Post New

https://www.nytimes.com/2025/10/06/opinion/ai-grow...


There Are Two Economies: A.I. and Everything Else

By Natasha Sarin, The New York Times, Oct. 6, 2025

…

The economy is being bolstered by a remarkable investment boom in artificial intelligence. A credible estimate suggests that A.I. capital expenditures may reach 2 percent of the gross domestic product in 2025, up from most likely less than 0.1 percent in 2022. To provide some sense of scale, that means the equivalent of about $1,800 per person in America will be invested this year on A.I.

Without these investments, economic growth this year might have clocked in at around 1 percent. Instead, it is likely to land at almost twice that. Just seven large technology companies are responsible for nearly 60 percent of the gains in the S&P 500 this year….

There are signs that the non-A.I. economy is under duress. As economists predicted, tariffs are pushing up inflation and dragging down growth. Hiring has stalled. Jobs are particularly hard to come by for young people entering the labor market; youth unemployment is at 10.5 percent, a level not seen in nearly a decade, absent the pandemic….

If there were no data centers to build, dollars would flow into other types of investment. It’s possible that other parts of the economy are being held back by A.I.’s dominance. That’s what happened in the 1990s internet boom. Smaller manufacturing companies had a hard time getting access to capital that flooded instead into every dot-com company on offer (some more successful than others)…. [end quote]

As the most established supply chain association, Institute for Supply Management® (ISM®) leads the way in research and reporting for the industry. The Services report represents about 80% of GDP. The Manufacturing report is also important.

https://www.ismworld.org/supply-management-news-an...

Economic activity in the services sector was unchanged in September, say the nation’s purchasing and supply executives in the latest ISM® Services PMI® Report. The Services PMI® reading of 50 percent was at the breakeven point between expansion and contraction for the first time since January 2010. …

The Business Activity Index moved into contraction territory in September, registering 49.9 percent, 5.1 percentage points lower than the reading of 55 percent recorded in August. This is the first time the index has entered contraction territory since May 2020. The New Orders Index remained in expansion in September, with a reading of 50.4 percent, down 5.6 percent from August’s figure of 56 percent. The Employment Index remained in contraction territory for the fourth month in a row and the fifth time in the last six month…

https://www.ismworld.org/supply-management-news-an...

Economic activity in the manufacturing sector contracted in September for the seventh consecutive month, following a two-month expansion preceded by 26 straight months of contraction, say the nation’s supply executives in the latest ISM® Manufacturing PMI® Report…. [end quote]

There are are many details embedded in these reports. The section “WHAT RESPONDENTS ARE SAYING” in the Manufacturing report is a must-read.

Real GDP grew 3.8% in 2Q2025. Personal Consumption Expenditures (PCE) Growth was 2.5% and Real Final Sales to Private Domestic Purchasers was 2.9%. This is very solid growth. The Atlanta Fed’s GDPNow model estimate for real GDP growth (seasonally adjusted annual rate) in the third quarter of 2025 was 3.8 percent on October 1.

Contributions to Percent Change in Real GDP by Industry: Gross Domestic Product (CPGDPAI) | FRED | St. Louis Fed
https://fred.stlouisfed.org/series/CPGDPAI

The Trump tariff and immigration policies are taking time to gradually impact the economy. The current growth rate is very solid but there are signs of slowing.

There are really not two economies – AI and everything else – because the real economy is $30 Trillion while AI is not generating significant profits. Nvidia is making profits from selling its chips but the actual profits generated by AI end-users are much smaller and even losses.

According to Google Gemini: It is generally recognized that at this time (late 2025), the global AI industry as a whole is likely operating at a negative net profit or a net loss, if one were to combine the massive profits from chipmakers like Nvidia and the profitable cloud segments of Big Tech with the enormous, multi-billion-dollar investments and losses from AI model developers and start-ups. [end quote]

The real economy exists in goods and services. The AI economy doesn’t exist yet – but the stock market is hugely invested in it. All the eggs are in one basket – a basket which may not pay off for years, if ever. (As the Chinese are developing low-cost alternatives to expensive chips.)

I am old enough to remember how tech advances have caused once-leading companies, like Wang Computers, to vanish.

An economy is not the same as a stock market. The Trump administration is actively undermining the real economy but it’s still strong. The stock market is in a bubble that could burst at any time.

Wendy

Generative AI is experimental.
To calculate the percentage of US GDP spent on Y2K costs in 1999, you must first determine the total cost and the GDP for that year. Based on reports from 1999, the Y2K costs were approximately 1.2% of US GDP.

Wendy,

Makes sense: this bubble is bigger than the 1999 bubble.

from your post
"A credible estimate suggests that A.I. capital expenditures may reach 2 percent of the gross domestic product in 2025,"

Wendy, you make a good point. For every dot.com company that hit a home run, there were likely a dozen casualties. Some "legacy" tech companies, like Microsoft, HP and Apple (after some fumbles) were able to survive the transition, others like Wang, DEC, Boroughs, even IBM and Xerox either didn't make it or were significantly diminished.

The first quantum change in datacom technology was the advent of the personal computer (and most of the pioneer companies died within a decade), followed by the advent of networking and the internet. Again, Darwin was an optimist and the dot.com crash wiped out most of the contenders.

I was intimately involved (professionally) in both of these extinction events and they had similarities:

At the beginning of each "era", if a company could "fog a mirror" they had nearly unlimited funding. The stock market grew to irrational levels. Companies didn't have to be profitable for their stock to rise, they just needed a good story. Then, it was discovered that some of the promises made by some of the companies were bogus and money stopped flowing in indiscriminately. When the stock market hit the fan, there was a big sucking sound and companies tanked in droves. The ones which made it through might not have had the most worthwhile products, but spread across them, most of the boxes were checked and the rest of the companies became redundant. In some cases the successful ones were led by those with extreme talent, but in many cases they lived because they had the good luck to be in the right place at the right time.

We are more than rhyming with the past. Like before, we have the profile of a future goal with the promises of quantum-sized profits. If you can insert the meme of "AI" into your description, people will pour money into your threshold. It is clear (to me, at least) that the majority of the pure AI (software) companies will not survive the next few years. None are making money (in that division, if a broader company like Meta, Amazon, Google, etc. - in which case it is simply a drag on their current profits). As long as they don't get sucked into receivable losses when it hits the fan, the hardware guys are already making a profit (Nvidia, AMD, data center/electrical infrastructure companies). Those account receivable loses during bankruptcies can be pseudo-random and a big deal.

Now, AI be a big deal and can likely cause dramatic upheavals and changes in our society, but before that happens, taking Wendy's explanation of the huge sums currently sequestered in AI-associated stock, my opinion is that there will be a significant "cleansing event" when there will be a "Sorting Hat" party and (assuming the big guys will drop precipitously, but survive because of their substantial other businesses) a large number of companies (and investors) will bite the dust. (OK, so it's a bit of a run-on sentence).

While we are probably a year or two from this coming to a head on its own, it can be accelerated by a coincidental by a coincidental loss of liquidity or other "black swan" type of event and be a sizable component of a resulting crash.

Despite what you've heard, it is not going to be different this time.

Jeff

While we are probably a year or two from this coming to a head on its own,

I think sooner: the housing market in many regions is crumbling. That will gain momentum.

Different? Most of us are far too old to believe capital flows are different.

If you can insert the meme of "AI" into your description, people will pour money into your threshold. It is clear (to me, at least) that the majority of the pure AI (software) companies will not survive the next few years. None are making money (in that division, if a broader company like Meta, Amazon, Google, etc. - in which case it is simply a drag on their current profits).

One thing that seems to be different is the amount of money. I don't have any stats, but back in the dot.com era it seemed like early investment rounds were in the low millions of dollars, with maybe tens of millions in the later rounds.

With AI it is more like tens of millions in the initial rounds and up to billions, even tens of billions, in the later rounds. And that doesn't include what the tech giants are all spending directly.

I'm curious to see what the copywrite law says about a "derivative" AI which is tasked with asking your question to a number of other available AIs and then consolidates an answer for you (sort of taking a vote to eliminate hallucinations and ignorant AIs which don't know the answer).

Jeff

Post New

Macroeconomic Trends and Risks FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.