• Welcome to the Internet Infidels Discussion Board.

AI and the economy

I was listening to NPR this morning. It was about AI and data centers. The expert said that 90% of it is financed by loans.
Seems risky.
He likened it to the expansion of trains in the 1870's. The main financier needed to raise $150,000,000. He could not. So, that sent a shock wave though the industry. No trains lines were expanded for some years.
 
The debt-financing angle actually connects to pood’s demand problem. If AI capex is largely loan-financed, the first crunch may be financial rather than mass unemployment: lenders expect future revenue, but automation that suppresses wages also suppresses the consumer spending needed to repay those loans. The Jay Cooke/railroad parallel is loose but relevant—overbuilt capacity on credit can collapse before the promised economy arrives. That doesn’t disprove AI disruption; it just suggests the transition is more likely to be a messy boom-bust than a smooth billionaire-run utopia.
 
And look at what's happening in China. They weren't willing to slow down as it became obvious that they would saturate the demand for capacity in China, now their economy is a mess because the whole thing wasn't built to reality. They get ghost cities, we're probably going to get ghost data centers.
 
I was listening to NPR this morning. It was about AI and data centers. The expert said that 90% of it is financed by loans.
Seems risky.
He likened it to the expansion of trains in the 1870's. The main financier needed to raise $150,000,000. He could not. So, that sent a shock wave though the industry. No trains lines were expanded for some years.
I have said before this is a bubble ready to burst.
 
I was listening to NPR this morning. It was about AI and data centers. The expert said that 90% of it is financed by loans.
Seems risky.
He likened it to the expansion of trains in the 1870's. The main financier needed to raise $150,000,000. He could not. So, that sent a shock wave though the industry. No trains lines were expanded for some years.
The analogy of AI and railroads works on many levels. Most early railroads were fairly short and there simply wasn't enough demand for passenger seats and freight to pay operating costs. Naturally, maintenance was the first thing to get cut from the budget. The result was thousands of train accidents across the country, with thousands of deaths. Rails and trestles deteriorated. Passenger cars were wooden boxes which collapsed, so derailments and collisions had a very high casualty number. The death knell for surviving short line railroads was the introduction of cars and trucks.

I don't imagine unprofitable AI Data Centers will kill people when they collapse, but we will see what we saw in railroads. The small ones will simply be padlocked while the largest merge and combine until sales volume generates enough profit to cover costs. Just as today, there are only seven major railroad companies in the US.

Today, there are probably more abandoned shopping malls than operating ones. The big difference between a mall and a data center, from the real estate angle, is malls are always in a population center. There is some use for the real estate. In my neighborhood, a mall that dwindled for twenty years before they turned out the lights, was bulldozed and is now an Amazon fulfillment center. AI data centers in rural areas will probably be left to collapse and return to nature, in the manner of Mayan ruins.
 
Back
Top Bottom