While Wall Street seems bullish that Big Tech’s bet on AI is justified—especially with a recent rise in AI services and cloud revenues—some economists are worried that the boom could end in a bust.
If AI doesn’t deliver the productivity gains its biggest boosters are hoping for, the Great Data Center Buildout could be a repeat of the early 20th-century railroad boom or the dot-com bubble, when tech optimists overestimated demand.
**But…**some experts say that the effect of the AI spending frenzy on the US economy might be overstated since much AI-powering semiconductor production happens overseas.—SK
An engineering friend formed a good description of AI. He said, (paraphrasing) plenty of apps need a layer of coding to work in new ways on top of their original function and with other apps. While that might seem a very imperfect definition, it is the effect of having AI.
To Bob’s downstream comment, there is little to no profit in it for the AI companies other than some of the chip producers and designers.
The chip producers are not downstream.
The coding is not economically worthwhile.
Google, Apple, Microsoft, Meta and Nvda/Tesla have the budgets to do that coding regardless. Tesla may have shifted Dojo resources to Nvda because their financials are problematic.
Are you seriously asking what is the potential of AI? A couple of here-and-now examples. Microsoft has told us that already a third of its coding is done by AI. The oil majors are using AI to examine their huge seismic databases.
It is these increases in productivity and resources that help are economy, more than the fact that the chips are manufactured in Taiwan.
Here is the problem Bob. Any massive capex, benefits economy only when it results in increased employment, higher wages. The reality is, the employment scenario, under the hood is bad. The AI investment boom, and stock market are hiding the following:
Housing in recession, Home prices are contracting; Builders are reducing their inventory; Commercial construction is down;
Look at banks lending, they are hardly growing; Banks are making money from treasury investments, M&A fees, etc. Very few regional banks are actually increasing their interest income from loan book growth;
Consumers are resisting price increase; While the unit prices are going up, the volume is going down; so you may not see it in inflation, but Consumer is declining; This is in fact what Scott bessent argued that tariff will not result in inflation;
The coding benefits are well documented and one of the relatively easy to achieve, because of the streamlined workflow in coding; Real life is not that well defined workflow’s.
On Oil Majors using AI, that is c3.ai bread and butter; The stock is down today 23%; The productivity gain on this is mostly squeezed;
The AI investment cycle might slow down, not necessarily similar to dot.com, because the big players are very profitable companies, but they need to see return on their investments to continue to invest. For now, they are doing land grab by investing heavily. When the valuations are stretched, that could create trouble in these names.
I don’t want you to think, I am all negative, but I have listed the negative points. Of course there are worries, and concerns at all times. But, I am not seeing AI is going to create a productivity boom immediately. Over medium to long term, sure.
When the volume goes down… growth stops. You can preserve margin, but the growth premium in your stock price and PE will come down. At some point, slowly, but surely, inflation will creep up. The administration is counting on slow inflation, not post-COVID like melt-up and melt-down, but permanent inflation around 3%.
If you are retired, living off savings, watch out. Inflation is your worst enemy and when it is combined with SS not keeping up with inflation and medicare/ medicaid cuts… “it is no country for old people, or poor people”
The big three have been losing market share for years. Their response has been to drop their lower priced models, jack up the price of what is left. As volume drops, they close surplus factories and lay off tens of thousands of workers. But that’s OK. It works for the CEO.
You can use AI to review programmer created code changes. Takes almost no extra effort and can find bugs before they cause a problem. Also very useful for creating additional test cases…same idea.
But beyond that asking AI is far better than a conventional web search (Google, etc) because you can ask follow up questions if your query wasn’t specific enough. Most everyone will like this.
Professionals doing things like CGI movies will be able to use prompts and follow ups to create scenes that would have taken many hours, previously
Yes, an automatic Jack Welch emulator, capable of gutting a company in a fraction of the time it took Welch. I recall him saying, as he carved off one division after another, he didn’t want to be in a market where GE wasn’t already #1 or #2. Seems to me that a smaller player in a particular market has more room to grow, by sharpening their business and taking share away from others. But, no. Welch apparently wanted to sit on segments GE already dominated, so he could put everything on autopilot and go play golf.