsh unhappy with what has been happening.
• Commodities have broken out, but they are now short copper. Credit spreads are widening.
• They are virtually flat equities: not long, but reluctant to be short with some things oversold and so much innovation underway.
• After 50 years of investing, he has less clarity on where we’ll be in 12 months than ever before.
• They have some small positions in the dollar and are largely staying on the sidelines.
• Bonds and currencies tend to make their biggest moves when equities are falling.
• They are in capital preservation mode. The priority is to avoid losing money.
• After a great three or four year run, they’re taking a break. It could last two days or six months.
• Their risk management is unusual: they don’t use VaR. In periods of chaos, the usual correlations across asset classes can break down.
• He believes in streaks. A big part of the job is recognising when you’re seeing the ball clearly and when you’re not.
• Given how commodities are behaving, he thinks the second half of 2027 could be disappointing.
• He’s looking for AI use cases and doesn’t think infrastructure will be the main engine forever. AI labs could be among the best investments if they begin selling completed tasks rather than tokens.
• He worries about the deterioration of US K–12 education. The US has worked around this by importing talent, but immigration policy could make that harder.
• He also worries about US fiscal challenges and sees the possibility of a crisis forcing major fiscal reform within the next decade.
• Investing edge isn’t simply about IQ. Intellectual humility matters: smart people can become attached to a thesis and be too slow to adjust.
• Put your eggs in one basket and watch the basket very carefully.
• What separates good investors from great ones? Curiosity, passion, discipline, an extreme competitive drive and, often, little work-life balance.
• Every buyer has a seller. Understand who is on the other side of your trade.
• He’ll make a meaningful initial bet with perhaps 15% of the information others might want, then increase the position as his conviction grows.
• His longer-term predictions: a 20% chance of fusion energy, a 70% chance of data centres in space, and cancer becoming a managed disease within ten years.
Minerva Capital retweeted
*DRUCKENMILLER: I DON'T PERSONALLY BELIEVE THE LABS ARE GONNA BE IN THE TOKEN BUSINESS. THEY CAN START SELLING TASKS INSTEAD OF TOKENS
*DRUCKENMILLER: IN EQUITIES WE'RE LOOKING FOR AI USE CASES AS OPPOSED TO INFRASTRUCTURE
*DRUCKENMILLER: I'M REALLY ALERT TO THE FACT THAT THE SECOND HALF OF ‘27 COULD BE ECONOMICALLY DISAPPOINTING
*DRUCKENMILLER: I'M JUST TAKING A BREAK, WHICH COULD LAST TWO DAYS OR SIX MONTHS, I DON'T KNOW
*DRUCKENMILLER: WE'RE NOW SHORT COPPER. THE FED IS NOT CUTTING, THEY'RE HIKING
*DRUCKENMILLER: WE'RE NOT LONG EQUITIES. WE'RE NOT SHORT EQUITIES. I WOULDN'T DARE GO SHORT
*DRUCKENMILLER: I CAN'T REMEMBER A PERIOD WHERE I HAD LESS CLARITY ON WHERE WE'RE GONNA BE IN 12 MONTHS
*DRUCKENMILLER: WE'RE IN CAPITAL PRESERVATION, TRYING TO NOT LOSE MONEY AS OPPOSED TO MAKE MONEY
Minerva Capital retweeted
Staples and Apple trading at 45x PE while all else falls apart is NOT a sign of a healthy market
These stocks are the last resort for funds that want/need to stay long and have rotated out of Tech
Historically this has often not ended well for those stocks (nor the market)
Minerva Capital retweeted
Got so much crap over this post.
I wasn’t crazy, things were cracking, the market is weakening and the rotation is still not healthy
I am pretty tired of all the "breadth is improving bros" and I want to show you what I am seeing.
As always things can change, mend, news, tacos, FOMC can improve things, this is a day by day assessment.
BUT everyone saying this is normal rotation and things are improving are smoking something or actively denying what is happening under the hood.
Lets have a look at a couple of things, this is a long one.🧵
1. Lets start with the simple stuff, I think both the 2024 top and the 2015-2016 market is similar to now, lets look at the pre-tariff decline.
Same pattern, roughly the same chop time. We see trend fatigue as the 50 SMA no longer gives bounces. And if the 50 SMA can't attract buyers, you need buyers lower down -- a first cautionary sign.
We also see the low volume when we top, always a bad thing to see -- there is no committed buyers at higher prices.
2. Small caps losing and living under the 200 SMA is a huge signal, everyone is taking it way too lightly. Only one time since the 2022 bear market has seen small caps living under the 200 SMA.
And just like now small caps topped before the market and was down about 10% when the decline in the S&P 500 started.
Minerva Capital retweeted
Replying to @KobeissiLetter
Leaving this here for the folks that want to understand why that data isn’t representative of the street’s real positioning.
Everyone who keeps sharing these short interest charts is missing the flows behind them. Most shops are not taking outright short bets. This is another effort to ease margin requirements across the street: banks providing balance sheet for more speculative bets are asking clients to post more short exposure.
The sequence on the street looks like this:
HF manager: “I want more margin.”
Bank: “Things look risky in public AI bets, and we’re already lending a lot against private AI bets. For some clients, we’ll extend more financing, but only if they add more short exposure to improve their collateral scores.”
HF manager: “Okay, I’ll short more stock against my main bet.”
That is how you end up with short exposures that look astronomically high while many end users remain heavily long tech. The additional shorts are supporting the financing of their main bets.
Which also tells the bigger story on why raw equity financing rates haven’t blown out the way they did ahead of the big SpaceX IPO in June, despite the fact that lending is slowing.
Minerva Capital retweeted
SemiAnalysis cut its 2028 AI accelerator spend forecast by 28%, mostly from AI chip price deflation, not volume collapse. NVIDIA GPUs take most of the hit, while Broadcom TPU/ASIC revenue stays the most resilient.
404kresearch.substack.com/p/…
Minerva Capital retweeted
ODE TO ANTHROPIC: It’s the year of the IPO, but the performance of 2026’s class of IPOs have been sluggish.
People get a lot of FOMO around IPOs in an attempt to buy around listing day. But a few things to consider about listings generally:
- There are buying opportunities that often arise soon after. With earnings season around the corner, lockups expiring, macro uncertainties remaining and questions around the direction of the AI trade, buying opportunities will likely emerge that are nearly as or more attractive.
- The private markets are becoming ever more robust. The ability to get exposure to companies like this has been rising, it’s not perfect yet, but with more abilities to access earlier stage companies, investors are realizing that they should be getting in way earlier in the themes and names they love. Institutions don’t need to be the only ones who benefit from an IPO.
2018 vs 2026 after the midterms - Similarities
• New Fed Chair: Powell was eight months into his tenure in 2018. Warsh is now in a similar position.
• December FOMC: In 2018, the Fed hiked into a falling market.
• Divided government, same president: The post-midterm rally in 2018 lasted just one session (So far still true).
• Crowded trades unwinding: FAANG then, AI now?
• Oil falling alongside equities: Oil is already well below its October highs. A US-Iran agreement could accelerate the decline.
Differences
• The Fed’s room to maneuver: In 2018, inflation was around 2%, giving the Fed room to pivot. Not this time?
• Returns on capital: Today, the market is questioning whether unprecedented AI capital expenditure can generate adequate returns. The election outcome introduces another risk: greater political and regulatory scrutiny of AI infrastructure, particularly data centers, their energy requirements and financing.
- Did I miss anything?
Minerva Capital retweeted
With Anthropic's IPO likely later this week, good time to remember that huge pops are normal on Day 1, but so are enormous drawdowns during that first year.
$ANT
Been bearish crypto for a while personally. I mean if you follow price you’re led to the same conclusion naturally that it’s in a protracted bear. The recent rally got very close to convincing but ultimately failed exactly where should have for continuation bear.
But more worryingly I would say for bag holders, the long term fundamental thesis (to the extent that fundamentals have any meaning in relation to crypto) has materially worsened. Mass inflows have come and gone, the reality is outside of some store of value which has some unique properties, it’s not a technology that’s driven incredible productivity growth like AI has. Decentralisation thesis turned out to be a fugazi because ultimately the main counterparties are highly centralised but with more obscure controls and less transparency than tradfi. regulation on on-ramps off-ramps outside of US has become alot tighter making it much more difficult to buy significant amounts of crypto, crypto asset providers themselves are increasingly subject to same controls on KYC AML as banks.
In terms of use cases promised in 2017 vs what’s been delivered AND practically useful to society or even practical, is a small fraction.
So long term thesis remains largely pure spec. It’s hard to be bearish any asset in USD terms because of the massive devaluation of central bank issued currencies in developed economies.
But it doesn’t have any convex properties from a fundamental perspective that a favourable vs something like a company that actually makes real money.
I’ve made life changing returns personally from being long of crypto and trading crypto primarily from the long side. But I think an objective and sombre takeaway from the “cyberpunk” and decentralisation and Web 3.0 promises from 2016-2017 era turned out to be empty.
When you look at crypto vs a technology which has delivered actual societal usefulness like AI it pales in an even more stark comparison. AI related companies have delivered 1000%+ revenue growth etc. in crypto, outside of pure speculation related and grift type marketing level businesses, those profitability metrics are missing. The most profitable businesses have been exchanges which just monetised speculation. Which is what crypto ultimately solely is. Hence another reason market has been selling hot air pure play themes which have no backing from
Actual corporate profitability economics: quantum computing, space etc .
If market is hot and animal spirits are abound, crypto can go into bull, but it’s difficult to see in their absence why Btc long term fundamental value is 60k rather than 30k. So value investor’s argument will ultimately prove right about terminal value of crypto.
So we are in a position where market action is saying bear and something is different, and when you look at “unseen forces” and how they are stacking up for crypto you don’t get any encouragement either.
It’s possible it’s peak bearishness, doesn’t feel like that to me. Feels like disbelief . In any case standing in front of such moves trying to get long is big -EV over the whole distribution of such events. Are there any special factors that make it particularly attractive in crypto? Not to me, I don’t see them.
Joining the rout at declining MAs, have been short ethereum primarily due to RW vs btc, via options and ETFs and some perps across accounts.
Market has experienced extreme rotation on this selloff in stocks: defensives have gone intk top 95%+ valuation extreme on forward P/E basis while tech has deflated to way below average.
I decided to look at extreme rotations of this nature as a signal and see whether it on its own predicts future 3mo market returns.
Distribution splits into pre 2010 period where these rapid dislocations in valuations have preceded further drawdown in indices and 2010+ period where they have led to rallies. 2016 episode led to a rally, 2018 episode led to final further leg down before bottom and rally in the new year.
Pre 2009 episodes are heavily concentrated on dotcom unwind, with several instances in early bear leading to further downside, and the November instance leading to market bottom.
Record a signal when consumer staples, represented by XLP, have beaten the Nasdaq-100, represented by QQQ, by more than 10 percentage points over the preceding 21 trading sessions.
We wait 63 sessions before recording another signal to avoid overrepresentation of one period in the overall sample.
Average 3month future market returns are negative, however there’s huge amount of variance.
Some extremely negative periods like dotcom “color” the 2000-2009 sample
2000–09
13 observations
−11.9% avg 3mo Nasdaq return
2010 onward
10 observations
+7.9% avg 3mo Nasdaq return. If you exclude 2020, the avg is closer to +4.8%.
So it’s fair to say the distribution is heavily left tailed (contains a bunch of huge drawdowns in markets) and some bigger than avg upswings too but less so than the negative ones, thus skewing the average to negative.
Fundamentals are quite different.
The resolution really comes down to whether we are still in a 2010+ regime, in which case this is more likely to play out like the 2018 instance. Or there’s a deeper unwind incoming which will probably take much longer to heal.
A lot of charts of single names look similar to dotcom 2000, ie longer bear and a multi leg unwind. Failures at key declining MAs the nature of rallies being and feeling like counter trend so far is further evidence that skews odds to this being the more likely scenario as things stand.
What this sample does confirm is that volatility is likely to remain high regardless.
Volatility subsiding for several days with HIGHER lows in place, OR another rotation signal in the opposite direction seem like useful signals to watch for to signal a key change in conditions.
Change of character has to come at the 20day EMA and with force, in which case should be willing to pounce on the V.
Am staying open to both scenarios and updating perspective on day to day basis.
Minerva Capital retweeted
This data lines up with what we see on our side as well.
Really hard to deem last Thursday a “clearing of risk”, when exposures came right back four days later.
BREAKING: Global equity funds attracted +$98 billion in inflows in the week ending November 11th, their 2nd-largest weekly inflow on record.
This marks the 5th consecutive weekly intake, totaling ~$350 billion.
As a result, the 4-week moving average of inflows is up to ~$75 billion, its largest reading on record.
By comparison, during the memestock mania in 2021, the 4-week average of inflows peaked at ~$45 billion.
Meanwhile, global equity technology funds attracted +$2.6 billion in the week ending November 11th, bringing their 3-week inflow up to +$9.1 billion.
Investors are aggressively buying the dip.
Minerva Capital retweeted
NEW ODD LOTS: NVIDIA EPISODE
What the Megaspesd scandal is really all about.
@tracyalloway and I talk to @mackhawk, who has been covering Megaspeed for a long time, and knows more details about chip smuggling and export controls than anyone you know. podcasts.apple.com/us/podcas…
Minerva Capital retweeted
A year ago, we asked Nvidia about a Shanghai data center with links to Alice Huang. They declined to answer our questions and insisted that when it came to Megaspeed, there was nothing to see
We now have filings showing there are indeed banned Nvidia chips at that site
Thread:









































