Former HFM | Trading & Investing | Rigorous Quantitative Research | Sophisticated Strategies | Institutional-grade for all Investors | Not Investment Advice
Signal #001: 20-Day Final
Published March 30. Tracked publicly every single day for 20 trading days. Every pick timestamped. Every update posted, the good days and the bad.
20-Day Results: Longs +20.50% | Shorts +0.03% | L/S Spread: +20.52%
SPY did +13.15% over the same period. Our long book beat it by 735bps. Our short book finished flat in a market that rallied 13%. That's the definition of a clean signal.
$STX +64.4%, $MOH +36.3%, $AMAT +25.3%, $PANW +18.5%, $BAX +17.0%. 7 of 8 longs positive. $RSG +7.0% and $MCD +5.8% led the short book.
This was our first public signal. We showed the work. Now it's closed.
Signal #002 is live. Future signals go exclusively to email subscribers at entry. Sign up link in bio.
Longs: $CRM $STX $PANW $GDDY $BRO $AMAT $MOH $BAX Shorts: $CMS $RSG $EVRG $ESS $CPT $MCD $DTE $VTR
$SPY $SPX $QQQ
The pressure zone is already here. The critical threshold is not a single price. It is duration. On MSTZ specifically, the 2x inverse daily reset means it is a momentum instrument, not a hedge. In a slow grind lower on MSTR it underperforms expectations due to volatility decay. In a sharp sustained drawdown it outperforms them.
bitcoin:native $MSTR
32 BTC out of 843,706.
That is 0.004% of holdings.
Same week he raised $128M through the ATM program and retired $1.5B in convertible notes. A $900M cash reserve is sitting untouched. BTC needs to appreciate just 2.3% annually to cover STRC dividends indefinitely.
This is not a crack in the thesis. This is treasury management.
Saylor told you he would buy 10 to 20 for every 1 sold. The people panicking over 32 coins are telling you they never read the filing. $MSTR bitcoin:native
Selling 32 Bitcoin proved several important things:
✅ Posting "Buying Bonds" was a bad idea
✅ Selling any amount of BTC was a bad idea
✅ Cutting to 6mo of dividend coverage was a bad idea
✅ Posting ads on how STRC held its peg was a bad idea
✅ Odds of SP500 inclusion: 0%
Bitcoin is down to $66K and nobody has a convincing explanation.
Here is one.
SpaceX. Anthropic. OpenAI. The biggest IPO pipeline in a generation is about to come to market. We estimated forecasts yesterday of $225B in 2026 IPO volume plus $500B in expiring lockups.
That capital has to come from somewhere.
Bitcoin trades 24/7 with zero lockup and instant settlement. It is the easiest asset on earth to liquidate when you need to raise cash for a new allocation.
This is not a coincidence. This is liquidity competition. The smartest money in the room is not selling because it lost conviction. It is selling because something better just showed up.
bitcoin:native $MSTR
WWDC 2026 is Monday. Here is what the data tells you heading into the event.
What Apple is expected to announce:
iOS 27 will let users choose between AI models for the first time. Bloomberg reported this on May 5. Apple is opening Siri to third-party models including Gemini, which Google confirmed has 900 million monthly active users as of May.
The OpenAI relationship is in flux. Bloomberg reported on May 14 that OpenAI has engaged outside legal counsel on options to detach from the Apple partnership, including a potential breach of contract notice.
Apple is also exploring chip manufacturing beyond TSMC. Early-stage talks with Intel and visits to a Samsung factory in Texas. No orders placed yet but the signal is diversification.
New consumer features include a bill-splitting tool in Wallet and iMessage using receipt photo scanning. Expect a full iOS 27, macOS, and watchOS reveal across the keynote.
The wildcard: Nvidia's new ARM-based N1X CPU announced at Computex targets the same AI PC market Apple dominates with M-series silicon. Apple's response at WWDC will set the tone for the next generation of the Mac lineup.
The 10-year WWDC price reaction:
T-1 (Friday): +0.15% average. Positive 80% of the time.
Day 0 (keynote): -0.34% average. Positive only 50% of the time.
Day +1 (Tuesday): +1.71% average. Positive 80% of the time.
The market bids Apple into the event, sells the keynote, and buys the reaction. Eight out of ten years, Day +1 is green.
Cumulative by Day +3: AAPL averages +2.09% from the pre-event window. QQQ averages +1.23%. SPY averages +0.65%.
The alpha is not in the event. It is in the reaction.
AAPL at $310.34. Friday is T-1. Monday is the keynote. Tuesday is where 10 years of data says the move happens.
$AAPL $SPY $QQQ
We flagged INGM yesterday and the day before on dark pool activity.
Today Ingram Micro presented at the William Blair Growth Stock Conference. Here is what management disclosed:
Memory supply constraints driving 2-3% revenue uplift in Q1 and Q2 2026. Cloud infrastructure-as-a-service as a key growth driver. AI-enabled PC refresh cycle with a long tail of demand. SMB AI adoption accelerating over the coming years.
Capital return: management committed to dividend growth and intends to initiate a traditional share buyback program. Platform deployment completes by end of 2027, after which opex and capex normalize.
The signal identified the positioning. The conference just confirmed the fundamental story behind it.
This is what institutional data coverage looks like.
$INGM $DRAM $MU
SPX GEX LEVELS: Jun 4
AVGO guided below expectations and dropped 12%. The gamma structure broke in a single session.
Net GEX halved from +$1.23B to +$622M. The 7,000 accelerator tripled from -$25M to -$79M, the deepest single-strike reading in the history of this series. The center of gravity collapsed from 7,743 to 7,612. And the 7,400 strike that was support for six weeks flipped to -$24M. The structural migration that carried this market from 7,400 to 7,600 just reversed direction.
Here's what changed mechanically. The zone from 7,000 to 7,475 is now entirely negative gamma. That's 475 points of accelerators below spot, the widest and deepest accelerator field we've ever recorded. The 7,000 at -$79M, 7,200 at -$35M, 7,150 at -$33M, 7,300 at -$28M, 7,100 at -$27M. If spot breaks below the flip at 7,535, there is no meaningful positive gamma support until 7,500 at +$24M, and even that is thin.
The AVGO selloff matters because it reprices the AI capex narrative. AVGO is a bellwether for AI infrastructure spending. Weak guidance from a company that supplies custom AI accelerators to hyperscalers raises the question of whether the AI investment cycle is decelerating. That's not a one-stock story. It's a sector repricing risk that applies to NVDA, AMD, MRVL, and the entire semiconductor complex.
The 7,000 strike is where the structural story gets interesting. Call OI crossed 1M for the first time, and put OI sits at 1.07M. The strike that was a distant floor at -$25M for weeks just became a -$79M gravity well. That kind of concentration doesn't appear randomly. Institutions loaded 7,000 puts aggressively today, and the gamma effect tripled. If the market ever sells toward 7,000, the accelerator at that strike would amplify the move violently.
Above spot, the magnets still function. The 7,600 max magnet at +$129M is 46 points above. The 7,700 at +$100M. The magnets above haven't deteriorated — what broke was the foundation underneath. Yesterday the structure was magnets above and magnets below. Today it's magnets above and accelerators below. The asymmetry is the sharpest it's been since the May 19-20 NVDA eve readings.
Put buying dominated today's flow. 7,500 absorbed -$66M. 7,450 absorbed -$26M. 7,400 absorbed -$26M. The entire zone below spot was loaded with new put exposure in a single session. Volume P/C hit 1.44. Institutions aren't selling. They're hedging, and the hedges they're building are deeper and closer to spot than at any point since the OPEX cliff.
The flip sits at 7,535, 19 points below the close. That's the fifth knife-edge reading. If AVGO's selloff continues into Thursday's cash session and drags semis, the flip breaks and the accelerator field fires. If the market absorbs the AVGO repricing and holds 7,535, the 7,600 magnet pulls and the structure stabilizes.
Tomorrow's expected range: 7,450 – 7,625. The widest range since the post-OPEX cliff readings. Direction depends on whether the AVGO repricing is contained to one name or spreads to the sector.
solana:J3NKxxXZcnNiMjKw9hYb2K4LUxgwB6t1FtPtQVsv3KFr $SPY $AVGO
@SyndicateTrades The near-term regime shifted, not the structural regime. Three weeks of bullish structure doesn't evaporate in one session. But it also doesn't mean one session can't change the near-term character. Both are true. $SPY
SPY UPDATE:
The structure shifted again towards the end of the trading day. Here's where we stand at the close.
GEX: -$693M. The most negative since the May 15 gamma cliff. The blanket that suppressed every selloff for three weeks flipped to amplifying. IV jumped from 13.8% to 17.4%, the largest single-session move of the cycle. The market is repricing risk.
The Iran strikes overnight broke the near-term regime. The afternoon build we tracked for three weeks didn't hold. The selling returned in the final hour and overwhelmed the midday recovery.
AVGO reported mixed earnings and is down 5% after hours. The AI supply chain was three for three (NVDA, MU, HPE). That streak is in question tonight.
The deep floor at $683 (9.1% below) is intact. The structural insurance hasn't broken. The near-term regime has.
We said this morning: "This is the turbulence, not the destination." That thesis gets tested tomorrow. The data at the open determines the next move.
$SPY $QQQ $AVGO $VIX
$SPY ALERT
Three significant shifts in 45 minutes:
Flow flipped bullish. -6.7M at 2:00 PM to +21.3M at 2:47 PM. A 28M share reversal. The afternoon build arrived. Not at last week's 80-90M intensity but the direction flipped. The institutions are buying the close.
Our Composite Score jumped 30 points. -7.0 to +23.2 Lean Bullish. The largest intraday composite recovery since the May 19 NVDA eve reversal. The flow component flipping from max bearish to max bullish drove the swing.
Dealers rebuilt 45M shares. 97.6M to 142.8M. The engine that shed a third of its capacity this morning is rebuilding into the close. The forced-buying loop is being restored.
$SPY $QQQ
AVGO just reported. We called BEAT. Here are the receipts.
Street EPS: $2.40
Alphatica EPS: $2.48
Actual EPS: $2.44 ✅
Street Revenue: $22.13B
Alphatica Revenue: $22.5B
Actual Revenue: $22.187B ✅
Both numbers beat. No revenue miss. Total revenue came in $57M above the street. EPS beat by $0.04.
The headline the market is reacting to: infrastructure software missed. Software revenue $7.18B vs $7.32B street, a $140M shortfall. Semiconductor solutions beat by $289M ($15.01B vs $14.72B). AI semiconductor revenue hit $10.8B, up 143% YoY, above the company's own forecast. The AI story delivered. The software segment did not.
Q3 guidance: $29.4B revenue. The street was at $28.25B. That is a $1.15B raise above consensus with AI semiconductor revenue guided to $16.0B, up 200% YoY. The forward trajectory just accelerated.
Stock is down 2.93%. We told you this was the risk. The June quarter is AVGO's weakest by stock reaction, ranked #4 with last 4 quarter moves of +5%, -11%, +9%, -5%. The alternating pattern continues. The BULLISH disclosure signal and the three green circles were right about the numbers. The seasonal pattern was right about the stock.
Our BULLISH Earnings Risk Signal (IV spread +2.94) had the direction wrong on the immediate reaction. The numbers beat, the guide raised massively, and the stock sold off on a segment mix miss. This is the first time our BULLISH risk signal did not produce a positive after-hours move this week. HPE +3.50 = +32%. PANW +5.74 = +10%. AVGO +2.94 = -2.93%. We own that.
The Q3 guide at $29.4B is the real number. $16B in AI semiconductor revenue growing 200% YoY. The software miss is a speed bump. The trajectory is intact. The receipts are the receipts.
Not Investment Advice.
$AVGO $NVDA $SMH
AVGO reports tonight after the close.
3/6 on our Alphatica Earnings Quality Signal. BULLISH on our disclosure signal, the strongest in our entire scoring universe. Seven straight beats. 92.5% lifetime beat rate. One of the most complete signal alignments we have published this earnings season.
Street EPS: $2.40 (Non-GAAP)
Street Revenue: $22.13B
Alphatica Estimates:
EPS: $2.48
Revenue: $22.5B
OUR CALL: BEAT ✅
Broadcom has beaten in 62 of 67 quarters. The recent cadence is small but automatic: +1.5%, +4.3%, +1.8%, +0.6%, +6.0%, +2.2%, +2.5%. Hock Tan guided Q2 revenue at approximately $22 billion, representing 47% YoY growth. The street at $22.13B is slightly above that guide. We are sizing our estimate at $2.48 EPS and $22.5B revenue, a 3% beat consistent with the recent pattern.
OUR FOUR SIGNALS:
Earnings Quality: 3/6 NEUTRAL. VMware transformed the business: revenue +92% in three years, op income nearly doubled, OCF growing to $27.5B annually. Three BULLISH flags for profitability, operating trends, and capital position. Two BEARISH: VA% reversed sharply from 30.9% to 42.3% in FY2025 with $1.77B built in a single year, same systemic pattern as GOOGL, META, AMZN, AAPL, CRM. SBC jumped to 11.9% of revenue post-VMware, absorbing VMware's SaaS compensation culture. One NEUTRAL: OCF clean at 1.19x net income but SBC elevated.
Disclosure Signal: BULLISH. The strongest signal in our entire scoring universe. Tone improved +0.011 over two years to -0.001, nearly positive territory. A level only 2-3 companies in the Russell 3000 have ever reached. Risk factor rewrites are extreme (0.329) but driven by the VMware integration narrative, not deterioration. Uncertainty dropped 53%. Litigious dropped 60%. Every sentiment metric improving simultaneously while the filing undergoes structural transformation. When a company massively rewrites its filing and every metric improves, it signals a fundamental business upgrade that management is confident about.
Prediction Model: BEAT. Seven consecutive beats. 92.5% lifetime beat rate across 67 quarters. 19 of 20 revenue beats. 20 of 20 EPS beats over the last 5 years. The June quarter is 4 out of 4 beats over the last four years. Beats are small (1-4%) but the base rate is overwhelming.
Earnings Risk Signal: +/-6.4% BULLISH. IV spread at +2.94, calls at premium. EVRP at 110.99. Expected move +/-6.4%. Avg PMIE over 3 years is 9.2% with a max of 24.4% (the December 2024 AI blowout). Options positioned for upside.
THE FLOW PICTURE:
6 of 7 dark pool triggers firing. The magnitude dwarfs recent HPE and PANW prints. $735M plus $300M in mega block prints, totaling $1B+ in two sessions. DP VWAP climbed from $423 to $483 over 10 days. 35 blocks plus 7 megas on 6/2.
The warning: today's intraday action. AVGO gapped up to $495, sold off to $473 (down 1.8% from yesterday's close), then recovered to $485. The stock is now flat on the day. Zero mega prints today after 7 yesterday. DP% spiked to 41.4% on the selloff. The flow was overwhelmingly bullish through yesterday. Today the buying stopped and the stock gave back the entire gap.
THE FORWARD VISIBILITY:
Hock Tan on the Q1 call: FY2027 revenue will be "well over $100 billion, largely driven by chip sales." AI semiconductor revenue hit $8.4B in Q1, growing 106% YoY. Six XPU customers, all strategic multi-generation programs. Anthropic deploying $20B for 1 gigawatt this year, tripling next year. Meta MTIA custom accelerator program in production. AI networking 33-40% of total AI revenue. Stacy Rasgon counted 8-9 gigawatts of deployments at $20B per gigawatt content.
Q3 consensus already at $28.25B revenue and $3.16 EPS. The forward guide tonight is what moves the stock, not the Q2 print.
THE RISK:
The June quarter is AVGO's weakest by stock reaction: ranked #4 with only +4.0% average return despite consistent beats. Last 4 quarter stock moves alternate: +5%, -11%, +9%, -5%. AVGO beats every quarter. The stock does not always go up. The intraday reversal today raises the question of whether the $1B+ in institutional flow was the last buyers before a sell-the-news event.
$482 was yesterday's close. The stock gapped to $495 this morning, sold off to $473, and sits at $485. $22.0B is the guide. $22.13B is the street. 7 straight is the streak. Three green circles and a yellow. The signals have never been more aligned on a single name. What matters tonight is the Q3 guide.
***Not Investment Advice.
$AVGO $SMH $NVDA
CRWD just reported. We called BEAT. Here are the receipts.
Street EPS: $1.07
Alphatica EPS: $1.12
Actual EPS: $1.10 ✅
Street Revenue: $1.36B
Alphatica Revenue: $1.39B
Actual Revenue: $1.39B ✅
Both numbers beat. Revenue was exact. 29 straight quarters without a miss.
Stock is down 9%.
The signals told you this was possible. Look at the card we published: two red circles and a yellow. 0/6 quality. BEARISH disclosure. NEUTRAL risk signal with an IV spread of -0.67, meaning the options market had zero bullish conviction. Compare that to the names that went up this week: HPE was +3.50 BULLISH, up 32%. PANW was +5.74 BULLISH, up 10%. When the risk signal is NEUTRAL on a stock that has run 82% in eight weeks, that is the signal.
CRWD beat on every guided metric. Revenue $1.39B. ARR $5.51B. Record net new ARR of $256M. Record operating cash flow of $591M. Record free cash flow of $468M. First GAAP profitable quarter ($27.8M net income). 4-for-1 stock split announced. FY27 net new ARR guidance raised to 27.7% growth. George Kurtz called it the AI inflection point.
The numbers were strong. The stock was priced for perfection. A 3% EPS beat on a stock up 82% in eight weeks with no bullish options lean is exactly the sell-the-news setup we described in the preview. We said last 4 quarter stock moves were +4%, +1%, +5%, -6%. Tonight is -9%. The pattern held.
The BEAT call was right. The two BEARISH signals and NEUTRAL risk were right. Both sides of the card mattered.
Not Investment Advice.
$CRWD $PANW $QQQ
CRWD reports tonight after the close.
0/6 on our Alphatica Earnings quality signal. BEARISH on our disclosure signal. 28 consecutive quarters without a miss. Up 82% in eight weeks. An institution just loaded $23.3M into the stock hours before the print on a down day. The accounting says one thing. The flow says another. The stock is about to tell us which one is right.
Street EPS: $1.07 (Non-GAAP)
Street Revenue: $1.36B
Alphatica Estimates:
EPS: $1.12
Revenue: $1.39B
OUR CALL: BEAT ✅
CrowdStrike has beaten in 27 of 29 quarters with only one miss in its entire public history. The recent cadence is compressing: +19.8%, +10.6%, +12.0%, +2.1%, +1.8%. But the June reporting period resists the compression: 4 out of 4 beats over the last four years averaging +16%, with the smallest at +4.5%. Management guided Q1 at $1.36-$1.364B revenue with non-GAAP EPS of $1.06-$1.07. The street is right at the guide ceiling on EPS and at the low end on revenue.
OUR FOUR SIGNALS:
Earnings Quality: 0/6 BEARISH. Score unchanged from original assessment. VA at 80.6% with the absolute build accelerating: +$47.6M (FY2024), +$234.7M (FY2025), +$320.3M (FY2026). Management reserved against 81 cents of every deferred tax dollar even as revenue grew 115% in three years. SBC at 22.9% of revenue. Four years of data. No release signal at any point. The gap between the revenue growth and the accounting confidence is the widest of any name in our database.
Disclosure Signal: BEARISH. Tone has deteriorated for two consecutive years, reversing the post-IPO improvement. The July 2024 global outage is the inflection point. The 10-K is quiet (LOW distance 0.038) but the latest quarterly filing spiked to HIGH distance (0.242) with an MD&A rewrite of 0.292. Quarterly tone stuck at -0.019 to -0.021, nearly double the negativity of the annual filing. Management presenting a calmer annual story than the quarterly reality supports.
Prediction Model: BEAT. 28 consecutive non-negative surprises. 93.1% lifetime beat rate. 100% beat rate over the last 8 quarters. June quarter: 4 out of 4 beats averaging +16%. The compression trend from +20% to +2% is real but the June quarter historically delivers larger beats than the recent Q3/Q4 prints. We are sizing at $1.12, a 5% beat, conservative relative to the seasonal average.
Earnings Risk Signal: +/-6.5% NEUTRAL. IV spread at -0.67, puts and calls at parity. No directional lean from the options market. EVRP at 62.09, the lowest of any name we have covered this week. Expected move +/-6.5%. Avg PMIE over 3 years is 6.56%. Last 4 quarter stock moves: +4%, +1%, +5%, -6%. This stock barely moves on earnings.
THE FLOW PICTURE:
The dark pool setup materially changed today. An institution loaded $23.3M at $750.61 at 12:49 PM, the largest CRWD dark pool print of the window. Same institution stacked two 9,580-share blocks one minute apart for $14.4M combined. DP% hit 51.7% today, the highest reading in the pre-earnings window. This is buying on a down day, hours before the print. High conviction.
Yesterday: $12.3M block at $759.91. Two sessions, $35.6M in identified institutional flow into the print.
THE OUTAGE RECOVERY:
George Kurtz on the Q4 call: 23% of Flex customers have re-Flexed (up from 5% in Q1), with average ARR lift of 26% per re-Flex. Nearly 100 customers have re-Flexed multiple times. Q1 pipeline grew 49% YoY. AI security (AIDR/Pangea) up 5x quarter over quarter. Falcon platform adoption deepening with 8+ module customers now 68% of ARR.
FY2027 ARR guidance of $6.47-$6.52B implies 22.5% net new growth, above prior commentary. Brad Zelnick flagged this as notably aggressive off an even higher base. George and Burt framed it as confidence in durable growth driven by AI security demand, Flex adoption, and platform consolidation.
THE RISK:
+82% in eight weeks. +64% YTD. The stock has more than doubled from its April low of $342.72. 0/6 quality with a BEARISH disclosure and VA build accelerating. Last 4 quarter moves are +4%, +1%, +5%, -6%. Even when CRWD beats, the stock barely moves. An 82% run into earnings with a 5% EPS beat and flat options positioning is a sell-the-news setup on paper.
The dark pool says otherwise. The question is whether $23.3M of conviction at 12:49 PM today knows more than the 82% move already priced in.
$769 was yesterday's close. The stock is at $749 today. $1.36B is the guide. $1.07 is the bar. 28 straight is the streak. The flow loaded into the dip. Tonight we find out why.
Not Investment Advice.
$CRWD $QQQ $PANW
Fair point. Cisco was the picks and shovels of the dot-com era and still fell 86% peak to trough. The infrastructure layer is not immune to sentiment. It just recovers.
As for timing, the dot-com analog is instructive. The Netscape IPO was 1995. The Nasdaq peaked in March 2000. That was five years of excess before the reckoning. ChatGPT launched November 2022. We are probably three and a half years in.
$SPY $QQQ
Ray Dalio says AI is a bubble.
He is half right.
The companies building AI infrastructure are generating real earnings, real backlog, real cash flow. Dell just posted 750% AI server revenue growth. NVDA has billions in backlog.
The bubble is not in the infrastructure. The bubble is in the companies consuming it without a business model.
Every gold rush ends the same way. The miners go broke. The people who sold them the pickaxes retire rich.
The question is not whether AI is a bubble. It is which layer of the stack you own.
$SPY $QQQ $NVDA
BREAKING: Ray Dalio just said the AI market is a bubble and it will burst.
"All great technology changes produce bubbles," Dalio told Bloomberg. "The pricking is the converting of wealth into money" right now, every major tech company is pouring hundreds of billions into AI infrastructure and booking it as investment.
The moment investors demand actual returns, companies will have to show that the money spent is generating real profits from real customers. If the revenue is not there, valuations collapse and right now, the revenue is not there.
AI companies are spending $800 billion in capital expenditure this year alone. OpenAI spends $60 billion annually on cloud infrastructure against $25 billion in actual revenue.
Less than 1% of executives globally report meaningful ROI from their AI investments. 95% of enterprise AI pilots have failed to deliver measurable returns according to MIT.
The entire $2 trillion cloud backlog held by Microsoft, Oracle, Google, and Amazon is anchored by two unprofitable companies: OpenAI and Anthropic.
By 2030, the industry needs $2 trillion in annual revenue to justify what is being built today. Bain estimates it will fall $800 billion short.
Dalio is not saying the technology is fake. He is saying the economics do not work yet and every bubble in history has ended the same way when that moment of reckoning arrived.
The companies buying pickaxes on credit are the startups raising at $10B valuations with $200M in revenue. That is where the credit risk lives.
When people say "AI bubble" they are usually describing the startup layer but pointing at the hyperscaler layer. The balance sheets are not the same.
$SPY $QQQ
Dalio is right that all great technology changes produce bubbles. He is also right that the economics do not work yet for most AI companies.
But this post is mixing up two different things.
The hyperscalers spending on AI infrastructure are not unprofitable. MSFT, GOOGL, AMZN, ORCL are printing cash from their existing businesses and allocating capex forward. That is not bubble behavior. That is R&D funded by monopoly margins.
The bubble is in the layer above them. The companies consuming the infrastructure without unit economics. OpenAI spending $60B against $25B in revenue is not an NVDA problem. It is an OpenAI problem.
When this shakes out, the picks and shovels survive. They always do. The companies renting the picks and shovels with no revenue model do not.
Dalio knows the difference. This post doesn't make it.
$SPY $QQQ $NVDA
Tomorrow in Washington. Three events that move markets.
Treasury Secretary Scott Bessent testifies before Ways and Means at 10:00 AM. With the 30-year at 4.98%, deficits widening, and the AI infrastructure boom requiring hundreds of billions in capital, every word on fiscal policy and debt issuance matters. This is the man who controls the Treasury market.
Fed Vice Chair for Supervision Michelle Bowman testifies before Financial Services alongside FDIC Chair Travis Hill at 10:00 AM. Bank regulation, capital requirements, and the state of the banking system. Last week the H.8 data showed banks funding loan growth through borrowings, not deposits. Bowman will be asked about it.
The AI hearing no one is watching: Homeland Security Committee at 10:30 AM on "The AI Security Landscape: How Frontier Models, Agentic AI, and AI Coding Tools are Reshaping Cybersecurity and Critical Infrastructure Resilience." This comes days after the White House signed the executive order giving the NSA authority to define and review frontier AI models before release.
Also on the calendar:
Richmond Fed's Barkin speaks at 8:30 AM. San Francisco Fed's Daly on a Bloomberg Technology Summit panel at 1:10 PM. Two regional Fed presidents speaking on the same day that Bessent testifies.
The defense authorization markup for FY27 is also in session. Armed Services Committee at 10:00 AM.
Tomorrow is not a quiet day in DC.
$SPY $QQQ
The S&P has averaged +1.85% in June with a 90% hit rate over the last 10 years.
July is even better. +3.37% average return with a 100% hit rate.
Seasonality says buy everything and go to the beach.
The $225B in IPOs, $500B in expiring lockups, an $80B Alphabet equity offering, and a potential USMCA auto tariff overhaul say check the forecast again.
Seasonality works until it doesn't. And it usually stops working when everyone is counting on it.
$SPY $QQQ
$SPY ALERT
Three significant shifts in 45 minutes:
Flow flipped bullish. -6.7M at 2:00 PM to +21.3M at 2:47 PM. A 28M share reversal. The afternoon build arrived. Not at last week's 80-90M intensity but the direction flipped. The institutions are buying the close.
Our Composite Score jumped 30 points. -7.0 to +23.2 Lean Bullish. The largest intraday composite recovery since the May 19 NVDA eve reversal. The flow component flipping from max bearish to max bullish drove the swing.
Dealers rebuilt 45M shares. 97.6M to 142.8M. The engine that shed a third of its capacity this morning is rebuilding into the close. The forced-buying loop is being restored.
$SPY $QQQ
🚨SPY VOLATILITY ALERT | UPDATE 2 | Wednesday June 3, 2:00 PM
$755.36. Down 0.55%. The selling is exhausting. The afternoon is stabilizing.
Our Composite Score: -7.0 [Neutral]
THE PROGRESSION:
11:45 AM: $755.26, flow -15.7M, GEX -$177M, premium 66% call-heavy.
1:30 PM: $754.52, flow -35.8M, GEX -$417M, premium 55% call-heavy.
2:00 PM: $755.36, flow -6.7M, GEX -$386M, premium 58% call-heavy.
Read the trajectory. The selling peaked at 1:30 PM. Flow improved 29M shares in thirty minutes. Price bounced 84 cents from the session low. The GEX improved $31M. The premium is recovering.
The afternoon build isn't the explosive 80-90M sessions from last week. But the selling wave crested and the buyers are showing up.
WHAT IMPROVED:
Vanna flipped supportive. Dealer effective vanna: +32.7K. Was -66.7K this morning. The mechanic that sells into IV spikes reversed as IV settled from 14.4% to 13.8%. IV is now compressing on the selloff. Dealers are starting to buy, not sell.
Flow: -6.7M. Down from -35.8M at 1:30. The bear flow is fading. Call volume overtook puts (3.22M vs 3.33M). 2.88M new call positions opened. The institutions are positioning into the close with calls, not exiting with puts.
Composite: -7.0. Improved from -10.2 this morning. Still neutral but moving toward the bullish side. The structural components (dealer delta, vanna) are recovering while the flow component drags.
WHAT HASN'T IMPROVED:
GEX: -$386M. Still negative. Still amplifying. Improved from -$417M but the regime hasn't flipped back to positive. The blanket remains thin.
Accelerators at price: $754 (-$64M), $753 (-$66M), $750 (-$61M). Still loaded below. If the close fades, $750 is the risk level.
Dealers: 97.6M shares short. Down from 143.5M this morning. The engine shed 46M shares today. The forced-selling into the negative GEX regime reduced the engine's capacity. Less firepower for the recovery.
THE SETUP FOR TONIGHT:
$AVGO reports after the close. The fourth AI infrastructure name this cycle. NVDA, MU, HPE all beat and guided up. Three for three.
The market absorbed an Iran bombing and a gamma regime flip today. It's down 0.55%, not 3%. IV moved 0.6%. The structure bent but didn't break.
If AVGO beats tonight, tomorrow's open gets the supply chain catalyst into a market that just absorbed geopolitical shock on muted volatility. The institutions who bought calls all afternoon are positioned for exactly that outcome.
If AVGO misses, the negative GEX amplifies the downside into $750 and the $740 accelerator (-$107M) becomes the next stop.
OUR READ:
The three-week afternoon build pattern didn't repeat at full strength today. The flow improved but didn't flip to the 80M+ bullish sessions that defined the last three weeks. The structure is stabilizing, not rebuilding. That's a different character.
The Iran selloff was contained. The gamma flipped. The deep floor held. But the engine is smaller and the blanket is thinner heading into AVGO and the June 18 countdown.
The bottom line:
Selling peaked at 1:30 PM. The flow is recovering. Vanna flipped supportive. The structure stabilized but didn't rebuild. The deep floor at $684 (9.5%) held. AVGO tonight is the catalyst that determines whether the selloff was one day or the start of a wider range regime.
$760 is the recovery target. $755 is at price. $750 is the risk. $740 is the accelerator. $684 is the floor. 10 days to June 18.
$SPY $QQQ $AVGO $VIX
BREAKING NEWS:
The Fed just published the Beige Book. The consumer data is the story.
Higher-income households: resilient, less sensitive to price increases.
Middle-income: "squeezing more life out of every dollar before deciding to spend it." That is a direct quote from the Federal Reserve.
Lower-income: greater financial strain. More credit card usage. Fewer retail visits. Spending shifting to necessities.
The economy is not slowing uniformly. It is splitting. Ten of 12 districts reported slight to modest growth. One reported a decline. The headline looks stable. The composition does not.
The inflation data matches what Core PCE already told you:
Prices rising at a moderate to strong pace. Most districts reported higher inflation than the prior report. Energy costs are the primary driver, spilling over into shipping, packaging, groceries, and fertilizer. Businesses have mixed ability to pass costs through.
The labor market: 11 of 12 districts reported little to no change in employment. Low-hire, low-fire. Companies are not expanding headcount but they are not cutting either.
Auto dealers saw softer new vehicle demand. Buyers substituting toward used and hybrid vehicles. That is the consumer telling you in real time where the price sensitivity is hitting.
Business outlook for the next six months: little changed in growth expectations. Elevated uncertainty. Signs of weakening labor market weighing on sentiment.
The economy is growing. The consumer is fracturing. Inflation is accelerating. The Fed sees all of it.
$SPY $QQQ
🚨🚨 DARK POOL ALERT:
14:40:19 QQQ BUY $500,734,195 NASDAQ Carte
14:39:08 IWM BUY $ 18,349,037 NASDAQ Carte
14:38:09 QQQ BUY $238,254,546 NASDAQ Carte
$QQQ $IWM