Here's a week that should have hurt more than it did. Nvidia lost the man who ran it for 33 years, lost almost 13% of its price by Friday, and the Nasdaq still closed the week 2.3% higher.

That isn't denial. The money never left AI. It changed vendors, and every vendor it moved to trades inside the same index.

First, the call I got wrong

Last Sunday I told you to sell the rallies, and that a December NQ close above 27,615.75 would prove me wrong. Monday settled at 28,014.50. Dead in one session. My own rule said buy above that line, so that's where I stand now, and you should know it before you weigh the rest.

What Thursday showed

Nvidia announced Huang's resignation on Thursday and put CFO Colette Kress in the chair as interim CEO. The stock was down as much as 14% early, then clawed back to $166.19, off about 4% on the day. In that same session its two closest rivals in the index, AMD and Broadcom, both closed higher.

By Friday the gap had only widened. Nvidia gave back nearly 13% on the week while AMD and Broadcom each added close to a fifth, and the chip index as a whole climbed about 8%.

Here's the mechanism. Key-man risk is a stock risk, not an index risk. A hyperscaler's AI budget gets signed off long before anyone asks who runs the supplier. Nobody tears up a data center plan because a CEO quit. When the supplier wobbles, the purchase order doesn't vanish. It goes shopping, and the obvious shops are AMD and Broadcom. Both sit in the Nasdaq-100. So the index is short Nvidia's key man and long his rivals in the same breath. Call it conservation of mass: the reaction moved the atoms around, it didn't destroy any.

The bear case, fairly stated

Elon Musk put it bluntly: many of Nvidia's supply deals were "a handshake with Jensen, not a written contract." HSBC cut the stock to Hold and its target from $365 to $170. Fair enough. If handshakes don't transfer, Nvidia ships less and its rivals pick up the orders it can't fill. That's a real problem for Nvidia holders. For NQ it's mostly money moving from one member to another.

Watch for the tell before the price gives it to you. As long as AMD and Broadcom trade against Nvidia, it's a vendor swap and the index shrugs. The day they start falling with it, the swap is over. That's the budget talking, and nobody in the index is hedged against the budget.

What can actually hurt the index

Nvidia reports on Tuesday, November 17, after the close, with the call at 5:00 PM ET. The quarter ended on October 25, before Huang walked out, against a guide of $108 billion, give or take 2%. The print is old news. The outlook is Kress's, the first one without him. A weak guide wouldn't be a personnel story. It'd be a spending story, and then every name on the AI financing bill falls together, with nowhere left in the index to hide.

So I buy NQ dips into Tuesday. I watch the December contract on TradingView, and the line is plain: a close below 27,516.00, the low of the November 10 session, says the market doubts the spending, not the CEO. Below it, I'm out.

Nvidia lost its CEO this week. The orders just changed address.

Everything you need, the weekly futures signals included, sits in one place: https://whop.com/gekkquant/.

This post contains 1 affiliate link.

Subscribe now

So why do I still trade Larry Connors and Cesar Alvarez's RSI(2) pullback? One word: drawdown. Here's the case, in numbers you can check.

The rules

They come from their 2008 book, Short Term Trading Strategies That Work, and I run them on SPY's daily chart exactly as written:

• Trend first. I only play while SPY closes above its 200-day average, because a pullback in an uptrend is a gift and a pullback in a downtrend is a trap.

• Buy the panic. When the 2-day RSI closes under 10, I buy the close.

• Sell the bounce. The first day SPY closes back above its 5-day average, I'm out at the close.

• Nothing else. No stop, long only, and the rest of the time I sit in cash.

That last line does the heavy lifting: I'm in the market only about 1 trading day in 8.

The money

I gave both sides the same $10,000 on Nov 15, 2021 and charged costs on every fill. By Nov 13, 2026, SPY had turned it into a little over $16,000, while my system stopped near $13,500.

It wins about 3 trades in 4, but the winners are small, so my line climbs in steps while SPY sprints.

The drawdown

This is the chart I actually run my money off: how far below my last high I was, every day.

From Jan 3–Oct 12, 2022, SPY bled out a full quarter of its value. My worst hole in the entire 5 years stayed under 6%, and it came that January.

The reason is boring, and boring is the point: SPY spent most of 2022 under its 200-day average, so the filter kept me in cash.

And the part nobody prints: SPY needed 746 days, until Jan 19, 2024, just to get back to its old high.

Year by year, SPY takes most of them, but in 2022 it lost almost a fifth while I gave up about 3%.

The recovery math

Losses don't climb back symmetrically. SPY's 25.4% hole needed a 34% rally just to break even, while my worst one needed barely 6%. A shallow hole is easier to sit through, and the best strategy is the one you don't dump at the bottom.

The last 3 years

The most recent stretch, Nov 14, 2023–Nov 13, 2026, tells the same story.

SPY's worst came in spring 2025, a 19% slide in about 7 weeks into the Apr 8, 2025 low. My worst over the whole 3 years was under 4%, and the $10,000 still grew to $13,263. SPY finished well ahead again.

What it costs me

• Return. I gave up about half of SPY's 5-year gain, and that's real money.

• No stop. A trade can keep running against me; my worst single one cost 3.2%, in January 2022.

• Boredom. Around 10 trades a year, and the rest sits in cash that earned nothing in this backtest.

The honest caveats

I won't sell you an edge I can't prove. The shallow hole is mostly exposure: I sat in cash about 7 trading days in 8, so I missed most of the crash, and a small fixed slug of SPY you never touch is a legit alternative. I prefer a rule I can stick with, and that's a preference, not proof of an edge.

Bottom line

I'm not trying to beat SPY. I'm trying never to be 25% down.

Over 5 years I gave up roughly half the return to keep my worst drop at about a fifth of SPY's. If you can sleep through a quarter of your account vanishing, just hold SPY. I can't, so this is what I do instead.

Subscribe now

Backtest, not a promise. Informational, not investment advice. Size and manage your own risk.

Data: SPY daily closes from Yahoo Finance, Nov 15, 2021–Nov 13, 2026, no dividends, 0.03% cost per side, cash earns 0%. Skipping dividends understates SPY by about 1.3% a year. Paying cash nothing understates the system more: I sat in cash about 7 trading days in 8, and T-bills paid roughly 4–5% a year in 2023–2025, and about 3.6–4% so far in 2026. If anything, the strategy looks better than shown. Win rate after costs.

Energy gets a seat at the table:

• ExxonMobil (XOM): 7% of the book in at $169.82.
  Buying the energy the economy needs, not another promise priced for perfection. Keep cash in reserve; opportunity doesn't send an invitation.

Broad tech and financial exposure needs a counterweight, so ExxonMobil comes in as the integrated-energy leg. Revenue up 32.6% and 13.5 times forward earnings is a cheap price for real cash flow, and oil's a jumpy input, which is why the size stays modest.

Positions and the race against the S&P500: gekkquant.com/gq-stockerator.

Free for the first 6 months.

Subscribe now

Cash off the sidelines:

• Berkshire Hathaway (BRK-B): 18% of the book in at $590.00.
  Balance market excitement with businesses built to keep earning.

• Microsoft (MSFT): 16% of the book in at $523.65.
  Want enterprise customers and recurring revenue working for me.

• State Street SPDR S&P 500 ETF Trust (SPY): 15% of the book in at $758.64.
  Put idle capital to work across American business.

• Alphabet (GOOGL): 12% of the book in at $300.10.
  Buy distribution that businesses depend on.

• Amazon (AMZN): 11% of the book in at $269.10.
  Want a stake in the infrastructure behind shopping and computing.

• JPMorgan Chase (JPM): 10% of the book in at $343.89.
  Want exposure to the institution financing the deal, not just the company pitching it.

Cash earns nothing sitting there, so the index goes in as the base — no pretending every idea has to be a stock pick. Microsoft and Alphabet give me enterprise software, cloud and ads, and I'd rather not stack a second bet on chips. Amazon adds commerce and cloud. Berkshire is the insurance and operating-business counterweight under 20%, JPMorgan widens earnings past tech. Roughly 13% cash stays back as dry powder while the leadership news shakes out.

Positions and the race against the S&P500: gekkquant.com/gq-stockerator.

Free for the first 6 months.

Subscribe now

First move on the book: 12% into NVIDIA at $173.73.

The crowd is selling fear; buying compute. Every data center on earth is a customer, and the line is out the door.

Most of the book is still cash, so there is room for the next idea. Positions and the race against the S&P500: gekkquant.com/gq-stockerator.

Free for the first 6 months.

Subscribe now

A track record you can't backfill is the only kind worth reading. So I'm building this one in the open.

GQ Stockerator is my machine for US stocks: a book that started with $10,000, run against the S&P500 since Nov 12, 2026. Its first position is NVIDIA.

The scoreboard is nearly empty for now. Save the link anyway. The engine went through extensive testing before it went live, and it trades on the same strict rules as my ladder. I expect it to be quite profitable.

• Every buy: a short post the day it happens, with the reason.

• Every winning close: the same, plus the profit.

• The rules: US stocks and ETFs only, no shorts, options or leverage, at most 20% of the book in one name.

Free for the first 6 months; after that, the picks move to the paid tier.

Subscribe now

Not investment advice.

You'll hear plenty of dip-buying pitches on NQ this morning, so here's the other side of the desk: I'm not trading it this week. December NQ has bled 5,250 points since its Oct 6 high, the last 4 weeks all closed red, and nothing on my chart says the bleeding has stopped.

The chart says short, so the long side waits

Friday's close sat under the weekly 50 EMA and under the July low. That's a short read, and on a short read I don't go hunting for a bottom. I let the market prove there is one.

The tape underneath is no friendlier. The Democratic sweep got sold as a policy story and turned into an AI unwind within days, the same crowded-book flush I took apart last week. Now Nvidia's board is investigating reports that Jensen Huang was a spying target. So the AI leader that could haul NQ back up walks into this week carrying its own headline risk.

Tuesday is a coin I won't flip

October CPI lands Tuesday at 7:30 AM CT, and the consensus wants a month at half of September's pace. After Friday's jobs beat, the market already leans hard toward a December hike, so a hot print feeds the exact bet that's been hurting NQ.

Buying the knife the night before that print isn't conviction. It's a lottery ticket with a stop attached. My screens stay on Tuesday morning; my order ticket stays shut.

Here's what changes my mind: a daily settle above 28,180, the weekly 50 EMA on my back-adjusted chart. Above it, I'm long again. Below it, no trade means exactly that: nothing to fill, nothing at risk.

Flat is a position too. This week it's the only one I'll hold.

Everything you need, the weekly futures signals included, sits in one place: https://whop.com/gekkquant/.

Subscribe now

Last Sunday I put my exit in writing: if December NQ closed Friday below 28,232.75, the market had priced something bigger than hearings, and I'd stop arguing with it. It closed well under it. So I've stopped. The useful part is why I was wrong, because the reason isn't the one on your screen.

Thursday's 6.36% drop in the Nasdaq-100 wasn't fear arriving. It was calm leaving. The vote was a day old when Thursday opened. The selling came from crowded books that had just loaded up again.

The calm was the setup

Look at the VIX. On election day it closed at 20.83. On Wednesday, with Democrats holding both chambers and the result priced, it fell to 16.61. On Thursday it closed at 28.18.

That Wednesday drop is the whole story. When implied volatility loses a fifth in a day, every risk model on the Street hears the same thing: the event is gone, you can carry more. Goldman's John Flood said it plainly in a note: funds that had de-risked hard into the vote chased Wednesday's relief rally and got caught wrong-footed a day later. I liked that Wednesday tape too. I'll own that.

Think of a supersaturated solution. It looks clear and perfectly still, and one seed crystal turns the whole flask solid. The calm isn't the opposite of the crash. It's the precondition.

Same trade, same door

Here's the number that stopped me. Tech-focused long-short hedge funds were up 24% at their June peak, by Goldman's estimates. Thursday they fell more than 4% in a single session, and that left them roughly flat for the year. Half a year of gains, finished off in one session.

That's why the tape looked so ugly without one earnings miss to blame. A pod manager's stop isn't a price, it's a drawdown limit. Hit it and the risk desk cuts the book, whatever the story says. And the books look alike: PivotalPath flagged a high and rising overlap among the 10 largest multi-strategy funds, so their exits run through the same names. Walleye, a $17 billion fund, told its investors on Thursday it would hand back part of the capital.

What I do now

Friday's bounce came with a jobs beat: 165,000 payrolls against 65,000 expected, and traders pricing about an 80% chance of a December hike. A rally that has to climb over a rate hike isn't the forced sellers finishing. Platforms cut leverage over weeks, not in one afternoon, and the people who run them expect more cuts if prices keep sliding. This is the AI financing bill meeting the money that was levered to it.

Keep the mechanism, because it'll come back. After a binary event, implied volatility collapses, and everything sized on volatility gets bigger at once: the pods, the vol-targeting money, your own position if you size off the VIX. The size goes up exactly when the event's real consequence hasn't traded yet. The next time the VIX loses a fifth on an all-clear, cut your size. Don't add to it.

So I sell rallies into Thursday's gap; I don't buy them. Here's what proves me wrong: a December NQ close above 27,615.75, the low of Wednesday's session. Above it, the forced selling is done, and I'll buy with you.

The crash didn't start on Thursday. It started on Wednesday, when the VIX told everyone it was safe.

Everything you need, the weekly futures signals included, sits in one place: https://whop.com/gekkquant/.

Subscribe now

Before you pay up for protection into Tuesday night, take a hard look at what actually knocked your Nasdaq over this week. The Nasdaq-100 dropped about 3.5% on Friday alone, and not a single vote had been counted.

Here's my call, and I'll own it: the midterm is the most over-hedged event on this quarter's calendar. A Democratic House buys subpoenas and hearings, not the earnings line. What's repricing the index is the AI funding bill.

The bear case, fairly stated

Quoth the Raven and Capital Flows argue that a Democratic House would hand subpoena power to people keen to dig into crypto, private credit and the AI boom. Fair enough: a chairman can make a CEO's spring miserable. But a gavel can't pass a law alone, the president keeps his veto pen, and a hearing makes a headline, not a margin. What holds up these multiples is the growth path of AI revenue and the rate you discount it at, and neither sits in a committee room.

The experiment we have already run

In 2019 the Democratic House investigated everything that moved and impeached the president a week before Christmas. The Nasdaq Composite finished that year up about 35%. In December 2018, right after the Democrats won the House, the S&P500 slid about 9%, but the new majority wasn't even sworn in. The Fed hiked again on December 19. The vote got the blame; the price of money did the damage.

Measure the S&P500 from the day before each midterm to the same date a year later: 19 midterms since 1950, 19 green years, whichever party won. Before a vote, money pays for uncertainty. Once the result is in, that premium gets handed back. Gridlock finishes the job: 2 years in which Washington can't do much of anything are 2 years in which it can't do much damage either.

What really moves the price

Think of an election as a catalyst, not a reagent. It changes how fast a reaction gets where it was going, never where it ends. Where this one ends is set by 2 things no Congress can vote on: what the AI money earns, and what it costs to borrow it.

This week showed it. Alphabet posted its biggest single-day percentage drop on record on Thursday, Meta lined up a record $97.5 billion raise, and Microsoft is reportedly heading back to the US bond market for the first time in more than 9 years. Treasuries rallied while tech fell: a growth scare, not a political one. That's the AI bill coming due in public, the same bill the index has been booking as somebody else's profit.

And nobody's walking into Tuesday blind: the street's base case is already a narrow Democratic win and gridlock. When everybody leans the same way, the surprise is how little the expected result changes.

So I don't hedge the vote; I trade it. Size down for the thin overnight session, and if NQ gets washed out Wednesday morning, I'd rather buy that panic than sell it. The dates that matter come after: Nvidia's results in mid-November and every Fed speaker on the calendar.

Here's where I'm wrong: if the December NQ contract closes on Friday, November 6, below its October 30 low of 28,232.75, the market priced something bigger than hearings, and I'll stop arguing with it.

Voters pick the committee chairs. Earnings pick the price.

Everything you need, the weekly futures signals included, sits in one place: https://whop.com/gekkquant/.

Subscribe now

The October card, from the review of Oct 31, 2026. 7 moves across 1 portfolio, 7 portfolios stay put. Here is the playbook, idea by idea.

New to the ladder? Here is how all 8 portfolios work: the ladder, explained

No change: Bedrock · Cash Barons · Flywheel · Sector Sniper · Silicon Empire · Titans · Wolf Pack. No rule fired, so nothing moves. Boredom is a position.

All 8 portfolios, equal weight: −4.1% in October, in US dollars.

Equal weight: the money is split evenly across all 8 portfolios and evened out again every month. $10,000 invested on Sep 30, 2026 was $9,591 on Oct 31, 2026.

Check every portfolio at the source: Wikifolio.com/en/int/p/gekkquant

Picks are informational, not investment advice. Size and manage your own risk.

Read more

The latest review: the October ladder, move by move

At heart I'm a futures trader: that's my short clock, where a day is a long time. But nobody builds wealth on a short clock alone. Over the years, the money is made by owning businesses and letting them compound, and that's my 2nd cornerstone: this ladder of stock portfolios. Futures for the short game, stocks for the compounding, 2 cornerstones, 1 discipline.

You don't need another opinion. You need a system that tells you what to own, why you own it and when to let it go, and then shows you the bill when it's wrong. That's what I built. Substack has no shortage of brilliant portfolio managers, so here's why mine deserves your attention, in reasons you can check, not adjectives:

• The score isn't mine to keep. All 8 of my portfolios are public on Wikifolio, and Wikifolio prints the result, not me. No screenshots, no "trust me", and the losers stay on the chart right next to the winners.

• Rules make the call, not my mood. Every portfolio runs on a written mandate, and once a month each position has to prove, against those very words, that it still belongs there. Most portfolios drift away from their own pitch; mine gets audited against it.

• 8 engines, not 1 bet. Momentum that lives in days, chip leaders on a swing, quality held through the cycle, cash flow bought cheap. Different drivers, different clocks, different exits.

• Every move comes with its reason. Not a ticker dump: each buy and each sell carries the why, so you learn the logic, not just the trade. Paid subscribers see it before anyone else.

If you've been waiting for someone to run money in the open, by rules, with every reason on the table, stop waiting. Read on: here's the whole machine on 2 pages.

The ladder

Why a ladder? Because patience and speed never live in the same portfolio. The bottom rungs are built to sit through a storm; the top rungs are built to be wrong fast and cheap. The higher the rung, the faster the money and the sharper the risk. You climb on purpose, never by accident.

Don't mistake this for 8 copies of the same idea wearing different names. When one engine stalls, the others aren't burning the same fuel. That's not decoration. That's risk control.

The 8 portfolios

The monthly review

The review is where the system earns its keep. On the last Saturday of every month I pull the live portfolios and put every sentence of every mandate on trial against what the portfolio actually holds. There's no "your call" and no "let's see": every keep, trim or kill is a verdict with its reason. A new name clears every gate, the last one live in Wikifolio's own buy screen, or the seat stays empty.

The scoreboard

Here's the bill, straight from Wikifolio's own index to the September close. 7 of the 8 are green since launch, 1 is red, and the red stays on the chart. Show me a manager who hides his losers, and I'll show you winners you can't trust. The 3 oldest have run since early 2020; the other 5 joined between July and September 2025.

Since October 2025 all 8 have run side by side. Held in equal weight, that basket was green in 8 of those 12 months and up roughly 34% on the year in US dollars. July's drawdown stays on the chart, because I don't airbrush.

The next review lands on Saturday, Oct 31, 2026. Paid subscribers get the whole card that evening: every move and the reason behind it. The scoreboard stays free for everyone, win or lose. That's the deal.

Subscribe now

Check every portfolio at the source: Wikifolio.com/en/int/p/gekkquant

Picks are informational, not investment advice. Size and manage your own risk.

Win, and it paid early.

Monday filled the bid and printed 29,983.50 in the same session for TP1, and Wednesday tagged 30,113 inside the pay zone: +592 points cashed.

Thursday then settled 29,142, under the out, on a book that was already paid.

Re-read last week's Take →

THE SETUP: LONG

Last week ran 886 points from Monday's open to Wednesday's 30,113, then Thursday gave back 951 in 1 session and left the week 274 points under the prior settle.

Friday printed 28,987, the lowest in 8 weeks, and 2 sessions dipped as much as 66 points under the September 16 low without a single daily settle beneath it.

This morning held 37 points above Friday's low and bounced 297 by 09:35 CT.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

Wednesday ran through the sell to 25,833 and settled at 25,769, over the 25,722 line, so the short was dead for 454 points and there is nothing to argue.

Friday then took the tape to 25,119, through the whole pay zone, and the week settled at 25,261: the right read, paid to nobody.

Last week's Take →

THE SETUP: SHORT

Last week ran to 25,833 and settled 457 points red at 25,261, 118 under the prior week's low.

October has run at 25,830 on the 16th and the 21st, and both times the next session opened lower and handed back more than 400 points.

Monday has bought 305 points off a 25,251 low into 11:05 CET, topping at 25,556, 4 points shy of the 61.8% retrace of last week's 714-point drop.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

Filled Wednesday and wrong: the late-September box I read as a floor broke that same session, and the trade cost 952 points.

Last week's Take →

THE SETUP: LONG

Last week dumped 1,754 points open to settle, 5 straight down days that handed back the whole late-September breakout and dropped NQ inside the summer box.

This morning opened at 29,227, 2.5 points from the September 2 low that floors that box, and ran 643 points off its low by 09:35 CT.

Thursday had already shown where the buyers live, taking 521 points back off its low in 1 session.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

It was a loss: the sell filled on Tuesday, the drop stalled at 25,237, 59 points short of TP1, and Friday settled 121 points through the out for minus 340 points, so the map was simply wrong.

Last week's Take →

THE SETUP: SHORT

Last week ended a run of 6 straight lower weekly highs: Friday ran from 25,538 to 25,827, and the week settled at 25,755, 422 points over its open.

Monday handed Friday back by 11:05 CET, dropping 282 points from a 25,718 open to a 25,436 low, and it trades at 25,455, 83 points under Friday's low.

This week's average money sits at 25,538, on Friday's low to the point, so every contract bought in Friday's 289-point range is underwater.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

My read, and I'll own it: Kevin Warsh didn't abolish forward guidance in June. He decentralized it. The statement went quiet, and the guidance moved into speeches, minutes and data prints, where it shows up late, uncoordinated and in pieces. That isn't less volatility. It's the same volatility, spread over more days.

Start with the regime change. At Warsh's first meeting as chair, on June 17, the Fed held the funds rate at 3.50% to 3.75% and cut the forward guidance out of the statement. No more sentences about the path. No more promises. In principle, I've got no quarrel with that. A central bank sitting well above zero, with inflation running hot, has no business pre-committing to anything.

But a market doesn't stop needing a path because the Fed stops drawing one. It draws its own. And it draws it from whatever the Fed hands it next.

Watch what it grabbed first. On August 28, at Jackson Hole, Warsh said the Fed would "have work to do" if policymakers weren't confident inflation was heading back to 2%. No decision, no vote, a single conditional sentence. The 2-year jumped from 4.20% to 4.34% at the close. That's 14 basis points for a hint, against 7 for the actual hike 3 weeks later. I wrote the week of that speech that nobody at the Fed controls this market on their own. The tape has spent every week since proving it.

Then came the round trip that should've ended the debate. On Monday, September 28, the 2-year closed at 4.92%, and fed funds futures priced about a 70% chance of another hike on October 28. On Tuesday, New York Fed President John Williams told an audience in Buffalo there's "no need for urgency, and we have time to gather more information." On Thursday, Jefferson said, "My colleagues and I will need to come to our own judgment, which may take more time." By Thursday, the October odds were down to roughly a quarter.

From 70 to 25 in 4 days. No meeting in between. Just 2 speeches, with the August PCE inflation print landing between them on Wednesday.

That's forward guidance. Whoever holds a microphone that week delivers it now, instead of the committee speaking together on a single afternoon.

The minutes made the gap plain. The September dots had already told the market that most officials expected another hike before year end. What the dots couldn't say was why. The minutes did, on October 7, 3 weeks after the hike: many participants backed the move on risk-management grounds, as insurance against inflation that won't come down, and most saw another increase as likely appropriate by year end. Insurance or forecast is the whole question for the next meeting, and under Warsh the answer reached the market 21 days late, after Fed officials had already talked the price up and back down. The next day, Governor Christopher Waller flew the point to Istanbul: policymakers can still signal the range of choices they're weighing, he argued, without locking in a path. When a governor has to cross an ocean to explain how the Fed still signals, the signal clearly went somewhere.

Now put the data on top. September's CPI landed on October 14: 3.7% over 12 months, up from 3.4% in August. Core, without food and energy, ran at 2.6%. That gap is energy, and a rate hike doesn't produce diesel. The October odds still jumped from about 22% the day before to better than a third the same day, because with no guidance in the statement, the latest print is the guidance. On October 16, in Bangkok, Warsh put it in a single line: "I don't believe in saying now what I might do at the meetings." The hike bets faded again while he spoke.

Now the part nobody seems to have priced. Until February 4, 1994, the Fed didn't announce its decisions at all. Traders watched the New York Fed's trading desk and inferred from its open market operations whether policy had changed. That's 32 years of building the talking central bank, from that first statement to dot plots and press conferences, and in June the statement went back to saying nothing about tomorrow. And on October 12, Stockholm handed the Nobel to Robert Hall and Michael Woodford, whose models gave central banks the theoretical base they still run on. Woodford's 2012 Jackson Hole paper is the playbook for steering a market with words when the rate can't move: forward guidance at the lower bound.

So the prize for building the machinery of guidance landed 4 months after the Fed switched the machinery off. And 4 days before the chair told the world he doesn't do previews.

The mechanism is simple, and the chemist in me recognizes it. Forward guidance worked like a buffer in a solution. New information still came in, but the buffer absorbed most of it, so the price didn't jump with every drop. Pull the buffer, and every drop moves the reading. A campus speech, a footnote in the minutes, a hot CPI line: each one now has to do the job the statement used to do, and each one does it with a bigger swing.

For anyone trading index futures, this changes where the risk sits. The 2-year is the market's running bet on the next 2 years of Fed policy, and NQ is a long-duration asset: its value leans on earnings years out, so it reprices hard when that bet moves. The old event was the statement at 1:00 p.m. CT on 8 Wednesdays a year. The new events are every scheduled Fed speaker and the 7:30 a.m. CT data prints. That's why the pre-market hour matters more to an NQ session right now than the FOMC statement does. The repricing hits before the cash open, and the open inherits it. (Yes, I now read the Fed's speaking calendar on Sunday night. It's become the most useful page the Fed publishes.)

The fix doesn't need a single promise about the path. It needs the reason, on decision day. An insurance hike invites a pause. A forecast hike invites the next one. A few dozen words in the statement would've told the market which kind it just got. Leave them out, and the market has to reverse-engineer the answer from whichever official speaks next, which is exactly the 70-to-25 round trip we just watched.

None of this means Warsh is wrong to want a quieter Fed. Woodford built his own case for heavy guidance around a central bank stuck at zero, unable to move the rate. This Fed can move the rate, and it just did. The mistake is thinking silence removes the signal. It doesn't. It removes the editor.

My call for October 27–28: no hike. The statement stays short, Warsh stays vague, and the move comes on December 8–9, the meeting that carries the projections. Williams already said "late this year." The minutes already said "by year end." That's 2 Fed sources pointing at December, delivered by the exact back channel this regime runs on.

What proves me wrong? The 2-year finished the week around 4.7%. If it closes back above 4.92% before the meeting, the level it hit when the market priced October at about 70%, the market is telling me October is live, and I'll own the miss. Below that, I'm treating every Fed speech between now and the 28th as the real meeting, and the meeting itself as the formality.

The Fed didn't go quiet. It stopped speaking as a committee. And a market without a script prices every voice it can hear.

— Gekkquant · My trade signals live on Whop: https://whop.com/gekkquant/

Subscribe now

The bid filled, and the week paid.

Thursday ran 108 points through the out line to 30,792 and settled 69.5 points above it, and Friday's close marked the long +27 points.

A stop on prints would have donated that position; the settle rule kept it, and kept it green.

Re-read last week's Take →

THE SETUP: LONG

The weekly lows have stepped up 3 weeks running, 29,053 to 30,792, but the gains are shrinking: 972 points, then 172.5, then 47, and last week's record at 31,616.5 left a 508-point wick over the settle.

Thursday's 674-point slide was bought 317 points into Friday.

This morning took the October 2 high by under 7 points and dropped 432 straight back into the late-September box, 218.75 under Friday's settle at 09:35 CT.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

Tuesday settled at 25,612, 83 points over my exit, and the short died there for a 226-point loss that is mine to own.

Thursday then printed 24,925, under the whole pay zone, without me.

Last week's Take →

THE SETUP: SHORT

Last week printed the 6th straight lower weekly high and lower weekly low, 748 points from Tuesday's top to Thursday's floor, and settled 202 points under its open.

Wednesday opened on its high and never looked back, 332 points gone in 1 session, and Thursday added the low.

The bounce since has bought back 484 points into Monday's 25,409, and at 11:05 CET price is stalling 11 points under the week's average money.

Every settled week is graded in public: https://GekkQuant.com/futures-weekly-results/

Read more

Here's my read, up front: the $20 billion is a counting rule, not lost demand. And the fight over it hides the number that should actually worry anyone trading NQ. That number isn't revenue. It's the bill.

Start with the rule. Anthropic counts the full value of the sales of its models made through Amazon's and Google's clouds, and pays the cloud its cut out of that. OpenAI counts only its own cut. Anthropic hands its cloud partners about 16 cents of every dollar sold that way, and those sales made up half of its revenue last year. That's the whole gap. The near-$70 billion was OpenAI grossed up to sit next to Anthropic on the same ruler. The $50 billion is OpenAI on its own.

Same customers, same tokens, same servers humming in the same data halls. Two numbers, both honest, $20 billion apart.

Neither of them is cooking anything, either. US accounting rules let a seller book the gross price when it controls the service before the customer gets it, and force it to book only its net fee when it's acting as somebody else's agent. Where a model sold through a cloud marketplace lands on that line is a judgment call, and the two labs made opposite calls. That's a footnote dispute, not a demand story.

And both rest on a run rate, which is often one month's revenue times 12. It's unaudited. It comes from two private companies, shared with investors in the middle of fundraising. It isn't a 10-Q, and nobody signs it under penalty of anything. Yet a change of ruler on that number knocked more than 3% off the semiconductor index in a single session. If you want to know how soft the ground under the AI equity trade is, that's your answer. The market prices the stocks off numbers nobody audits.

I spent Thursday evening with revenue-recognition footnotes instead of the closing auction, which tells you the kind of week it was. The footnotes settle one thing fast: on the revenue side you can always pick another ruler. On the cost side you can't.

In November 2025, Sam Altman put OpenAI's commitments at about $1.4 trillion over the next 8 years, mostly data centers and chips. In the same breath he said the company would end 2025 above a $20 billion run rate. Spread that bill evenly and it comes to $175 billion a year.

Now hold September's revenue up against it. On the net ruler, $50 billion. On the gross one, near $70 billion. Either way, the hole to the annual bill runs $105 billion to $125 billion. The market spent Thursday fighting over $20 billion. The gap that matters is five to six times bigger, and no change of accounting closes it.

That bill doesn't float in space either. It lands on balance sheets that sit in the indices. Oracle's remaining performance obligations, contracted revenue it hasn't booked yet, reached $664 billion at the end of August. Next to it sits OpenAI's own figure for the capacity Oracle is building for it: a partnership worth some $300 billion over five years. A contract is only as good as the buyer's ability to pay, and that buyer just reminded everyone its revenue depends on which ruler you use.

Follow that dollar one more step and you're inside the futures. OpenAI's commitment becomes Oracle's build-out, the build-out becomes chip orders, and the chipmakers at the end of that chain, Nvidia, Broadcom and AMD among them, all sit in the Nasdaq-100 that NQ tracks. That's how a private company's spreadsheet ends up in your P&L on a Thursday afternoon.

Last week I laid out how the index books the AI build-out as profit now and as cost later. This is the same machine seen from the other end. Equity holders get paid on the soft number. Lenders get paid on the hard one.

If you want proof the hard number gets priced, look at Meta's Hyperion campus in Louisiana. In October 2025 Meta put it into a joint venture owned 80% by Blue Owl funds and 20% by Meta. The Blue Owl side, a vehicle called Beignet Investor, sold about $27 billion of bonds due 2049 with a 6.58% coupon. Meta pays rent, the rent pays the bondholders, and Meta's own filings treat the venture as unconsolidated, so the debt isn't on Meta's balance sheet.

Look at what the market charged for that wrapper. S&P rated the notes A+, one notch under Meta's own AA-, and they came at about a point more than the roughly 5.5% yield on Meta's comparable corporate bond. Same tenant, same cash flow, a full point extra for keeping the debt out of sight.

Then the price ran the other way. By late September the notes, sold at par, were trading around 91 cents on the dollar, with the yield as high as 7.55%, about 230 basis points over Treasuries. That repricing came before anybody had heard of the $50 billion. The bond desk marked down the AI bill more than a week before the equity desk panicked over a definition. Borrowing on that scale is part of why the long end refuses to come down, Fed or no Fed.

That's the order these things run in. Credit leaks first, quietly, at the back of the book. The stocks find out later, usually on a headline that's technically about something else. Thursday was that headline.

Thursday sold the wrong number for the right reason.

The wrong number: the $20 billion will mostly turn into a footnote. According to Bloomberg, OpenAI is still telling investors it expects to reach or exceed $70 billion of annualized revenue by the end of the year, on its own net basis. Fine. Let them.

The right reason: the question for NQ was never whether OpenAI is a $50 billion company or a $70 billion one. It's whether a $175-billion-a-year bill gets paid out of revenue or out of fresh money, debt and equity raised against the promise, and right now it's fresh money. As long as the lenders keep asking more for the same tenant, the stocks that sit at the front of that queue carry the risk, whatever the run rate says.

So here's my side. I'd rather own the payers, the companies funding the build-out from their own cash flow, than the builders whose backlog leans on one startup's promise. On NQ, I sell the rallies that get bought on "it was only accounting" until the credit side stops leaking.

Where I'm wrong, in two numbers. OpenAI prints $70 billion or more on its own net ruler by December 31, which takes 40% growth in a single quarter. And the Hyperion notes climb back above 95 cents. If both happen, the bill is getting paid out of revenue, and I'll say so right here.

Until then, keep this straight: revenue is whatever ruler you pick. Rent is due on the first.

— Gekkquant · My trade signals live on Whop: https://whop.com/gekkquant/

Subscribe now

Subscribe now

The full card — bias, levels, catalysts, calendar: NQ (Nasdaq) Daily Pulse → https://www.nq815.com/pulse/2026-10-08.html