Is this the peak of the “AI won’t pay for itself!” narrative and a bottom for stocks?
Hello. It’s Friday. Thanks for signing up. I’m Brent Donnelly.
The About Page for Friday Speedrun is here.
friday speedrun
Cracked 2
Here’s what you need to know about markets and macro this week
Global Macro
Before we start: Please complete this quick 6-question survey about the current state of markets and the AI boom/bubble. Should take 39 seconds. I will publish the results on Twitter on Monday. Thanks!
Click here to complete the survey
There has not been all that much global macro to chew on of late because the US shutdown blocked the flow of data and nobody really knows what’s going on in the US economy. The story has been one of changing sentiment as META’s earnings on October 29 marked the turning point where markets decided more AI Capex can be bad instead of good.
The disbelief around Sam Altman’s increasingly ludicrous promises has built to a crescendo now as investment banks downgrade Oracle’s debt, Chanos cries foul, and CRWV and NBIS both tossed up bricks at earnings time. The low-margin and heavily debt-funded nature of the AI capex from Oracle, CoreWeave and NBIS has become a concern. Partly a case of “it matters when it matters” but also partly a case of these companies showing disappointing revenues and margins despite continuing promises written on napkins and assurances that it’s all gonna be fine by 2030.
The gap higher in ORCL after the OpenAI MOU was broadly derided and mocked, but at the time it was hard to tell if that was sour grapes from shorts or just rational people pointing out how stupid the whole thing was.
The whole thing was, indeed, stupid.
I was on the down move in stocks with great alacrity but sadly around mid-week this week I turned bullish as I felt that the AI revulsion trade had become mainstream. We were talking about the ORCL situation being a joke in mid-October and now it’s mid-November and investment banks are finally downgrading Oracle’s debt in response to rising CDS levels. Twitter is alight with anti-AI malice as most of the poopcos like RGTI, MSTR, and so on have dropped 50% or more from the highs. I felt like it was a good chance to buy the dip. That was early/wrong.
Since Wednesday, we’ve seen another huge leg lower and a lot of my gains on shorts have been blasted by these badly timed longs. This is trading. There are 52 weeks in a year, and one of those weeks is going to be your worst week of the year. This week is mine.
The US data should come back onstream next week, but the BLS is still playing coy and haven’t quite had time to hit SHIFT-F9 on nfp.xls or cpi.xls.
Elsewhere, British markets are bobbing and weaving around UK headlines as everyone tries to gameplan the 26NOV UK Budget. It’s a bit of a damned if you do and damned if you don’t situation as the UK has two choices:
- Choose austerity and a weaker economy and political pain.
- Choose no austerity and a bond market implosion and market pain.
The structural sovereign debt issues are pretty much impossible to solve without a gut-wrenching Reset, and so policymakers will keep making whatever moves possible to keep kicking the can past the bond vigilantes. The moves today in gilts and GBP are nothing like the ones in 2022, but they are concerning given the lack of good policy options.
It’s that time of year again.
The fourth annual handbook has arrived…
The 2026 Spectra Markets Trader Handbook and Almanac is live
Use The Spectra Markets Trader Handbook and Almanac as a trading journal/day timer, and as your guide to seasonal patterns and key economic events throughout the year. It will anchor your process on proper journaling and planning—and help you thoughtfully assess your performance at the end of each month. Goals and plans written down are much more likely to come true than a flurry of ideas and thoughts and hopes swirling around in your head.
Stocks
The market went from all-in bullish every single momentum trade to fully bearish embracing a view that the AI bubble has made its ultimate top—in four weeks. RGTI and OKLO and the other unprofitable (zero revenue!) companies peaked 15OCT (along with two other momo and retail favorites, gold and silver) and the NASDAQ peaked on 29OCT. Sandwiched in the middle was one last attempt by WSB to “corner” a stock: BYND on 22OCT.
Today smacks a bit of a final swoon as everything is ripping back and the NQ chart looks like a false break and double bottom.
I’m writing this at 10:43 a.m. and feel that the close today will be super critical. My guess is we keep squeezing, but you will probably know the answer by the time you read this.
Part of the reason I think we squeeze this afternoon is that the market knows that everybody knows that Mondays have been incredibly bullish in this (and the last) Trump administration. When markets are political utilities, you see frequent weekend announcements in an attempt to stabilize or pump stocks and crypto.
The stock market has been a political utility ever since the TCJA as that was the moment that the US officially abandoned fiscal orthodoxy. So the government is always looking to help the stock market and that frequently comes via weekend announcements. Or, could be random. But I doubt it. Trump yells about stuff mid-week then he, or Bessent, or someone else rolls it back on the weekend. And the big crypto pumps have also been weekend affairs.
The takeaway, as more and more people are aware of this Monday effect, is that we will start seeing frequent short squeezes from 2 p.m. to 4 p.m. on Fridays.
If you’re wondering why the negative bars for Biden on the left hand chart, the SPX win rate in his term was much lower than average. SPX win rates by president: Bush2 54.2%, Obama1 55.2%, Obama2 53.6%, Trump1 56.8%, Biden 52.8%, Trump 2 58.0%
Here is this week’s 14-word stock market summary:
Bears too confident but timing the bottom is hard. I am bullish and scared.
https://www.spectramarkets.com/subscribe
Interest Rates
Bond markets have been moribund as the lack of US data makes it hard to take a confident view on yields. The slowing labor market could be contracting to below-break-even job creation rates, or it could be normalizing as the immigration crackdown hurts labor supply and AI makes for less hiring at the low-end analyst level in some occupations.
Inflation is not gone, but it’s mostly forgotten as the Fed’s inflation target has moved from 2% to 3% and markets don’t care. Inflation breakevens are definitely not sounding any alarm bells, but then again, those are just markets like any other market. They could be wrong. Inflation markets are not oracles, they’re just a bunch of people and computers trading. Much as people overvalue the accuracy of prediction markets, many seem to think inflation swaps are great predictors of the future. They’re not. They’re simply an aggregation of market participants’ bets and are mostly extrapolating the current inflation regime.
To quote from a NY Fed paper:
This study examines the forecasting performance of inflation swaps and survey-based expectations for one-year inflation. Conducting this exercise helps determine if one set of expectations can provide a cleaner signal about future inflation. The study finds that, overall, inflation swaps more frequently provide better forecasts of future inflation. Previous studies that found poor performance of swaps were strongly influenced by liquidity issues during the financial crisis and the pandemic. When these periods are excluded, swaps have superior predictive ability.
So yea, if you exclude the financial crisis and the pandemic, inflations swaps were pretty good predictors—relative to survey-based expectations! Survey-based expectations are purely an extrapolation of the current inflation environment. So, to put that quote from the academic paper into different words, 1y1y inflation swaps extrapolate the current environment slightly better than surveys do. Extrapolation is empirically a decent forecasting methodology, except across regime changes. Here’s 1y1y inflation swaps vs. actual realized 1y1y inflation.
My point is simply that inflation could come ripping back in 2026. We don’t know. If there is to be a catalyst, I suppose it could be that some of the AI Capex will start to see real-world impacts, creating shortages of energy and labor in various US counties where the big megaprojects start to get built. It’s interesting that the United States is the only country in the world doing these AI megaprojects. Everywhere else has said: “pass”, or is doing AI using much less and much cheaper infrastructure.
Fiat Currencies
Quite a few little subthemes in the FX market right now, but no broad narrative around the USD. A kind-of-scary selloff in stocks should have been maybe bearish AUDJPY or bearish USD (capital flowing out?) or something but we haven’t seen much impact in G10. In EMFX, the impact has been completely invisible. BRL and EWZ continue to rock and everyone that’s sitting in Brazil long carry just keeps collecting their nickels—the steamroller is nowhere to be seen.
The Swiss franc signed a 300-billion-dollar deal with OpenAI early this week and the currency gained sentience by Friday, rallying to a new all-time high against the JPY. EURCHF also made a new daily closing low and an all-time low if you exclude the flash crash on SNB day in 2015. Switzerland agreed to tariff relief with the U.S. and this forced those using the CHF as a funding currency to get out.
CHFJPY is not a currency pair, it’s a swarm of cyber hornets serving the goddess of wisdom, feeding on the fire of truth, exponentially growing ever smarter, faster, and stronger behind a wall of encrypted energy.
That chart is crazy but take a look at what interest rate differentials have done as the BOJ hikes and the SNB cuts. The red line would normally move in line with the currency pair unless that currency pair has achieved AGI.
The CHF is like the JPY in the 2000s before Abenomics. Deflationary macro with a strengthening currency, high real rates, relentless inflows, and no policy prescription. Japan broke the JPY with Abenomics, but the Swiss have zero interest in breaking the franc.
Crypto
Crypto remains soggy as there is no bullish narrative.
- Whales are selling.
- Volatility is low.
- Crypto treasury companies stopped buying because the MNAV premiums came down too much and vol is too low to sell converts.
- We are in the bad part of the halving cycle.
But! I foolishly decided this week to take the other side anyway. It felt to me like crypto sentiment was getting too damp and tech would find a base. Both ideas were wrong and the buyer’s strike in digital coins continues. It is truly difficult to think of who is going to buy crypto right now, but sometimes that’s actually the smartest time to buy. When you can’t imagine it basing, that’s when it bases.
I bought MARA, MSTX and some other crypto treasury type things because the NAVs have all pretty much contracted to 1.0 or so. This gives you some margin of safety if you are bullish, but there is nothing stopping them from going to 0.7X NAV! TONX already did! At the time of writing, I have already stopped out of much of my bullish crypto DAT exposure as my timing was horrible, but if I had to pick a side I would say I am still bullish. This tech and crypto selloff feels long in the tooth and the bears look overconfident to me.
Bulls need a recapture of 98400 in BTC and then maybe (just maybe) shorts will be forced to run for cover. I posted a similar chart showing 98400 last week and you can see that every single line you could possibly draw on a BTC chart is broken, regardless of your crayon’s girth.
That’s generally bearish in the normal world, but I feel BTC has more false breaks than most products so if we get back above 98400 it’s time to get busy for the squeeze.
The idea of giant crayons pleases me.
Tammy Trendline, Head of technical analysis, Spectra Markets
Commodities
Gold and silver went mad again this week, pumping right back to the old highs. Silver made a supreme double top at $54 and cratered once again. The silver chart rivals Troy Polamulu for best Head and Shoulders advertisement.
$49.40/$50.50 is the pivot. Below there, trend dead / fun’s over.
BCOM is getting some attention this week as it makes a new 3-year high. I will say the chart does not look particularly impulsive at this point, but it is kind of impressive given it’s composed 27% of energy commods, and oil is barely moving. This is something to watch because commodities drive inflation. Note that there was a series of lows around 109 after the Russia-induced oil spike in 2022 and we are approaching that now. Get BCOM on your radar.
That’s it for this week.
Get rich or have fun trying.
Some stuff
*************
For Pearl Jam fans (silly)
*************
Man steals bus, goes on joyride
If you know me well or attended UWO in 1991/1992 you get why I’m posting this link.
*************
A fantastic analysis of Fake Plastic Trees via Gittins
A cool cool video. Perhaps the most beautiful song ever written. And how about Phoebe’s version!? Gives me goosebumps and brings me back to COVID.
If I could beeeeee… Who you wanted.
*************
An excellent academic paper about gold
Academic papers scream “BORING” but if you go in with an open mind and ignore the parts you don’t understand and the formulas, they can yield some interesting insights. Just don’t assume the results can be replicated. They probably can’t. :]
*************
*************
Thanks for reading the Friday Speedrun! Sign up for free to receive our global macro wrap-up every week.
