Artificial Intelligence, real Money
February 6, 2026

I think it was 2023 that I listened to Jensen Huang speak where he outlined the future of computing in the face of the advent of AI, and what he said made sense to me: basically, that the mountains of data collected by the hyperscalers would be used as training models for an entirely new form of large-scale computing infrastructure built largely around his companies' products. As a matter of competition, the hyperscalers would have no choice but to engage in this build-out at fantastic expense. No one could afford to be left behind. He was correct and Nvidia is now the most valuable company in the world.

At the time, I took a position in Nvidia and later, after that investment did well and Mr. Huang revealed that his company couldn't build chips fast enough to satisfy demand, I sold out of Nvidia and switched to AMD, where I believed growth would accelerate as overflow demand moved in AMD's direction. Last year, I also invested in memory-maker Micron since, well, all those GPUs require memory.

I also had good-sized stakes in companies like Meta, Amazon and Google. Until recently. Last year, Meta announced their upcoming capital expenditure forecast, what the projected forward costs of their AI build-out would be and the numbers were shocking. So much so that I sold out of Meta last fall. Tough decision as it's a very well-run company with a premier CEO...but, in my opinion, they were spending too much money. I also sold off Amazon as, at that time, the stock had not been outperforming the technology index.

For this earnings season, I expected things to be fine...and one might argue that results in many respects were positive...but maybe not good enough; as more earnings results were revealed, we saw more evidence of massive spending. I was caught flat-footed when Alphabet revealed its capital outlay would be more like $175 billion dollars when analysts were expecting something in the $115B range. I sold off my Alphabet position the next day. Then, yesterday, Amazon reported they expect to spend $200B this year. Folks, these numbers are truly stupendous...and they raise a question; how long will it take see profits that justify this level of spending? The market sell-off this week says to me that this question, and others, are being asked.

And another thing, an observation I did not make but found alarming: what happens to net income/EPS when these expenditures need to be depreciated? Reportedly, AI hardware might be depreciated over a 5 year schedule, so, roughly speaking, Amazon's $200B 2026 capex spend might affect earnings to the tune of $40B per year for 5 years. But these add up: if Amazon spends $200B this year and next year and on and on, what happens 6, 7, 8 years from now? Amazon had about $77B in net income last year...even if this doubles in the years ahead, much of their earnings could "disappear" in capex depreciation. Of course, this may be an oversimplification. But still...

Traditionally, companies like Alphabet, Meta, Microsoft et al. were asset-light businesses with very high margins. All that seems to be changing as they spend vast sums on hardware, town-sized computing centers and get into energy deals with nuclear power plants of all things to supply energy to these costly projects. I don't really want to own individual shares in any of these right now.

But what about the Nvidias or AMDs or Microns? Aren't these companies akin to selling picks and shovels during a Gold Rush? Maybe yes...but the overall negative market reaction to all this spending also tells me that the market is thinking that something has to give. I have a feeling that, at some point, one of the hyperscalers is gonna blink and report that they're trimming all this colossal spending. Once one goes, the others might, after which the pick and shovel sellers won't look so good. Recall what happened after Mark Zuckerberg announced the "Year of Efficiency" at Meta? The stock took off like a rocket. Yes, the situation is different here...but how different?

Please bear in my mind these are merely off-the-top-of-my-head musings; I write them here only to attempt to organize my thoughts.