Jeremy Grantham — We Asked Jeremy Grantham Why AI Won't Boost Profits — and What It Will Do Instead
Jeremy GranthamexcessreturnsLong-term
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
Rejects the "permabear" label — distinguishes baseline caution from rare "abandon ship" calls (2008, 2021).
Mean reversion remains the master variable; high returns on capital eventually attract competition, regardless of how durable a moat looks today.
Monopoly power explains the Mag 7's margin profile, but AI is the tool most likely to dissolve that monopoly — turning today's winners into brutal competitors.
New technology historically becomes a cost of doing business, not a permanent profit boost; AI follows that template.
Bubbles defined as two-sigma deviations from trend (Japan, dot-com, housing, 2021). Current US market shows classic bubble signatures.
Institutional career risk is the reason value strategies break down inside bubbles — being early is indistinguishable from being wrong.
ChatGPT's late-2022 arrival short-circuited what would otherwise have been a normal bear-market completion.
The one lesson for ordinary investors: long time horizons + realism about pessimism, not optimism, are the edge.
Mike Gyulaimacro-opsNear-termNQ1!ZR1!CC1!ZM1!ZMZ2026ZLZ2026RSX2026KC1!+2
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
NQ printed a Doji above weekly trendline — next directional break (up or down) determines whether v-shaped rally continues or reverses sharply
Broad commodity complex trend-change theme playing out: Rice hit 1X measured move, Cocoa retracing to breakout level, Soybean Oil already in uptrend, Canola similar
Coffee broke below weekly support — MO is short July contract; Soybean Meal attempting breakout at ~323 on December contract
Key data5▾
Rice futures (ZR1!)hit 1X measured move target; MO long risk-defined
Coffee (KC July)below weekly support; MO short risk-defined
Soybean Meal Dec (ZMZ2026)breakout attempt at ~323 level
Nasdaq (NQ1!)weekly Doji at trendline; 2-3 week directional resolution pending
Cocoa (CC1!)retracing to breakout level after hitting 1X target
Marc Rubinstein — He Studied the Financial System for Decades: Where the Real Risk Is
Marc RubinsteinexcessreturnsLong-termKBWBDCOBDCGSJPMPNCBAC
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidence
Post-2008 rules pushed risk from banks into private credit, hedge funds, trading firms, and exchanges — creating a "layer cake" of leverage connecting banks, private credit, and borrowers.
Private credit redemption risk is not classically systemic (no bank run possible) but still matters via headline/reputation/liquidity contagion — Blue Owl as poster child.
Insurance/private credit entanglement (Apollo/Athene model) has accumulated fast, with incentive conflicts and regulatory gaps that represent latent risk.
Rubinstein's "growth is bad" axiom: rapid growth in financial businesses always degrades underwriting quality — SVB and Revolut both illustrate this.
A "golden age of arbitrage" is underway, driven by geopolitical fragmentation pricing assets differently across markets and a widening gap between public and private asset valuations.
Most misunderstood: investors still reach for the 2008 bank playbook on every wobble; the actual risk is now in NBFIs and government bond market structure.
Key data12▾
Blue Owl Capital Corp IIredemption requests exceeded 5% quarterly tender cap
Blue Owl settlementfunded 30% of redemption queue, remainder deferred
JP Morgan$160B lending to private credit vehicles on balance sheet
Bank lending to non-depository financial institutions~5x growth over 10 years (~16% CAGR)
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidence
As AI drives supply of commodity goods and services toward zero marginal cost, economic value concentrates in two places: unrepeatable objects/experiences (the "totem" tier) and the irreducibly human relational core inside jobs (the "maker" tier).
Manidis's portfolio call: short Amazon-deliverable commodity goods; long first editions, beachfront, pilgrimage-style experiences — the K spread will widen further.
Imas's counter: jobs are O-rings, not separable task stacks — automate the reading but not the explaining, and the radiologist's job holds; most working life will remain in "maker-scarcity" rooms, not totem rooms.
The synthesis: both are right on different time horizons and different population strata; the mix of object-scarcity vs maker-scarcity in any given sector is the decision variable.
Key data5▾
Starbucks reversalremoved baristas → customer walkoff → Niccol 2024 letter restored human element
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
GOOG and AMZN booked a combined $53B in Other Income in 1Q26, of which $49B was equity stake revaluations in private AI companies they also fund and sell cloud to.
The circularity now runs through the income statement (earnings), not just revenue — analysts upgrading on "earnings beats" are validating marks set by the companies themselves.
Hyperscaler purchase/lease commitments exceed 100% of operating cash flow (Oracle >700%), with most off balance sheet; the flywheel requires ever-rising private AI valuations and perpetual capital inflows.
Paulo draws direct parallel to PE mark-to-myth era; the tell is lagging free cashflow while "earnings" growth accelerates.
The unwind trigger: Anthropic/OpenAI IPO, or any participant flinching from the next up-round (Prisoner's Dilemma).
Key data7▾
GOOG + AMZN Other Income 1Q26$53B combined
Other Income from stake revaluations$49B of the $53B
Big 5 hyperscaler bottom lines~34% from Other Income
Alexander CampbellcampbellrambleNear-termZWZCSBSIGLDSLVTLTUSO+4
POSITIVEPositive sentiment — analyst is constructive / bullishHIGHHigh conviction — strong directional view held with confidence
The "dissonance framework": bad down days = internal dissonance with your worldview; good down days = dissonance with the market while your conviction grows. Today was the latter.
Added to wheat (calls + futures) after spec washout despite stronger WASDE fundamentals; added to energy exporters as Hormuz equity transmission; re-entered short equity because the rally ran on positioning/easing flows that are fading; sold more bonds to fix duration imbalance.
Sugar thesis intact but catalyst delayed by Petrobras pricing freeze keeping domestic gas cheap and cane-to-sugar ratios elevated; need to roll near-and-out calls.
Portfolio risk at 18 vol, targeting 15 vol by late summer for institutional conversations; biggest tail risk is a deflationary rip or unilateral Iranian capitulation on Hormuz.
Key data6▾
US wheat production (WASDE)42 MMT vs 54 MMT last season
Global wheat ending stocks275 MMT vs 281 MMT consensus
E15 House vote218-203 (standalone, passed)
Portfolio vol18 vol (cap), targeting 15 vol by late summer
Remaining risk capacity~20% more exposure before risk controls trigger
Paul PodolskypaulpodolskyNear-termSPYTLTGLDSLVEURUSDUSDJPYBRLNOK+3
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidenceFLIPPEDAuthor flipped — analyst has materially changed their view without a formal rating change
Iran war keeps energy elevated → US inflation at ~4%, EU at ~3% and rising → Fed tightening on the way.
Bonds are the clearest short: nominal growth ~7% vs. yields below 5%, budget deficit, war spending, tariff refund, and likely a dovish new Fed chair all add to long-end pressure (UK gilt analog).
Tech equities resilient (productivity boom, earnings keeping pace with prices, short-squeeze dynamics) but will face headwinds; precious metals (gold, silver, copper) bearish on tightening.
Currency view: elevated oil good for BRL/NOK, bad for EUR/JPY; Fed tightening adds USD-bullish pressure on top.
Key data6▾
US inflation~4% and rising
EU inflation~3% and rising
US nominal growth~7%
10Y bond yieldbelow 5%
S&P drawdown (2022 tightening)~15%
Iran war resolution timelineextended (no near-term end)
Kevin Muir — The Largest Mid-Cycle Earnings Acceleration on Record
Kevin MuirthemacrotouristNear-termSPYQQQRSPSMHIGVMAGSMU
NEGATIVENegative sentiment — analyst is bearish or sees downside riskMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertaintyFLIPPEDAuthor flipped — analyst has materially changed their view without a formal rating change
2026 Q1 was the 8th best S&P earnings beat in 22 years — but every other beat of that magnitude was a recession bounce-back. The only non-bounce-back parallel for NTM EPS revisions is late-2017/early-2018.
Trajectory was "average" for 30 days then exploded — 105% of the revision was from 10 stocks: Micron alone ≈50%, six energy stocks, three AI capex plays. Equal-weight S&P and Nasdaq beats are typical, not "out of the norm."
Forward earnings curves got raised across the next year — not just the front quarter — at a near-unprecedented rate. "Analysts were way too bearish in '23/'24 and are making the opposite mistake today."
P/E compressed across all major US indexes EXCEPT semiconductors, where investors paid more for the earnings (NTM PE 24.1x→27.0x YTD) — "the market is assuming that semis are no longer cyclical. Sold to them."
Concentrated short stays on the MAG7 (works, second-worst index after software); EPS changes in expectations drive prices, not earnings themselves, and the hurdle for the next leg up is now much higher.
Key data9▾
Q1 earnings rank8th best of 88 quarters since 2004
2026 full-year EPS growth expectation~24% (cited via Yardeni)
Q1 YoY EPS growth~18%
Top-10 stocks share of S&P EPS revision105%
Micron share of top-10 revision~50%
Energy stocks in top-106
AI capex stocks in top-103
SMH NTM P/E24.1x → 27.0x YTD
Only non-recession NTM EPS parallellate 2017 / early 2018
The Blind SquirrelblindsquirrelmacroNear-termSOXXKWEBFXIXLIXLBXLEXLYHYG+2
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
"Stealth QE" via Bessent/Warsh: JP Morgan reserves at the Fed down ~$200B as banks rotate reserves into Treasuries, absorbing ~$1T more in supply while QT is the nominal posture.
Alexander CampbellcampbellrambleImmediateUSOBNOCLDBA
NEGATIVENegative sentiment — analyst is bearish or sees downside riskMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertaintyFLIPPEDAuthor flipped — analyst has materially changed their view without a formal rating change
Sold 15% of book in oil futures against existing calls — portfolio now neutral to a down move in oil while keeping the convex upside. "I want the vol, I don't want the direction."
Trigger: Spanish hantavirus outbreak — cruise-ship R0 ~2.7, ~40% fatality, droplet (likely) not airborne. Ship environments historically ran ~5x hotter than population R0; rough mapping puts population R0 below replacement (~0.5), so self-extinguishing is base AND bull case. But.
Public-health response looks inept — WHO last update 6 days ago for a disease in 23 countries at 40% CFR. The narrative is the trade, not the virus.
Keeping the agricultural longs — "diseases and potential lockdowns may be bearish for fuel, they are not bearish for agriculture." Already regrets having sold June sugar.
"I gave you my best 'shorting a bubble' essay yesterday. Today, I am wedging a hedge into a transportation tail. Don't trust machines either — AI safety filters made the actual research nearly impossible to do."
Key data6▾
Oil delta hedge~15% of book sold in futures vs calls
Hantavirus ship R02.7
Hantavirus CFR estimate~40% (vs COVID ~1%)
Implied COVID-vs-hanta ratios~7x less contagious, ~50x more deadly
Population R0 mapping~0.5 if ship/population ratio matches COVID-era pattern
John Authers — Is Polaroid Coming for Today's Stock Market?
John Authersbloomberg.comNear-termQQQSPYSMHTLTVXX
NEGATIVENegative sentiment — analyst is bearish or sees downside riskMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
Top 10 S&P 500 stocks (all tech except BRK) = ~40% of index, comparable to Nifty Fifty's 45% but on far narrower sector diversity
PPI running >6% YoY; electronics components +27% YoY; PCE deflator forecast at 3.8% — too hot for Fed cuts
VIX and S&P moving in lockstep (not inverse) — rare signal that preceded Volmageddon (2018) and last year's software selloff; suggests unstable rally
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
Bill submitted last night; under a 90-day campaign requirement, August is logistically dead, leaving September 1 or September 15. September 15 is the bet.
The legal trick: the bill does NOT set a date — it delegates that to the House Committee chaired by the bill's sponsor. Netanyahu sets the pace; coalition holds until the last minute.
Likud primaries math: polls show Likud bleeding seats → fewer viable list slots → MPs won't whip for the unpopular Haredi draft exemption. Calling early is cheaper than survival.
Strategic read: Netanyahu had been waiting for an Iran regime-change image as backdrop. Pulling the date forward signals he either expects something by September 15, OR he's accepted the "big event" isn't coming by late October.
China sub-piece: Netanyahu's 60 Minutes answer on China was unusually cautious. Trump in Beijing this week; "every dollar the US invests in interceptors for Israel is a dollar not invested in submarines in the Strait of Malacca." The Iran war is China's way of draining American attention from Taiwan.
Key data5▾
Likely election dateSeptember 15 (Sept 1 = back-to-school, both Tuesdays compromised by High Holidays)
Campaign period (legal minimum)90 days
Likud pollingbleeding seats vs current Knesset count
Chinese share of oil imports through Malacca~80%
Vehicle penetrationmore Chinese cars in Israel than in US or Europe (per anonymous senior Israeli official, paraphrased)
Andy Constan — He Invested Through Five Bubbles: What They Taught Him About AI
Andy ConstanexcessreturnsLong-termSMHNVDAQQQSPY
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
Defines bubble regimes by three sequential phases: root conditions (regulatory change, easy money, new tech, exogenous shock), escalation events (leverage or crisis-driven liquidity injection), and peaking (expectation explosion + FOMO).
Maps five historical episodes: 1982–87 LBO bubble, 1995–2000 internet, 2005–08 housing, 2009–2022 bond bubble, and today's AI trade; draws direct timing analog — 2025 Liberation Day = Oct 1997 Asia crisis; 2026 Iran war = Oct 1998 LTCM.
The bear case is financing, not valuation: semiconductor P/S is extreme while P/E looks fine because earnings expectations doubled; the real risk is IPO supply (three frontier-model listings) + corporate debt issuance + cancellation of hyperscaler buybacks draining equity demand.
Key data7▾
LTCM bailout size$1.3B from 11–13-bank consortium
Greenspan irrational exuberancemarket +40% after the comment, then he cut anyway
NASDAQ post-LTCM+60% in 6 months into March 2000 peak
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
30yr gilt yields hit 5.81% (highest since 1998) and 10yr hit 5.13% (highest since 2008) within days of the local election results — the domestic political risk premium is being priced explicitly at the long end, where fiscal credibility lives.
Both leadership scenarios are gilt-negative: Starmer stays and loosens fiscal rules to survive; Starmer goes and a challenger (Burnham most likely at 42% in polls) arrives with an expansionary mandate.
The BoE cannot bail out the long end: held at 3.75% in April (8-1 vote), CPI at 3.3% and rising, hikes expected this summer — front-end tightening removes any bull-steepening support.
OBR elasticities: a 1% gilt yield rise adds £15B to 2030-31 borrowing; a 1% RPI inflation rise adds £11B — the fiscal feedback loop is now self-reinforcing.
Actual FY2025-26 borrowing came in at £132B (£0.7B below OBR forecast) — the fiscal data is not yet the problem; the political uncertainty about the fiscal path is.
Key data13▾
UK 30yr gilt yield (peak this week)5.81% (highest since 1998)
UK 10yr gilt yield (peak this week)5.13% (highest since 2008)
Gilt Trip spread trade67bps entry → 75bps (profit already running)
UK 10s30s spread~66bps; target 90bps; stop 59bps
Labour local election loss-1,229 council seats
Reform UK local election gain+1,372 council seats
Labour MPs calling for Starmer to go>80
Burnham polling lead in Labour members42% vs 11% (Rayner/Streeting)
UK FY2025-26 borrowing£132.0B (£19.8B lower YoY; £0.7B below OBR)
OBR1% gilt yield rise adds £15B to 2030-31 borrowing
OBR1% RPI rise adds £11B to 2030-31 borrowing
BoE April vote8-1 hold at 3.75%; CPI at 3.3% and rising
John Authers — Warsh Could Find Inflation Too Hot to Handle
John Authersbloomberg.comNear-termTLTIEFUSOBNOCLSPYSOXCSI300
NEGATIVENegative sentiment — analyst is bearish or sees downside riskMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
April CPI: 3.8% headline, 2.8% ex-energy; sticky prices and Supercore both re-accelerating
Real wages negative for first time since 2022 — affordability squeeze politically toxic
2-year yields at 4% (highest since June 2025); market pricing next move as hike not cut
Brent December futures retouched $90+ as Trump called latest Iranian peace proposal "garbage"
Trump-Xi summit backdrop: China exports +14% YoY in April despite Iran war freight costs; rare earths leverage intact
POSITIVEPositive sentiment — analyst is constructive / bullishHIGHHigh conviction — strong directional view held with confidence
The market move has gone "just far enough for long enough in exactly the wrong names" that underweights now have to buy — admission, not conviction. SPX 7400 target (his prior call) traded.
Dealer gamma flipped from short-gamma (every dip disorderly) two months ago to long-gamma now → dips dampened, realized vol below what Iran tape suggests, "even more irritating" grind higher.
Reflexive flows are bullish until reflexivity turns: overwriters covering and rolling premium into the move, discretionary underweights chasing, levered ETF AUM huge.
"Don't choose between real and mechanical — both are true." Earnings/capex/revisions real; flow mechanical; call buying reactive; long gamma stabilising.
Specific positions: cut long call leg of 6500/7000 SPX risk reversal; FSLY position (2% wt) hit post-earnings, not concerned; looking to add CRDO above prior level (target >$300 this year); looking to add Swiss equity upside via low-IV calls.
Edward Chancellor — He Wrote the Book on Bubbles: Capital Cycles, AI, and Anti-Bubbles
Edward ChancellorexcessreturnsLong-termGLDCOPXFCXNVDAAMZNTSLAAAPLMETA+2
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
Every major technology transition has ended in overinvestment and investor losses — railway mania, autos, aircraft, telecoms, dotcom — and AI fits the template: prisoners' dilemma dynamics force hyperscalers to keep spending even when aggregate returns are suboptimal.
LLM hallucinations are not a temporary bug but structurally inherent to token-probability inference, which Chancellor argues will cap the total addressable market for truly mission-critical AI applications; a "Hindenburg moment" (single high-profile failure triggering widespread second-guessing) is a tail risk.
Anti-bubbles are the better bet: during the TMT bubble, old-economy value stocks were the opportunity; during the energy-transition SPAC boom, traditional energy at 2% of S&P was the setup; today, AI-disruption-narrative selloffs in non-disrupted businesses (travel, exchanges, car auctions) may be the equivalent.
Key data9▾
Railway mania capex~10% of UK GDP (vs AI today: lower)
NASDAQ dotcom peak-to-trough~78-79%
Amazon dotcom peak-to-trough>90%
Energy sector S&P weight at SPAC-boom trough~2% (vs 8-10% historical avg)
Alexander CampbellcampbellrambleNear-termSPYHYGTLTEWCSLV
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
Three ways to short a bubble: (a) Wedge — go long the trend that kills it (rates, in this cycle), (b) Victim — short something non-convex correlated to the bubble where put vol is still cheap, (c) Confirmation — wait for fundamentals to weaken, vibes to exhaust, and the chart to break the line.
Directly shorting parabolic names is double-negative-convex: exposure grows with price, options are 100-160 vol so paying half the share price for premium, and a short who covers is "a buy that's coming back tomorrow to clear his tab."
Today's CPI print confirmed inflation re-acceleration; the AI bubble is a long-duration asset that compresses when rates rise. "If AI is the bubble, rates are the wedge."
Hedges added today: short 5% SPX, short 10% HYG, treasury put spreads (TLT puts), sold 5% of long Canadian bank position.
Canadian banks at 3x book stuffed with negative-amortizing mortgages, and private credit ("a roach motel — the money checks in and doesn't check out") are the candidate wedges; victim leg is "honestly, I'm not sure" yet.
Key data2▾
Today's actionsshort 5% SPX, short 10% HYG, treasury put spreads, sold 5% of Canadian bank long position
Lake Cornelia CommentarylakecorneliaNear-termSMHSPLVSPYQQQ
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
"AI will be bigger in 2 years than it is today. Many of these stocks will multi-bag. In the short term, we see all the signs of an imminent reversal and overstretched stocks."
Factor dispersion extreme: 10% spread between Momentum and Value on the day; Quality just as bad. YTD same pattern.
SMH closed $576 vs 200dma $370 — 55% premium to the 200dma. RSI hasn't been below 70 (overbought) since mid-April.
The "contra trade" — SPLV (low-vol) — has been in deep correction since Iran-war fears peaked. Suggests low-vol oversold as a long.
Breadth deteriorating: % of stocks above 50/200 DMA declining for 2-3 weeks; outright readings (60% / 55%) "pretty average — the bubble is localized and now sucking the air out of the rest of the market."
"Dollar Neutral PM" framework (his April X-post, reproduced): in jump/stress scenarios, betas don't collapse to 1 — they collapse to the cube root of the original beta (2β→1.26, 3β→1.44). Pod-shop dollar-vol budgets get pressured by expanding regression vol; pods can't be contrarian; almost all are offsides in the aggregate. "Tomorrow will be insane. Utter positioning bloodbath…"
Key data5▾
SMH$576 close vs 200 DMA $370 (55% premium).
SMH RSIabove 70 since mid-April.
Day-of factor dispersionMomentum vs Value 10% spread; Quality similar.
Arnaud Bertrand — The Neocons' Come-to-Jesus Moment
Arnaud BertrandarnaudbertrandLong-term
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
John Culver (former NIO for East Asia, CIA since 1985) tells Max Boot via WaPo: in a Taiwan war, "we need to get our high-value naval assets out of the theater, and then we would have to fight our way back in. From where, it's not clear. Guam is no bastion either."
China leads in "most military domains — and it's not even close." Submarines/undersea is the only domain Culver names as a remaining US advantage. In advanced munitions China leads "by magnitudes."
Shipbuilding: a single Chinese yard (Jiangnan, near Shanghai) "has more capacity than all U.S. shipyards combined." Total Chinese naval shipbuilding capacity = 232x the US (leaked US Navy briefing slide referenced).
"China deploys enough ships every year to replicate the entire French navy."
Despite the gap, Culver judges war unlikely: "Taiwan is a crisis Xi Jinping wants to avoid, not an opportunity he wants to seize." China gets what it wants by waiting until the US security guarantee is no longer credible enough to honor.
Trump's $1.5T defense budget (50% increase): "throwing good money after bad," per Culver. Trump-class battleships and the "Hellscape" drone strategy dismissed as nostalgia and impractical respectively.
Author's framing: two establishment hawks publishing the obituary of US military primacy in 48 hours. "The arsonists are now writing the fire report."
The Curious MindthecuriousmindLong-termUSOBNOCLSMHNVDA
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidenceFLIPPEDAuthor flipped — analyst has materially changed their view without a formal rating change
The 1973 playbook said: Middle East war → don't own anything that needs cheap power or cheap capital (memory chips need both). Since Hormuz closed Feb 28, oil +45%, Micron +80%, NVDA +15%. Memory chips outran a Middle East war.
Cut AI book from 30% → 15% at end of March, added Offshore Oil Services; sold the bottom; bought back higher in early April on a thesis he couldn't yet name.
Two mechanisms why the 1973 framework failed: (a) algorithmic marginal trader with no 1973 anchor (no FOMO, no fear, no memory; the bubble-shaped trader has been replaced); (b) absorbed shock — WFH and business-travel cuts can absorb 10-30% of a Hormuz disruption without recession-style demand destruction. The "air pocket" is real but it shows up as sticky inflation, electricity costs, capex crowding — not gas lines.
Cited: Dylan Patel — AI demand is a phase change, Anthropic from $9B → $35-45B ARR in months, gross margins 30% → 72% and rising, supply chain sold out two years deep; model release cadence compressed from 6 months to 2.
"I had been watching the right indicator for the wrong regime."
New operating rule: Do not be religious on your views. Markets pay you for "the accuracy of your reading of the system as it currently is," not for framework elegance. The slow variable is calibration of your own lens, not the world.
"Move 78" framework: Sedol's only win against AlphaGo came from a depleted body, not analysis. The body knows before the mind does — discipline is "the bandwidth to hear what the body is already saying" when the price/headline canary has been hollowed out.
Information-diet corollary: name your five sources (Substack/X/podcast/chat/newsletter), make sure two are thinking against the other three, at least one isn't on the same side of AI as you are.
Key data8▾
Oil since Feb 28 Hormuz closure+45%.
Micron since same date+80%.
NVDA same period+15%.
AI book sizing30% → 15% end of March, then re-added in early April higher.
Alexander CampbellcampbellrambleLong-termEWYEWTINDAUSOBNOCL
POSITIVEPositive sentiment — analyst is constructive / bullishHIGHHigh conviction — strong directional view held with confidence
The "peaceful" China-as-reserve-challenger thesis dies on "what do I do with my RMB" — closed capital account, broken banks, no Treasury-equivalent.
New construct: AI Tigers (Taiwan/Korea/Japan) are now the marginal global surplus generator on AI-cycle exports — Taiwan +$181B current account in 2025 (20% of GDP), Korea +$123B, Japan +$204B; half-trillion-a-year combined and accelerating.
The recycling channel quietly changed holders — private (Korean lifers, Taiwanese insurers, Japan pensions) now absorbs the surplus instead of central banks, making the dollar plumbing fragile (TWD moved 5% intraday in May 2025 on lifer hedge-ratio wobble).
Trade: long TWD/INR and KRW/INR via NDF risk reversal — equity ratios (EWY/INDA +240%, EWT/INDA +110%) have moved while FX hasn't (KRW/INR +7%, TWD/INR +7.6%). Premium-neutral, sized for 10-15% spot move.
Energy is the accelerant: Brent $150 right-tail blows out India's CAD toward 5% of GDP ("1991 territory") and forces RBI to hold/hike, widening the rate differential against the carry.
Key data11▾
Taiwan 2025 current account surplus$181B (20% of GDP)
Korea 2025 current account surplus$123B
Japan 2025 current account surplus$204B
Setser Korea+Taiwan customs surplus$400B trailing 3-month annualized, up from $100B two years ago (4x)
SSD pricesdoubled in 10 days; NAND +47% in a month
EWY/INDA equity ratio+240%; KRW/INR FX: +7%
EWT/INDA equity ratio+110%; TWD/INR FX: +7.6%
Short ratesIndia 5.25%, Korea 2.50%, Taiwan 2.00% — carry cost ~3%/yr on short INR vs long TWD/KRW
India FY2025 services surplus$180B; oil imports ~5mbd ≈ $140B at 2024 prices, $250B at $150 Brent
Alex Barrow / Macro-Ops — Straight Through the Top
Alex Barrowmacro-opsNear-termQQQIWMBTCSOLDAXSLVDBCKE+7
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidence
Portfolio +288bps last week, +53.8% YTD.
Core book: long AI/semis, long RTY, long BTC, short SOL/USD, short DAX, long silver, long Bloomberg Commodities ETF, long KC HR Wheat, long Dec Brent, long RBOB Dec gas, long 2y USTs, long lithium and rare-earth miners, plus a few idiosyncratic equities.
Qs blew through the upper bound of their megaphone pattern with six consecutive weekly bull bars — "firm evidence of a demand/supply imbalance that historically leads to further continuation."
INTC long, opened in early April at ~start-of-rally, +165% since; trailing stop on remainder after partial profit-takes.
Sentiment / flows at multi-year extremes: Citadel net buying 98th percentile since 2019 (April monthly 87th percentile); StateStreet equity allocation at 5-year high (biggest monthly shift since Nov 2020).
Negative divergences: LQD/IEF still diverging from equities, complement view from Jim Carroll's HY Bond Cumulative A-D line; Nasdaq-VXN correlation hitting an up-up extreme.
New trade idea: short coffee — weekly H&S continuation pattern broke, position/sentiment neutral, enters most bearish seasonal window in two weeks.
Key data5▾
MO Portfolio+288bps last week, +53.8% YTD.
INTC+165% since early-April entry.
Citadel platform net buying last week98th percentile since 2019; April monthly 87th percentile.
StateStreet equity allocationhighest in 5 years; biggest monthly shift since Nov 2020.
Arnaud Bertrand — America's Chief Warmonger Says the US is "Checkmated" by Iran
Arnaud BertrandarnaudbertrandLong-termUSOBNOCL
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
Bob Kagan (Atlantic, May 2026) — co-founder of PNAC, Victoria Nuland's husband, Iraq-war pre-9/11 advocate — concedes US suffered a "total defeat" in Iran that "can neither be repaired nor ignored." Bertrand argues Kagan's points match his own prior "First Multipolar War" essay nearly point-for-point.
Vietnam/Afghanistan didn't damage US position structurally; Iran is "of an entirely different character."
Iran will not relinquish Hormuz — Bertrand quotes Iran's parliament speaker Ghalibaf: "The Strait of Hormuz situation won't return to its pre-war status." Kagan independently: "Iran will be able not only to demand tolls for passage, but to limit transit to those nations with which it has good relations."
Gulf monarchies have no real choice but to accommodate Iran — the "distant superpower that can't protect them" vs "the regional power that just proved it can hit them whenever it wants" — and Bertrand reads this as US "paper tiger" status.
Kagan: "if the United States with its mighty Navy can't or won't open the strait, no coalition with just a fraction of the Americans' capability will be able to, either." Bertrand: same point, citing Germany's defense minister Pistorius.
"Just a few weeks of war with a second-rank power have reduced American weapons stocks to perilously low levels, with no quick remedy in sight" (Kagan, quoted).
Author and Kagan diverge on (a) human/moral cost, (b) the neocons' historical responsibility for forging Iran as an adversary, (c) Kagan's solution (full ground/naval war and occupation), which Bertrand mocks.
Key data2▾
Atlantic article publicationMay 2026; author: Bob Kagan.
AP ResearchalphapicksNear-termSPYQQQIWMTLTIGLTUSDJPYBNOUSO+1
POSITIVEPositive sentiment — analyst is constructive / bullishMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
S&P 500 rose 10.49% in April — 99th percentile monthly gain since 1990, best since 2020; RSI ~72, technically overbought but not automatically bearish in a high-momentum regime.
Small-cap tech (Russell 2000 tech) up 50.5% since March 30 bottom vs 29.3% for large-cap; 22 of 185 Russell tech stocks up >100% this year — Paul Tudor Jones explicitly invokes late-1999 comparison.
April CPI expected at 3.7% headline (highest since May 2023), driven by energy; core at 0.34% MoM; Fed on hold but equity markets can shrug off the print while rates markets face Warsh-led hike pressure.
UK gilts: local elections (Reform gained 1,372 seats, Labour lost 1,229) convert a macro bond sell-off into a domestic political fiscal risk story; AP Research sees 20-30yr gilt yields moving higher.
Key data10▾
S&P 500 April return+10.49% (99th percentile since 1990)
S&P 500 RSI (14-day)~72
Russell 2000 tech vs large-cap tech (27-day)+2,119bps outperformance
Russell 2000 YTD lead vs S&P 500+736bps
Q1 EPS growth (blended)25.8% YoY vs 12.4% forecast
Positive surprise rate81.7% (vs 78.3% 4-yr avg)
April CPI consensus3.7% headline, 0.34% MoM core
Gasoline expected+6.2% MoM
Europe rate hikes priced 202660+bps
UK Reform local election gain+1,372 seats; Labour loss: -1,229
Leonid Mironov — ROI not AGI (Guest: Leonid Mironov)
Leonid MironovthemarkethuddleLong-termFXIMCHIKWEBLITUSOBNOCLSMH+10
POSITIVEPositive sentiment — analyst is constructive / bullishHIGHHigh conviction — strong directional view held with confidence
China's anti-involution campaign is a real, technocratic, multi-year policy that ended PPI deflation in Feb-Mar 2026 (March PPI +0.5%) and will reverse Chinese consumer behavior — implication: China starts exporting inflation, not deflation, into a world positioned for the opposite.
Q1 2026 Chinese corporate earnings up 16%; 2025 was first year on record with negative net issuance (more buybacks/dividends than IPOs); only 7% of ~170 trillion RMB household savings is in equities (down from ~25-30% in early 2010s) — long runway as savers rotate in.
Mironov is structurally bullish Chinese stocks (running a new Gavekal China fund with Louis-Vincent Gave in Hong Kong); valuations on KWEB ~1 std dev below historic average ex-AI/semi names.
Lithium: anti-involution removes the "China floods the market" tail; storage + EV demand point to shortages H2 2026, certainly 2027-28; no more $1,000/ton ceiling.
Oil/Hormuz: physical and financial markets diverging is the physical market being forced, not futures being wrong; a real Hormuz closure forces ~6-month inventory rebuild even if reopened tomorrow; Saudi can ramp Red Sea exports 4-5mb/d to 7-8mb/d within 12-18mo; right-tail asymmetry favors long oil.
Ceresna's bold call: SMH semis blow-off top in May 2026 (SMH 53% above 50w MA in 4 weeks; NDX 13% above 50d only matched in Sep 2020 and March 2000). High confidence on timing, low on price. Expects sector rotation, not S&P crash — initial ~5% S&P pullback, 20%+ in semis.
KOSPI is the new Nortel/BCE story: SK Hynix (25.5%) + Samsung (24%) = ~50% of index; SK Hynix +400% YoY. Ceresna thinks the Korean drop will be worse than SMH.
USDJPY: post-intervention setup similar to prior episodes — took weeks last time for yen strength to kick in; Muir says intervention alone is insufficient, GPIF needs to actually repatriate (Korea-style tax incentives).
Bitcoin gets a "C-minus" grade for this rally — risk-on participation without risk-on enthusiasm; bad sign if the tape turns.
Copper out of flag, fresh closing highs; copper stocks lagging (Freeport disappointed, Ivanhoe mine issues, Southern Copper) — Muir frames as part of broader non-AI underperformance that should catch up post-semi blow-off.
Key data26▾
China PPI March 2026+0.5%
China PPI Feb 2026first positive print
Q1 2026 China earnings+16% YoY
China household savings~170 trn RMB
Equity allocation in China savings7% (was 25-30% in 2010s)
2025 China net issuancenegative (first ever)
Tencent peak market cap~$1T → ~$700B
BYD supplier payment terms360d → 90d
China pork prices YTD-36%
SMH above 50w MA53% in 4 weeks
NDX above 50d MA13% (matched only Sep 2020, Mar 2000)
S&P 500 stocks above 50d MA52%
SK Hynix 1yr+400%
SK Hynix + Samsung KOSPI weight~50%
CBOE oil vol index71% (was 125% in March)
Russia fertility1.3 → ~1.5
Hungary fertilitythrough 1.5
Korea marriage rate+10% YoY
Pronatalism spend threshold~4% of GDP
SEO short-oil ETF AUM$126M → $1.2B since war
Saudi Red Sea capacity4-5 mb/d → 7-8 mb/d in 12-18mo
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidence
Pre-trade carry-to-vol screener: how much will it cost me to own this for 30 days vol-adjusted? Forces "thinking in bets" discipline.
EUR/HUF "Orban implodes" trade: ~4% annual carry differential → ~30bp/month against you; against very-low monthly vol this is ~-0.6 monthly Sharpe before you're even right.
Hard-to-borrow stocks tend to drop, but the borrow cost almost perfectly negates the negative return in aggregate; CoreWeave intraday short (sell 9:30, cover 4:00) made more money than holding through the borrow.
Sympathy trades (silver short when copper collapses, Trade Desk short when AppLovin craters): valid but very short-lived because there's usually no economic reason for the correlation — most are forced-liquidation unwinds.
Excess Returns — The S&P 500 is Just 46 Stocks. 89% of the Economy is Flatlining
Jack ForehandexcessreturnsNear-termSPYIWMRSPQQQSMH
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
S&P 500 effective stock count (HHI-based) is 46 — lowest on record; investors taking on more concentrated risk than they realize
Jim Paulsen: today's inflation is supply-side and temporary, not demand-driven like the 1970s; labor force growing ~0.5%/yr makes excess demand structurally impossible
87% of the economy is growing at near-stall speed (2.1% real GDP); the entire bull market has been driven by the 13% "new era" slice
Bridgeway paper ("I Know What You Did Last Summer"): small-cap premium re-emerges when you exclude IPOs and fallen large-caps from the small-cap universe
Chris Mayer: vertical market software has a durable moat vs AI disruption because it is the system of record / mission-critical infrastructure; horizontal software faces real disruption
Ian Cassel: best stock pickers hit 49%; the edge is magnitude and patience, not hit rate
Goldman flagged the first weekly rise in HF gross exposure as a "stop-in" — forced covering, not conviction.
Options skew at extremes: call skew ~100th percentile, put skew ~zero on 10-year lookback — "all-offense, no defense; fragile if flow reverses."
AI capex / token consumption: Goldman's 24x token-growth projection by decade-end is a "dangerous linear extrapolation"; wastage and efficiency gains will compress demand; GPUs no longer the binding constraint (explaining NVDA's relative underperformance).
xAI rented entire Colossus Memphis data center to Anthropic — "Musk has effectively become a landlord and ceded foundation model leadership." Implied neocloud valuation: 5-8x EBITDA fair (CoreWeave 11x 2026 / 5.5x 2027 as comp).
Equity supply story: S&P 500 capex $1.3T (2024) → $2.1T (2027); LBO take-out slowed; SpaceX/xAI IPO pipeline could put ~$2T of net new paper in. "Bearish equities? The IPO pipeline is your catalyst."
Breadth deteriorating: 50% of S&P names in technical downtrend; Friday saw XLK +3.5% but breadth only 37%, 9 of 11 sectors negative-breadth.
Trade idea: SMH calendar strangle — buy June OTM ~38-delta calls + September OTM puts; roll June calls up as the rally extends and use harvested premium to "free-fund" the September puts; ends as a free put for back-half summer.
Consumer short book: Wingstop and Shake Shack both missed; thesis is real-estate/food/labor costs rising into eroding middle-income demand. Calls the regime a transition "from K economy to gamma economy" — income and gains concentrating at the top.
Key data8▾
S&P 500 YTD+8.5%
Citadel platformnet buying last week in 98th percentile since 2019; April monthly flows in 87th percentile.
StateStreetequity allocation at 5-year high; April was biggest monthly shift since November 2020.
Damped Spring — A Bubble Is a Challenge to Investors
Andy Constandampedspring101Long-termSPYQQQIWMTLTIEFGLDDBC
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsHIGHHigh conviction — strong directional view held with confidenceFLIPPEDAuthor flipped — analyst has materially changed their view without a formal rating change
Calling a bubble regime in equities; will be early, doesn't predict the path or duration. AI capex is the driver.
Bubbles are low realized vol with extreme trending — that combination lulls investors into leveraging up at exactly the wrong time. Risk models demand it (cf. SVB 2023, banks pre-GFC).
Passive playbook: maintain risk target, but lower the leverage cap (e.g. 130% → 110%). Keep balanced allocation. Slow rebalancing to let trends run. Re-examine own FOMO/rug-pull psychology before deciding lump-sum vs DCA.
Active/leveraged playbook: alpha quality falls; momentum beats mean reversion; relative-value relationships diverge dramatically; don't fade the move semi-martingale; post-pop contagion is universal (1987 exception).
All bubbles eventually pop — Main Street pain follows. The pain is not optional. Whether it happens to you is.
Key data3▾
Personal markers cited1987 crash at career start, internet bubble (~4 years inflating), housing/GFC 2006-2008, bond bubble blowoff 2021.
2020/21 bond example75bp 10Y UST treated as risk-free; banks levered up into it (SVB, BofA).
Concrete leverage cap shift130% → 110% in bubble regime.
PauloMacro — When Trends Go From Seemingly Linear to Exponential
PauloMacropaulomacroNear-termBNOUSOUCOSCOCL
POSITIVEPositive sentiment — analyst is constructive / bullishHIGHHigh conviction — strong directional view held with confidence
Pros are universally "in headshaking disbelief" while generalists don't care — a linear-vs-exponential thinking failure approaching a tsunami moment.
US commercial inventory at ~460 mmbbls headed toward sub-400 mmbbls by July; absolute usable floor is ~350-370 mmbbls (linefill ~150 mmbbls is unusable). Margin is days/weeks, not months.
Three forces suppressing price: (1) Asian physical traders sat out April fearing a Trump TACO and are now ~200 mmbbls behind; (2) retail aggressively short via broken 2x products (SCO ~$1.1bn AUM, short 8% of WTI Jun27 OI alone); (3) SPR auction structure forced selling of fronts/buying of deferreds, then tweet + rumored BoJ intervention crushed the trade.
Brent managed-money length dwarfs WTI length — market is voting with its feet on US export-ban (jurisdiction) risk. Diamondback now publicly betting on a wider WTI/Brent spread.
Key data11▾
US commercial crude inventory (ex SPR)~460 mmbbls; estimated drawdown to sub-400 mmbbls by July
Practical floor (linefill + working stocks + min refinery stocks)~350-370 mmbbls
Linefill alone~150 mmbbls
May 4 SPR auction size92.5 mmbbls as a loan, 24% repayment premium due Q2 2027
USO AUM~$1.8bn (down from $2.8bn peak in March)
BNO AUM~$800mn
UCO AUM~$400mn (round-tripped despite ATH price)
SCO AUM>$1.1bn (up from ~$100mn in February, despite >50% price decline)
SCO short position in WTI Jun27~10k contracts ≈ 8% of total Jun27 OI
Mike Gyulaimacro-opsNear-termZRCCHEKCAUDNZDEURGBPGFRC+2
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
Money-projection framework: "Money is freedom/choice/security" → "Without money I am imprisoned/coerced/unsafe." Traders protect these projected qualities, not capital; this disconnects them from process as the account scales.
Trader-development path (response to Collective member with ~1-2y trading experience): pick ONE domain → audit recent work for flow-state activities → study obsessively 3-6 months (build database, backtest, develop framework) → trade in small size for judgment not profit → share and ask specific questions to break next plateau.
"Iron sharpens iron" as Macro Ops community principle.
Commodity whiplash: US/Iran peace headlines stopped out Brent, Soybeans, KC Wheat, Corn for small P/L. Rough Rice and Cocoa breakouts holding.
Lean Hogs breakdown (possible secondary completion of April H&S Top); Coffee H&S Continuation lower is the surprise — was monitoring 280-420 range, now appears to be breaking down.
AUDNZD broke down the Rising Wedge — small wedge, price above rising MA, expectations low.
NQ weekly breakout flagged with "unfathomable potential for further acceleration of the trend."
MIXEDMixed sentiment — analyst tone balances bullish and bearish elementsMEDIUMMedium conviction — constructive view but acknowledges meaningful uncertainty
SPX options are as call-heavy as it gets (91% call delta equivalent, ~$1T notional); 60% of record SPX call volume is zero-DTE — positioning is transient, not structural
Short-dated dealer positive gamma is suppressing vol now; after May OPEX those positions expire and the market "breaks out of the straight jacket" — expect RV contraction into 5/15 then V expansion into 5/19-5/20
COR1M (CBOE single-stock-to-index implied vol ratio) going sub-8 into that window is the key bear signal for hedging; if it holds above 8, bull case stays intact
Oil-equity correlation has flipped positive (both up together) — barbell long energy + long semis emerging as a hedge strategy instead of equity puts
Top options meme: Tesla calls cheapest in the complex and shifting to call skew; likely next chase candidate after AMD/Micron exhaust
Key data9▾
SPX call delta equivalent91% calls
SPX notional call volumerecord (Goldman, as of ~5/7)
Zero-DTE share of record day60%
VIX drop magnitude2nd largest ever from 25-30 level
Ben Norton / Geopolitical Economy — Trump Boasts "We're Like Pirates" as China Challenges US Sanctions
Ben NortongeopoliticaleconomyLong-termUSOBNOCLFXIMCHIKWEB
NEGATIVENegative sentiment — analyst is bearish or sees downside riskHIGHHigh conviction — strong directional view held with confidence
Trump in West Palm Beach speech bragged the US military is "like pirates" — seizing Iranian cargo ships, oil, and vessels themselves; "It's a very profitable business."
The Hormuz blockade was described by EIA as "the world's most important oil transit chokepoint" — ~20% of globally traded oil pre-war (Feb 28 attack).
IEA describes the war as having "unleashed the largest oil crisis in history"; food crisis follows from fertilizer-input disruption.
China is the largest importer of Iranian oil (>80% of Iran's exports). US Treasury sanctioned Chinese companies (24 April) and State Department followed (1 May).
Sanctions on Hengli Group (Dalian Changxing Island, 20 mm tons/yr) were the line-crossing event — Hengli is the endpoint of NE China's green-hydrogen/petrochemical export terminal infrastructure.
May 2: Beijing ordered all Chinese companies not to comply with US sanctions on refiners — the first invocation of China's blocking statute. Henry Gao (Singapore Management Univ): "decoupling is coming."
Key data6▾
Hormuz pre-war share of globally traded oil~20% daily.
US-Israel surprise attack on Iran28 February.
Chinese imports of Iranian oil pre-war>80% of Iran's oil exports (Reuters).
Hengli Dalian refinery scale20 million tons/year crude.
US sanctioned countries~1/3 of all countries globally; 60%+ of low-income nations (Washington Post 2024).
Lancet 2025564,258 sanctions-associated deaths/year on average (1971-2021).