TL;DR: I am not as bullish on AI's potential as many. I also believe investors are using AI's potential to overlook a number of market realities that would normally take the market in a different direction, and are using that potential to anticipate higher valuations for the larger market.
I have written before about my skepticism about AI's potential, at least in the near-term.
- I believe AI has grabbed most of the low-hanging fruit. Dramatic improvements from here require the creation of different technologies.
- The required computer power for any significant improvements will not be as easy to build as people think. The numbers are outlandish (see the article that kdogg posted above), we are starting to see NIMBYism, there are regulatory constraints, etc.
- Monetizing AI is problematic. AI companies want market share, so they are limited in what they can charge.
- Even if AI succeeds, a major difference between the AI boom and the tech boom is that the tech boom promised to (and did) create a whole new class of jobs, while if AI succeeds as its proponents envision, it will destroy far, far more jobs than it creates. This piece is roundly ignored.
Market realities, many of which have not fully been realized:
- Inflationary pressures are inevitably going to weigh on the economy.
- Tariffs have settled in at a minimum of 10% (with a handful of exceptions), which will be a permanent inflationary bump.
- Electricity costs are going to continue to rise, between the government actively working to slow bringing renewable energy online and the increased demands from new data centers for AI.
- Health care costs continue to rise, with no real end in sight.
- etc.
- The government has taken a heavy hand on its reporting of macroeconomic data, and seems to be settling into a pattern of releasing positive data for the current period while revising down data for previous periods. This means we are flying a bit blind, and likely are getting data that has been massaged to be rosier than it really is for political purposes.
- We live in an age where companies have gotten very good at manipulating balance sheets to put up a good face for the market. I've seen some of this first-hand.
- Challenges with different timelines of AI progress:
- AI is successful in the short term: tons of jobs are destroyed. Job insecurity rises. Consumer spending plummets.
- AI is successful in the medium term: AI companies have to navigate monetizing products while trying to capture market share, while investing a ton in data centers. Employees go through an extended period of wondering when their job will be eliminated by AI.
- AI is successful in the long term: in the short term, the market punishes AI companies for their investments with no return.
- In any of these cases, the rest of the market gets punished.
Put differently: there is no such thing as a free lunch, and I see a lot of free lunches. The market is at an all-time high despite some of these headwinds.
I think the market is buoyed by the following beliefs:
- AI companies are going to become incredibly valuable, while AI is going to make a lot of companies more profitable. (My skepticism is outlined above.)
- Rate cuts are a given. (Inflationary pressures make this trickier. Political pressure may force cuts to happen when they shouldn't, but this would eventually be another free lunch that needs to be repaid, along with any distrust this creates in the market.)
- Where else are we going to invest? (This is bubble behavior to a T, though as I've written, I cannot discount that the growing wealth inequality may make the market a self-fulfilling prophecy for a while, because what else are the wealthy going to do with their money?)
I'll be interested to see what happens in January. I've been considering taking some profits in the new year to defer the taxes, and I wonder if there are others out there doing the same.
) of the Bloomberg article: 