Who's Coming For The Market In Q3?

Summary

• Avi’s core call: the regime that drove the melt-up may be closing. The hyperscaler-then-memory rally is now “baked into the expectations” — Micron’s “blow off top, you hit 1200, you’re back down at about 1000” — so he’s reducing Intel and Micron (with Intel a substantial chunk of his net worth) and considering a rotation toward indexes and AI’s downstream beneficiaries: biotech (ARKG up 15-20% in the two weeks since he flagged it) and Reddit ($191 today vs. $160 when first discussed). “I’m not saying go short the market. I’m saying that regime maybe is coming to a close.”
• The sharpest disagreement of the episode: Jonah says memory is still squarely in fundamentals territory — Micron trades at 7.5x forward earnings, and sell-side analysts are structurally disincentivized from bold calls, so the stocks “keep smashing analyst expectations” quarter after quarter; retail can ride that edge. Avi’s counter: earnings beats no longer hold — Micron crushed, popped 15%, and round-tripped, and “when you retrace a move like that… that’s probably a bad sign for things to come.”
• Jonah’s macro fear is the strengthening dollar. “Never underestimate just how terrifying a rallying DXY can be for your assets… we’re really just short dollars. Cash is trash… it’s all a short dollar trade with various degrees of beta.” Avi’s mitigation: the dollar bid may be yen-carry driven, and unlike 2021-22 the market is benefiting from AI-driven growth, not just Fed liquidity — “money is quite literally being made right now” — so dollar strength matters less for stocks but is “really bad for Bitcoin.”
• The Bitcoin fight, in the open: Avi sees no place for BTC to massively outperform over the next 1-2 years until the AI/robotics/biotech trades die down — Saylor’s dominance, Wall Street co-option, and the quantum question (why Zcash is doing well relative to BTC) have soured him. Jonah concedes the short term but calls long-run BTC “a one-way trade”: Saylor “is going to blow up, mark my words,” and from post-blowup lows (20-30k) he sees $1M — “50x from the lows,” possibly “the best portfolio hedge of all time.”
• Long HOOD over BTC is the scoreboard trade: since flagged a month ago, Robinhood +20% while Bitcoin fell 20%. It’s Jonah’s pure-play crypto expression — upside from any crypto recovery plus prediction markets, options growth, and Trump-account lock-in (Micron is contributing $250M to those accounts). He sees possible all-time highs and hopes not to sell for a year.
• Risk discipline for white-hot AI names: Intel’s new 20% two-day gyrations (118 to all-time highs and back) are new information — “if a stock’s vol goes up, by definition you should hold less of it.” Avi’s CTA framing: keep daily P&L variance constant — if position variance went from $350 to $3,500, peel some off. Context: Nasdaq’s Q2 +19.6% was the strongest quarter since Q2 2020.
• Both are hunting megatrends beyond AI for the wealth now being generated: Jonah toured Tel Aviv apartments (the shekel rallied 40-50% vs. the dollar on capital inflows; diaspora buying is secular), SF real estate is going “$2 million over asking,” and collectibles are financializing — an upcoming guest is raising a fund that buys dinosaur skeletons. The meta-advice: “get out of your little bubble.”

Deep dive

1. The melt-up regime may be closing — Avi considers rotating back to the index

• Avi’s framework for the moment: an analyst can collect 100 pages of evidence on why the 10-year moves, “but they can’t figure out what actually matters… who’s buying? Why are they buying?” What was driving this market — daily frontier-model releases, Anthropic “doing Claude for finance, Claude for this, Claude for that,” massive capex forward-purchases — is now “baked into the expectations.”
• The sequence he laid out on prior pods is playing out: hyperscalers led, memory followed, now value migrates downstream to what AI actually impacts. Micron: “blow off top, you hit 1200, you’re back down at about 1000.” He’s reducing Intel — which had grown to a substantial share of his net worth — because single-name downside risk is “much higher than it was two months ago.”
• The prescription is explicitly not bearish: “I’m not saying go short the market. I’m saying that regime maybe is coming to a close” — from his 80% single-names / 20% index mix back toward the index. Jonah, at portfolio all-time highs, feels the same unease without evidence for it: “it’s the gut that’s tingling… this ran way harder, way faster, and way longer than I ever would have expected.”

2. Fundamentals vs. flows — the episode’s real debate

• Jonah’s pushback — worth keeping: memory is not in meme-coin territory yet. Passive trillions index to analyst recommendations, and analysts are selected for hedging — the “weaselly guys and gals” who survive make “milquetoast middle of the bell curve calls” while the one who nailed 2008 got “written off into irrelevance.” So Micron and SanDisk “keep smashing analyst expectations precisely because the analyst community is disincentivized from making the bold call even when it’s correct.”
• His conclusion: the retail listener has a structural fundamental edge — “you can make a bold call in your PA” and ride the serial beats, because “the entire capital market system around these stocks is set up to misprice extreme events like AI demand blowing through a DRAM bottleneck.” Checked live: Micron at 7.5x forward earnings — “very much tethered to reality.”
• Avi agrees earnings will keep beating but insists flows now set price: Micron crushed earnings, popped 15%, then gave it all back as winners took liquidity out. “When you retrace a move like that… that’s probably a bad sign for things to come.”
• Avi’s proposed edge: buy these names when they’re down into earnings — the day-before de-grossing (Nasdaq -3%) made Micron’s upside asymmetric, which Avi tweeted pre-print — “but it doesn’t necessarily mean it’s going to be up in a month.”

3. Track the server rack, not the analyst deck

• Jonah’s commodity playbook applied to DRAM: the leading indicator for any commodity is the price of the fully assembled product — for crude it’s travel demand; here “the finished product is the server.” A fully kitted Nvidia GB200 rack runs roughly $7.5 million on the high end, up from ~$3 million a couple of years ago, “and it’s probably going to go higher.”
• With MU down 16% from peak and 8% on the day (Intel also -8%), his tactical read: “a good time to yolo into a little bit of Micron here for a trade on this pullback” — the shortage is still ridiculous, and pullbacks are for buying while the rack price keeps rising.

4. The dollar is the thing that mulches everything

• Jonah’s scar tissue as an oil trader: “I’d have some bull trade, some idiosyncratic thesis, then the dollar would strengthen a bunch and it wouldn’t matter. I would get dildoed.” His warning: “Never underestimate just how terrifying a rallying DXY can be for your assets… we’re all patting ourselves on the back here, but we’re really just short dollars. Cash is trash… it’s all a short dollar trade with various degrees of beta.”
• The macro backdrop he cites: M2 “ripping like there’s no tomorrow” for four years, and any slowdown in central-bank and Treasury profligacy — plus a rates trajectory “having gone from cuts to hikes” — is a big asset problem.
• Avi’s rebuttal: the dollar bid may be the yen carry trade (borrow collapsing yen, park in dollars for the rate spread — hat tip to friend-of-the-show Capital Flows), and dollar strength is more worrying for markets driven by Fed liquidity. “If the core driver is that money is quite literally being made right now because of advancements in tech, then it doesn’t matter as much.” What it is bad for: Bitcoin.

5. Bitcoin: quantum, Saylor, and an unresolved disagreement

• Avi took heat for his take on Pomp’s podcast and repeats it: his 10-year bull case “really just comes down to one question — will the Bitcoiners manage to solve their quantum issue?” Add Saylor’s dominance (“the future of Bitcoin is owned by one man in many ways”) and Wall Street co-option, and BTC “doesn’t exist outside the system the way it did before” — which is exactly why Zcash is doing well relative to BTC.
• Jonah’s rebuttal on quantum: it breaks JP Morgan’s RSA encryption too, and if Bitcoin were hacked the devs would fork it post-quantum “like they’ve forked it before” — “revert to the last saved point where we were alive.” He assigns that an “overwhelmingly massive probability.”
• Jonah’s long-run case is categorical: with socialism, asset seizures, and vote-buying via printed money in democratic governments, “it’s just a one-way trade for Bitcoin over the long run.” The short-run problem is Saylor: “he’s going to blow up, mark my words.” Jonah says whoever’s alive to pick up the pieces will have an epic run; Avi later says that from post-blowup lows (30k, 20k, “COVID levels”) he still sees $1 million: “50x from the lows,” possibly “the best portfolio hedge of all time.”
• Avi’s timeline: he doesn’t expect BTC to massively outperform over the next 1-2 years — “ironically kind of lines up with the four-year cycle” — until the AI, robotics, and biotech trades die down; Nasdaq has beaten BTC over five years. His closing rule, which Jonah loved: “Don’t make it your personality. Don’t make it your entire portfolio. Don’t forget to look elsewhere.” Or, per his bird-cage meme: “just walk out of the cage, guys.”

6. Long HOOD is the crypto trade that’s actually working

• Back-slap segment, but earned: since Jonah’s long-HOOD/vs-BTC call a month ago, Robinhood +20% while Bitcoin fell 20% — “you would have made some great money following that trade.” He’s still long HOOD, not short BTC.
• The thesis stack: if crypto comes back, HOOD benefits massively — plus prediction markets, options growth, and revenue diversifying quarter over quarter. New kicker: lock-in from Trump accounts, with Micron contributing $250 million and Trump pushing American companies to fund them — “a big boon for Robinhood.” Jonah sees possible all-time highs and hopes to hold it for a year.

7. When vol rises, hold less — the CTA discipline

• Jonah’s new information on Intel: it sold to 118, ripped ~20% to all-time highs in two days, then sold off again — gyrations it didn’t have before. “If a stock’s vol goes up, by definition you should hold less of it.” That’s now his only real issue with memory stocks.
• Avi’s mechanical version: do what a CTA does and keep daily P&L variance constant — even max-long in an uptrend, trend-followers trim as moves get bigger. “If the daily variance of this position was $350 six months ago and now it’s $3,500, maybe peel a little bit off.” It enforces buying low and selling high; and as Jonah adds, “you can’t pay yourself unless you take profits.”
• The reason the conversation exists at all: Nasdaq’s Q2 was +19.6%, the strongest quarter since Q2 2020 (+13% YTD). Jonah’s wish is a fear-driven cleanse — “if I can buy Intel at $80 again, I would just buy it and literally not think about it for a year.”

8. Megatrends beyond AI: real estate, dinosaur skeletons, and the passion economy

• Jonah’s epiphany from touring apartments in Tel Aviv and Jerusalem: anti-Semitism is a secular trend, diaspora Jews (London, Canada, France) are buying in Israel, and the shekel “rallied 40 or 50% versus the dollar” on capital inflows through a war. The generalizable advice: “get out of your little bubble… there are all kinds of megatrends,” and nothing beats property for levered length (75% LTV) — though not in SF, where AI winnings have homes going “$2 million over asking.”
• The collectibles thread: an upcoming guest, AJ Scaramucci, is raising a fund buying dinosaur skeletons (a reportedly Ken Griffin-owned dinosaur skeleton makes the rounds at Citadel parties); another prospective guest is really into Roman coins. Avi’s thesis: “if you generate deep knowledge in a specific subset of an area, you can make a lot of money” — financialization now monetizes any top-0.1% niche.
• They disagree, then converge, on “follow your passion”: Avi hated it when Matthew Fox delivered it at his ‘07 Columbia graduation (“I decided to just be really ridiculously good-looking… and you should too”), but Avi argues social media changed the math — “you could not make money being an underwater basket weaver 20 years ago,” now a 15,000-person TikTok audience pays. In an AI world, value concentrates where people care a human is doing it.
• Avi’s own revealed preference: he’s been offered capital to start a hedge fund and actively declined — “being a real human on your screens talking to you is going to be infinitely more valuable than starting a hedge fund now.” Jonah’s coda from his own mistake of taking too long off: take one to three months, not one to two years — “you learn more from doing than from sitting and watching on the sidelines.”