AI Spend Going From Worst Case to Base Case for 2026 Stock Market

Yes, this week is likely to be quiet, but we should learn a lot about what to expect at the start of next year.

A Precious Metals Pullback?

I do not typically track precious metals closely. I pay attention and incorporate the moves, but rarely try to predict much in the space (except periodically trying to capture what as I see as lag effects in gold miner ETFs like GDX/GDXJ, versus the physical price of gold).

But recent returns, especially in things like silver, platinum and palladium, caught my eye.

Three things concern me include:

Parabolic moves have a tendency to unwind just as viciously as they arose. I think we can safely say we have seen some parabolic moves in the space.
Bid/offers are wide and liquidity is low. Some of the recent run-up has occurred with very low liquidity. My “go to” people in commodities are saying they see liquidity (in terms of bid/off and depth of bid/offers) as incredibly low. That will have helped the last leg of the run-up, but may have set us up to be in an “air pocket” where it won’t take much to release the pressure.
Precious metals futures are trading horribly already. While I thought the risk to a downside could start next week, as market making desks (and algos) return to full strength. I am concerned this has already started.

I’d take profit in the precious metals, and am likely to short something in the space (but haven’t figured out which ETF I will use).

The AI Story

As you know, I think there at two big stories to unfold in the first weeks of the new year, but we might get a sense of market sentiment this week, as traders set up for next year.

The AI Debt Diet

Will the hyperscalers start to refine under what conditions they will spend? Are there profit targets? Utilization targets? Or are we still in the “spend it and they will come” mode?

I think the latter possibility (spend no matter what) will result in more downward pressure on the non-chip side of the equation. It will hurt their credit spreads, which will drag their stocks lower. Chips and electricity providers could do well, but even then, I think that “game” is overplayed and markets will force the companies to go on some sort of “debt diet”

An attempt to mollify creditors should stabilize the hyperscalers, but will weigh on chips, electricity production and some of the “unicorns” in the space.

The AI Spend Diet

Will companies continue to spend more and more on AI, or will they pause, and slow it down while trying to truly evaluate the results?

I know that when I put “used AI” in a section, it used to be a “signal” that I’m “tech savvy” and used AI. Now I use it more as a caveat — I included this bit from AI, because it more or less fits my narrative, but it isn’t an important enough part, for me to do the work, so blame AI if it is wrong.

Please re-read the above. It might sound crazy, and maybe it is, but increasingly, this is the tone I’m getting when discussing AI. If correct, it tells me use continues, as we all try, but there is no longer an urgency and imperative to force AI into all facets of workflow.

An AI spend slowdown, if it starts to materialize, will hurt some of the biggest market gainers of the past two years.

I like AI. I will continue to use it. It is evolving rapidly. It is a question of cost/benefit, and I remain concerned valuations are too optimistic (who might emerge as competitors? China? Something new domestically?). I cannot get past the sense that we are spending 2030 prices and only getting 2025 technology (overpaying for what is currently getting delivered).

I will be trying to watch for signals this week. I went from tossing out the “AI spend” concerns as a possibility, to becoming increasingly worried that I should make it my base case.