Key insights
- The market anticipates a 25bp Fed hike in December, with focus shifting to October PCE data. If PCE meets expectations, it could signal the end of rate hikes, potentially boosting long-duration assets. However, risks remain from high PPI, potential hot PCE, and AI-specific concerns. Nvidia faces SEC probes and export investigations, while AMD is under scrutiny from a short report regarding its dealings with a blacklisted Chinese server maker.

Futures are pricing a near certain 25bp hike at the December 8 and 9 meeting, with only a small chance of 50. A hike that's fully priced doesn't give anyone a reason to sell, so the meeting itself matters less than two things, Wednesday's October PCE, which lands alongside the second estimate of Q3 GDP, and the 2027 dots that come out with the decision. Headline PCE was 3.6% in September and nowcasts have October at around 3.8%. The reason rates matter so much for equities right now, which most of you know, is the 10 year. It has closed above 5% almost 30 times this year, matching the 2007 run. When the risk free rate is 5%, earnings ten years out are worth a lot less today, and the stocks with most of their value far out in time like AI infra, high multiple software etc, take the biggest hit. That also flips very quickly, if PCE comes in on expectations, most of the Fed's work is done, and one last hike into year end with no further hikes on the dot plot is the cleanest path to the 10 year coming off 5%. I think that sets up long duration with a huge V shaped recovery. But but but, rates aren't the only problem, inflation isn't cooperating, with October PPI at 6.5% year on year, the highest since late 2022, a hot PCE print is a real risk. Then there's a month of AI specific fear that keeps stacking up. In October the debt market started refusing to fund AI without cash flow behind it, with CoreWeave's bonds yielding over 15%. On November 5 the midterms put a data centre moratorium in Texas on the table and the Nasdaq 100 had its worst day since 2020. Nvidia then lost Jensen Huang, the SEC opened a civil probe into its disclosures, the Megaspeed export investigation became a criminal case, and the Treasury Secretary wants answers from the board within about 10 days. AMD was hit by a Hunterbrook short report alleging its firmware kept reaching a blacklisted Chinese server maker, pointing to $48.8B in commitments against $15.8B in cash. Anthropic raised around $81B in the largest IPO ever, then the market fell 2% the day it listed. None of these need to fully resolve for a bottom, but until they stop getting worse, every rally in AI names is fighting the tape. I think the bull case here is that most of the FUD and extreme scenarios are priced. Nvidia beat, guided Q4 to $130B, and trades around 9.7x next year's earnings while AMD sits near 33x. That's the kind of multiple you'd put on a business you expect to shrink, which it obviously isn't. Breadth looks washed out, with 529 Nasdaq stocks making new 52 week lows on November 19 and decliners beating advancers 3.5 to 1. Bottoms tend to form on bad news when everyone is already positioned defensively. Risks are real and I've been hedging more this month, but I think the low most likely comes this week. If October PCE comes in at or below the roughly 3.8% nowcast, I expect the Nasdaq to close Friday above its November 20 close and not revisit those lows into the Dec 9 meeting. A sustained recovery also needs the Nvidia probe headlines to calm down without new findings. If PCE runs hot, or the dots add hikes for 2027 and the 10 year breaks higher, I'm wrong and I'd expect another leg lower in AI names. One caveat is that Thanksgiving week is thin, so moves in either direction can get exaggerated. How is everyone else positioning their book and what say catalysts/risks are the main priority here?