AI Bubble - crystal ball time

[Drastic]

Just catching up here (sup y'all) and wondering your thoughts on the next couple of years, economically overall?

[rcjordan]

Recession seems imminent, maybe already underway.  AI is already impacting tech jobs, serps, & traffic --effects likely to get worse.  Feds have made 3 cuts in a row and Debbie doesn't think that they've had much impact.

Tariffs have hurt small biz & farmers bigtime but the losses haven't been fully realized yet.

It's hard to say how much the hospitality industry is suffering. International tourism is down, but local travel may be taking up some of the slack.

Restaurants seem to be hurting pretty much across the board.

Dollar General is reporting puchase downshift by poorer customers but an increase of middle class customers.

All in all, pretty dark --but, hey, this is Debbie Downer speaking, hhh.

What does your Debbie say? 

[ergophobe]

The correlation of professional economists' predictions of recession and the occurrence of recessions are near zero.

That tells me that anything I would say is random guessing.

My goal is to try to have a plan that avoids catastrophe without doing too much catastrophizing.

My worry is less that a recession is here or imminent than the signs that our system is in an unusually brittle period. The problem with brittle is that everything looks okay until it shatters.

The way things are currently held up in a sort of house of cards means that things could spin out of control very fast, much faster than in 2008.

So we can spin out scenarios of the perfect storm that result in a crash bigger than anything we've seen since the Great Depression. I can't think of a time in my life where we had the confluence of so many troubling factors.

But that doesn't mean that it's going to happen. There was a time when the number one concern of elementary school kids was nuclear war. Sci-fi written in the late 1950s to late 1960s took it as a given that there would be an all out nuclear war by 2000 or, at the latest, in the early 2000s. We had a Civil Defense unit in 5th grade where we learned about washing off radiation dust and storing water. And yet...

So rather than betting on one future or another, my goal is more to reduce the brittleness in my own plan.

Having gotten all that out of the way... I've never been more nervous about the economy. On the other hand, I don't actually think much about nuclear war.

[ergophobe]

PS - an old friend of mine looks at the spending, consumer debt, government debt, concentration of the stock market, P/E values, consumer confidence index and says, "This can't go on."

He's been saying that as long as I've known him and always for the same reasons.

But to avoid being so wishy washy, I am sort of betting* on a sluggish economy, a slow recovery and a 20% drop in the stock market in the next couple of years, but I am betting that the big crash will not come for 3-8 years and still may be averted if people stop acting like it's impossible.

[Drastic]

Anecdotally, everywhere I go out to eat looks maybe half at best what they were 6 months ago. Commercial money has been super tight the last 4-5 months. And we're usually rather insulated from national swings here.

It seems the stock market is propped up by ai. Execs are finally figuring out it ain't what's been sold as. Nvidia, ram, etc. numbers are just crazy from the ai bubble. Tech people losing their jobs to "ai" when it's now just offshored. I've been thinking this thing was going to pop soon before the idea became national news and when fully realized, it's not gonna be fun for anyone.

The gap between the haves and the have-nots seems much wider this year. If millions lose ACA plan access, it will just compound.

I think we're in for a rough ride sometime soon. I just don't know how soon. Personally, I really hope we have a slightly above average year in 26 and we don't actually feel all of this until mid 27, but I'm just huffing hopium at this point.

I am also reminded of my pops talking about a bubble popping for more than 2 decades before it finally did, and not exactly for reasons he expected. It seems like uncharted territory and we've decided to head towards the storm. I'm definitely uneasy and packing on everything I can for now.

[buckworks]

>> goal is more to reduce the brittleness in my own plan.


That's about the most that we ordinary mortals can do.

Now to figure out what that actually means!

[Brad]

AI and data centers are becoming synonymous in the public eye.  Local opposition here is growing against data centers.

Besides having a big noisy concrete box next door there is the fear that costs for water and added electricity will be fobbed off on residential consumers by the utility companies.  So if AI data centers  cause the costs of water and electricity to go up at the same time ACA goes up and healthcare in general goes up plus inflation, it's like the perfect storm.  There will be trouble.

[littleman]

If you subtract AI we are already in a recession with AI being responsible for 92% of the US's growth in the first 6 months of 2025.  When we take into account a lot of the Round Robbin spending between OpenAI, MS, Nvidia, Oracle and the like I think we're very likely heading to a substantial crash.

I don't really trust the inflation and employment numbers being published right now to be accurate.

My Debby says there is going to be an AI bust, then the poor performances of the rest of the economy will be harder to ignore.  The average Joe has very little faith in the economy right now and that sentiment is just going to get stronger.

Even if AI doesn't bust and they figure out how to make it profitable, it is going to come at the expense of the middle class, as the ultimate "win" for AI is a "loss" for jobs.  We just have the Camel's nose in the tent in that regard.  This is inevitable imo.
 
 

[ergophobe]

It's easy to find evidence that a recession has started and that an AI bubble will pop. What's always hard is pinning it down to 2026.

So evidence rolls in....

Goldman Sachs makes unemployment prediction
https://www.msn.com/en-us/money/markets/goldman-sachs-makes-unemployment-prediction/ar-AA1Sp6BG

And then there is all the big structural stuff we have in the data center thread, which littleman alludes to as well.

And then there is the fact that the top 10% accounted for 49.2% of consumer spending in Q2 2025.

So you have a perfect storm where

 - all the growth in GDP and the stock market comes down to the Mag10.
 - consumer spending is overwhelmingly accounted for by the top 20%
 - the top 20% are particularly sensitive to portfolio health when it comes to spending
 - so a minor drop in valuations based on AI gains spreads quickly into consumer spending which spiral the economy down

In the bad case scenario, that has other follow on effects
 - people get spooked and pull out of the market, but since so many investors are in index funds, the drop in AI-based valuations then takes everything down
 - wealthy people pull back even more on spending
 - older workers decide to work longer, putting even more pressure on the increasing unemployment for the young
 - bond investors get spooked. One big change in the bond market is that more and more bonds are held by hedge funds, and that means that once they think it's not a good hedge, they pull out. So bonds have more potential than usual for volatility.
 - this keeps the spiral going.

In the really bad case, the government tries another bailout but finds out that investors just don't want to buy T-bills at the price offered and the government has to offer more interest and this sets off a really bad spiral.

Many people don't realize that government debt is not like a 30-year mortgage. A lot of it is held in short-term bonds, so when interest rates go up and those bonds mature, they get turned over into higher rate bonds, which means debt service can go up very rapidly in that case.

That ultimately means that the government cannot actually do a bailout because it doesn't have enough credit, so the mechanism that has saved us in every downturn since the Depression fails and things start to really spin out of control.

Whew!

Okay, so all of that is, I think, plausible. The problem is whether that will happen in 2026 or 2030 or 2036 or we get our act together and avert that scenario.

So I tend to look at all of this and spin out doom scenarios, but let's try to "red team" all of that.

Here's why we are not going to have a major downturn in 2026.

First, Fed rate cuts. We've already had a few and the Fed has announced another, but remember that Trump is going to appoint one of his lackey's when Powell's turn is up. The main requirement of this lackey will be to print money and juice the economy to try to keep the cadaver alive through the 2026 mid-terms.

Second, Redfin sees a "reset" in the housing market where things will start to pick up there.

The Great Housing Reset will take shape in 2026. It won't be a quick price correction, and it won't be a recession. Instead, the Great Housing Reset will be a yearslong period of gradual increases in home sales and normalization of prices as affordability gradually improves.
https://www.redfin.com/news/housing-market-predictions-2026/?msockid=28b967ad448061130d12713a45cf60e8

Meanwhile, all the TALK of an AI bubble means that investors are already spooked and already starting to take the bubble into account. The bubble alarm got sounded fairly early on this one, so it will be a fizzle more than a boom.

Finally, we have the unemployment problem. This makes a lot of headlines in tech and among young people, but this reflects overhiring during Covid, not a structural problem due to AI or a slowing economy. We are bringing the numbers in line with where they should be.

So all in all, 2026 will be a bit turbulent with some trouble signs, but in the end the economy will hum along for at least another year before a big downturn

[ergophobe]

Of course... that's getting harder and harder to steel man

 - Powell just said labor market continues to soften as inflation remains elevated
 - Florida leads nation in wave of foreclosures (Money Talks News
 - China has seen the first investment decline in 3 decades (NYT)
 - China faces deteriorating economy on three fronts (WSJ)
 - California LAO says deficit will hit $18b, $5b more than planned and is "structural" and will go through at least 2029
 - McKinsey plots thousands of job cuts in consulting (Bloomberg). But not that if you want to buy into the "red team" argument above: " From 2012 to 2022, the firm's employee count climbed from 17,000 to as high as 45,000. Since then, it has slid to around 40,000."

And those are just the headlines I saw in the few minutes since making the last post

It doesn't look good friends

[rcjordan]

US corporate bankruptcies set to hit 15-year high amid credit jitters, S&P data shows | Reuters

https://www.reuters.com/legal/litigation/us-corporate-bankruptcies-set-hit-15-year-high-amid-credit-jitters-sp-data-shows-2025-11-13/

[ergophobe]

If you look at Google Trends, search interest in "housing bubble" peaked in August 2005. By 2008, it had come most of the way back down.

Trends only goes back to 2004, so there's no similar data for "tech bubble" or "dot com bubble" or "internet bubble" but an analyst who was in telecoms in the late 1980s through the crash claimed that there was a lot of talk of a bubble in 1997-98, but by 1999 the "this time it's different" narrative had taken hold.

If you look at Google ngrams, occurrences of those bubble terms trail way behind the events, but that is probably due to how long it takes to get a book out.

Search interest in "AI bubble" is about half of what "housing bubble" search interest was in 2005.

https://trends.google.com/trends/explore?hl=en-US&tz=480&date=all&geo=US&hl=en&q=housing+bubble,ai+bubble&sni=3

Anyway, the fact that peak concern about a housing bubble was in 2005 means that I'm mistaken to believe that if so many people are worried about an AI bubble, maybe that will help us avert the worst of it.