Unless you can demonstrate that there is in fact an AI bubble, that is simply begging the question.
Notice how the term “housing bubble” is used much less frequently today than 10 years ago? That’s because that so-called bubble has been ballooning in size for three decades now, and almost nobody still believes that it will “burst” in any meaningful sense. The Dotcom bubble was in many ways an outlier.
People often use it instead of 'raises the question'. E.g "There was very little fallout to the Y2K bug, which begs the question: was the Y2K crisis real and well handled or not really a crisis at all?"
The longer it takes to burst, the worse the outcome.
Housing is being propped up by the governments of the west because it was already so problematic if it fails that millions of people would be severely impacted.
Even the housing backed mortgage crisis of 2008 was as large a shock as the great depression, the reason we’re not all using money as toilet paper is due to government intervention. Rightly or wrongly. Some people believe that this intervention makes something worse bound to happen later- and based on the cost of housing I tend to believe them. It is not sustainable to keep housing at its current cost, and the financial model requires that they continue to increase in price. If house prices fall it is a disaster for millions.
There are a few noticeable differences between housing as an asset and AI datacenters. Beyond the obvious difference that a house has a depreciation in the order of decades, whereas AI GPUs are a few years.
Then there is the fact that housing is a fundamental human need whereas AI, as frequently demonstrated, isn‘t even a want by many people.
I am not saying that AI cannot demonstrate todays value 20 years into the future. But there is zero reason to believe the short to medium term payoffs on AI investment will be proportional to the investment we‘ve seen over the past few years
> a house has a depreciation in the order of decades, whereas AI GPUs are a few years.
That’s no longer true for datacenter-class GPUs. The A100 came out six years ago, and it still sells for $15k+, with no meaningful drop in the foreseeable future.
I think housing as an appreciating is a weird Western capitalist socio-cultural artifact - in order to enter the middle class, you have to buy one, and the way to go about it smartly is to get a bank loan. Most people have no way of affording one otherwise, and if they did, paying sticker price for it would be dumb.
So everyone who's not considered a deadbeat shall be indebted to the bank for decades to come. Their prices are also supported by the collective power of the people - for whatever unreasonable price they're charging for it, you can be certain that someone (or some bank) has paid it, therefore it's in the best interest of significant chucks of regular folks that house prices don't go down.
Imo something similar is going on with retail investing.
Btw housing in China is not an appreciating asset, because you don't own the land, you merely lease it from the government for 70-100 years, which is about the natural lifespan of a lot of buildings, after which maintaining them becomes more expensiv than rebuilding them.
> That’s because that so-called bubble has been ballooning in size for three decades now, and almost nobody still believes that it will “burst” in any meaningful sense.
That's because Boomers live far longer than prior generations thanks to medical advances. The housing bubble will collapse (at least outside of the megalopolises) once the Boomers finally start to die en masse due to their over-representation in demographics.
But before that, the pension systems will crash hard. For people in systems with redistributions (like most of Europe), there simply aren't enough working age people contributing payments for the pensioners, and for people in stonk-based systems (e.g. US 401k), they will run into the issue that someone has to buy the stocks that the pensioners sell off to fund their retirement, and ain't no one of my generation buying stocks, thanks to us having to spend insane amounts of rent.
The problem is, people more and more can't manage to pay these rents, so they're either cutting back on any absolutely not necessary spending or going homeless outright. For now, there are enough desperate people that still have some money to pay rent... but retail, no matter which industry, is feeling the impact of people having no spending money hard.
Without the AI bubble artificially propping up the GDP, it is most likely the US economy is in a recession [1].
Notice how the term “housing bubble” is used much less frequently today than 10 years ago? That’s because that so-called bubble has been ballooning in size for three decades now, and almost nobody still believes that it will “burst” in any meaningful sense. The Dotcom bubble was in many ways an outlier.