Without Data Centers, GDP Growth Was 0.1% in the First Half of 2025, Harvard Economist Says (fortune.com) 83
U.S. GDP growth in the first half of 2025 was driven almost entirely by investment in data centers and information processing technology. The GDP growth would have been just 0.1% on an annualized basis without these technology-related categories, according to Harvard economist Jason Furman. Investment in information-processing equipment and software accounted for only 4% of U.S. GDP during this period but represented 92% of GDP growth.
Renaissance Macro Research estimated in August that the dollar value contributed to GDP growth by AI data-center buildout had surpassed U.S. consumer spending for the first time. Consumer spending makes up two-thirds of GDP. Tech giants including Microsoft, Google, Amazon, Meta and Nvidia poured tens of billions of dollars into building and upgrading data centers.
Renaissance Macro Research estimated in August that the dollar value contributed to GDP growth by AI data-center buildout had surpassed U.S. consumer spending for the first time. Consumer spending makes up two-thirds of GDP. Tech giants including Microsoft, Google, Amazon, Meta and Nvidia poured tens of billions of dollars into building and upgrading data centers.
You're not wrong but... (Score:5, Insightful)
Those are all temporary jobs with temporary benefit. If you want to see what those datacenters are like in the long term... check out Grok's Memphis facility. Less than a dozen vehicles in the parking lot.
Re: (Score:1)
Yeah, but they're getting incentives... (Score:3)
They're promising thousands of jobs, when it's literally eight permanent full-time jobs.
Re: (Score:1)
Re: Yeah, but they're getting incentives... (Score:4, Insightful)
Endless growth is called cancer.
There is no planet b.
We haven't developed space enough to move industry there.
Do the math, instead of meth
Re: (Score:2)
Re:Yeah, but they're getting incentives... (Score:4, Insightful)
I expect someone here to answer that we should just spend near infinite amounts on all of the above. Why not just give everyone a check for $1m?
I'm gonna go ahead and answer you seriously, because why not? The answer is sustainability. Believing everyone should be entitled to the requirements for survival (e.g. food, clothes, shelter, medical care) as well as education (most people disagree only about how much) is different from believing that everyone should get everything they want. For example, you could believe that without working, you shouldn't be able to own much. Instead of handing people a bunch of money, you'd provide them with housing and food, etc.
But it's actually not sensible to do it that way, because if you give people money, they spend it and you tax it, and then they spend it again and you tax it again, and you get it all back eventually but in the process it induces a lot of work. And isn't work what you want done? So yes, of course you hand people money. And then instead of handing them food and necessities, people just go buy those things from this fearsomely efficient crap production machine we call capitalism, and the wheels keep on turning until we deplete our natural resources and/or destroy our biosphere with industry... at least, if the road we're on now continued. But that's not inevitable, it's literally only because we're continuing to allow big oil and friends to decide where we get our energy, how we're going to transport people and goods and so on.
Re:You're not wrong but... (Score:5, Insightful)
Re: (Score:3)
Re: data center construction as blue collar jobs program
Too bad they are not building houses, something we know we need more of.
Re: (Score:2)
How many have you asked? Why?
Rent's expensive also; rent prices generally mirror housing prices because they are roughly exchangeable options.
Re: So the tech industry is giving back (Score:2, Troll)
We do not need more houses. We need to seize them from those keeping them empty and put people in them.
Re: (Score:1, Troll)
People seizing and living in unoccupied houses make the news occasionally. Most of them treat the houses they occupy poorly, some of them trash the houses -- what have they got to lose? Generally, people responsible enough not to damage the building they occupy are responsible enough to find a way to pay for a place to live.
"We need to seize them...". "We" is a pronoun that includes the person generating the statement, in this case drinkypoo. By your own statement you have put yourself into the category of
Re: (Score:1)
Corpos seizing and stockpiling unoccupied houses make the news occasionally. Most of them treat the houses they occupy as pork bellies, some of them hoard the houses -- what have they got to lose? Generally, people mortal enough not to be immune to the elements need no more than one place to live.
Re: (Score:2)
Why would someone liable to be removed from a property at any moment care about its upkeep? Having an ownership stake in the property would change the arrangement entirely.
Re: (Score:1)
And when the bubble pops, there will be lots of these huge buildings sitting empty and unused. They will be difficult to use or sell because very few people/companies need buildings whose size is measured in acres.
I expect that such buildings will find a use in a short amount of time, it will just take something to remove a few buildings from use.
I can recall a couple shopping malls in the area that had been largely vacant for quite some time. This doesn't appear to be unique since there's just been a cultural shift nationwide where people don't hang out in shopping malls like they used to. These shopping malls are in two different cities but they were both in an area impacted by a large flood. The malls survived
Re: (Score:2)
"Tradesmen"? Construction crews (many of them Hispanic) are "middle class tradesmen"? IT professionals (I'm a retired one) are tradesmen?
Jeez, can't you find a better source of crap?
Do you *really* think they're just going to keep building more datacenters? You don't think the market for them is peaking? And, for that matter, look at the economics of chatbots - I won't call them AI - and look at the increasing flow of articles warning that the bubble's on the way to collapse.
Why exclude data centers? (Score:1)
Seems fairly arbitrary.
Why not exclude whatever else contributed to growth too, so you can say there was no growth at all?
Re:Why exclude data centers? (Score:5, Insightful)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
Marxists claimed Trump would cause a total collapse of the economy. Once again, they were lying.
He's got years yet to destroy it. Once again, cowards are cowards. For example, you're afraid of facts.
Re: (Score:2)
Even though I supported Biden, he didn't create jobs. The U.S. economy created jobs. He merely didn't construct the clusterfuck that el Bunko has created. The U.S. economy sort of runs along just fine without screwing with it. Naturally U.S. companies decided to pull in their plans because they can never figure out what that idiot it going to do them next. The tariffs are enough to drive many small business out of business because they cannot source the parts they need reliably or cost effectively.
Add to th
Re: (Score:2)
Even though I supported Biden, he didn't create jobs. The U.S. economy created jobs. He merely didn't construct the clusterfuck that el Bunko has created.
This is true.
The U.S. economy sort of runs along just fine without screwing with it.
This is nonsense. There has never been a time when it has not been screwed with, so there's nothing at all to support that statement.
Add to that, el Bunko totally screwed the farmers.
He screwed them last time, then there were bailouts, so they figured they could get more bailouts. So they voted for him again. He screwed them much worse this time, though. Many of us tried to warn them, but they dismissed us as a bunch of libtards, because they are stupid fucksticks. Farmers may be smarter than the average nerd thinks, but only because they thin
Re: (Score:2)
I'm not sure "farmers" even exist the way they did 50 years ago. After big business had its way with agriculture, "farmers" are now Mexicans working on someone else's land for sub-minimum wage, mostly.
A lot of people might identify as farmers because their parents had one in the 70s, and a couple of zucchini plants on the porch. But they would be entirely disconnected from the business and economics of agriculture.
Re: (Score:2)
Last I looked there was still a surprising amount of acreage in family farms, even though it's steadily waning. Maybe you're right about mischaracterization, though.
Re:Why exclude data centers? (Score:5, Insightful)
Seems fairly arbitrary.
Why not exclude whatever else contributed to growth too, so you can say there was no growth at all?
While I'm not 100% sure myself, there is some speculation that a lot of datacenters being built today will never actually be used at capacity, and may in fact never be used at all if the AI bubble pops. I think there will be growth in AI at some point, but right now it's fairly unhealthy growth, and doesn't seem long-term sustainable at the rate the tech pushers keep saying. Maybe those datacenters can be utilized in other ways, or will be altered for whatever the next form of automation / AI will be, but I do believe there may be a bit of merit to those saying we've overbuilding for a trend that may end up not being able to sustain the projected current growth trajectory.
Space and power (Re:Why exclude data centers?) (Score:1)
While I'm not 100% sure myself, there is some speculation that a lot of datacenters being built today will never actually be used at capacity, and may in fact never be used at all if the AI bubble pops.
I'd wonder how long these data centers stay productive given how quickly the technology is still changing. What's the expected operational lifespan on a data center? The technology in the buildings would be "current tech" for how long? 3 or 4 years? The building will last longer than that, or at least I would hope so, which can mean the technology can be updated and not need to start from nothing. I don't know how data centers work on this scale, and it could be nobody really knows since this is a lot
Re: (Score:2)
there is some speculation that a lot of datacenters being built today will never actually be used at capacity, and may in fact never be used at all if the AI bubble pops.
So you're saying there will be a lot of cheap data centers available? Cool!
Because it's completely driven by AI (Score:5, Interesting)
If we had competent leadership in Washington they would be getting ready for that but well, you know.
Meanwhile the trade war is hammering manufacturing jobs. It turns out that just cranking tariffs doesn't magically create jobs. It costs us jobs because other countries do retaliatory tariffs cutting off our access to their markets. Corn and soybean Farmers might get bailed out but nobody else does. And fun fact those corn and soybean Farmers aren't salt of the earth they're millionaires who have been buying up land from small family farms going out of business for ages.
Basically it's a classic case of a combination of an unsustainable bubble and a weak economy that was barely recovering from the last round of damage done to it by incompetent stewardship. This next crash is going to hurt
Re: (Score:2)
>> Why not exclude whatever else contributed to growth too, so you can say there was no growth at all?
And what would those other things be that contributed the 0.1% remaining amount of growth?
Re: (Score:3)
Re: (Score:2)
Re: (Score:2)
Without detailed numbers, I suspect it's economist bullshit.
Probably, but what *is* important is that it demonstrates the linkage of Government Overspending, followed by GDP Growth, resulting in Inflation. That is, inflation can be delayed with low GDP Growth, Temporarily.
Re: (Score:2)
Seems fairly arbitrary.
Why not exclude whatever else contributed to growth too, so you can say there was no growth at all?
Because these are investments, not sales. Accounts are being drained and not refilled because the expected sales have not arrived yet. This cannot continue. It feels like late in the DotCom boom and many of remember that time well as well as what happened next.
Re: (Score:2)
because GDP numbers being almost entirely dependent on growth in data centers demonstrates that the economy outside of that one sector isn't in very good shape, which is important to know
US depends on bubbles (Score:5, Interesting)
USA's economy is addicted to bubbles. Without one bubble after another, we wouldn't know what business to be in. We can't manufacturer as extensively as China because we don't have a Xi keeping wages down, we have no fashion sense like France, we don't have Venice-like tourism because GOP scares off visitors (Iraq war era, not just orangeshirts), and we don't have great ruins like Machu Picchu because we tore them down to make casinos.
Our under-taxed plutocrats have giant piles of cash to invest in speculative shit, which often pops, yet we hope a new fad on the block replaces it in time: the Hype Industrial Complex.
Re: (Score:3)
The problem with bubbles is that they need others to believe in them. That may be comming to an end of the US now.
Re: (Score:1, Insightful)
Re: (Score:2)
It's sad to think that whoever wrote this probably thinks they're being clever
Re: (Score:2)
Making an ad hominem attack instead of making an actual argument against their opinion is sad.
Re: (Score:2)
Re: (Score:3)
What do the numbers say?
What the numbers say is that the top 10 companies in the S&P500 (i.e. 2% of those 500 companies) now account for a larger share of its total valuation - 38% - than ever before (since at least 1880).
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.visualcapitalist.c... [visualcapitalist.com]
Of those top 10, which are not tech? Only the smallest 2, Berkshire Hathaway
So what? (Score:1)
A job is a job.
Take out the fastest growing sectors.... (Score:3)
... and the economy isn't growing as fast. You need a Harvard degree to see that?
Re: Take out the fastest growing sectors.... (Score:2, Insightful)
Re: (Score:2)
It's "Investment in information processing equipment & software", which is at least two. It actually appears they're picking and choosing subcomponents from multiple sectors to get this result.
Don't forget... (Score:2)
"Tech giants including Microsoft, Google, Amazon, Meta and Nvidia poured tens of billions of dollars into building and upgrading data centers."
And buying lobbyists! Don't forget the lobbyists!
Still better than expected (Score:1)
The orange felon must be losing his touch. I would at the very least have expected the GDP to decrease.
Give him time (Re:Still better than expected) (Score:1)
The orange felon must be losing his touch. I would at the very least have expected the GDP to decrease.
Why would you think that? Did DJT cause economic problems the last time he was in the White House? I found a history of GDP growth and GDP appeared to be at a fairly steady 3% growth for some time:
https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fdata.worldbank.org%2Find... [worldbank.org]
There's that understandable hit and recovery in 2020 and 2021 because of COVID-19. Because of that outlier event it's a bit difficult to tell just how badly DJT could have impacted the economy otherwise. I do wish that Congress would not have handed POTUS so much unilateral aut
Re: (Score:1)
Then consider this is not even a year into DJT's term, this is still in part a "Biden economy".
biden's economy ended on "liberation day". you broke it you bought it.
Re: (Score:2)
Are you humor-challenged?
what's amazing (Score:4, Insightful)
Job growth in the US is in the toilet along with consumer confidence. Massive inflationary tariffs, the impending death of democracy and rule of law. What's amazing is that the stock market has continued to post gains in the face of all the bad news. It is almost entirely driven by investments in tech, which appears to be immune.
But not everyone is convinced this will hold. Gold hit $4,000/oz recently. I bought some a couple of weeks ago and I'm up about $2k already, I wish I had bought more.
Re: (Score:2)
In the 2000's, I remember hearing the exact same thing about real estate. That's a great sign you're in a bubble - everyone believes that the sector is immune from corrections. And it always is, right up to the point that it isn't.
Re: (Score:2)
Right, but the dotcom bubble collapsed at least partly because so many companies had huge valuations but no revenue or real product.
That's not the case here. The AI companies have real working product, hundreds of millions of active users, and significant revenue. You can argue that they are overvalued but that may only result in an eventual correction.
Re: (Score:2)
"have real working product"
A few of them do, but the product they have is not the product they claim to have.
"hundreds of millions of active users"
We know these user counts are typically inflated, and we know these users are mostly not paying.
"significant revenue"
I'm hearing a bigger share of this is "circular investment" (masturbatory illusion) and not money paid by customers for a product that they use.
Re: (Score:2)
>> we know these users are mostly not paying
Maybe you should show some evidence for all those things you know.
Re: (Score:2)
Evidence: Launch an updated Chrome browser, Microsoft product, conduct a Google search, or open a random page on Facebook. You will be presented with unsolicited AI-generated content. You are now a "user" of AI, and you have paid nothing. Industry-standard practice.
Re: (Score:2)
>> You will be presented with unsolicited AI-generated content
And you know this how? I'm not interested in baseless speculation.
Even if true, you will also have to show evidence that the AI companies are counting that in their user statistics.
Re: (Score:2)
I gave you the steps if you want to replicate my experiment. Or, you can ask anyone who has used a computer within the past 18 months.
Re: (Score:2)
So all you've got is baseless speculation.
Seems Hard To Believe. (Score:2)
It's hard to believe these numbers.
Government spending shows up in GDP. With the MASSIVE recent government spending, I feel like there should be more than 0.1% GDP growth without tech.
That's not to say that GDP growth due to government spending increases is good. Quite the opposite, that sort of artificial GDP growth is very bad. But, I do feel like the over spending from the last two administrations has been done with significant specific intent to drive GDP.
Re: (Score:2)
What theyâ(TM)re really saying (Score:2)
What theyâ(TM)re really saying is that we are in a massive recession?
Without the AI bubble the growth was 0.1%? (Score:2)
When will this bubble start and make us come to our senses.
My toiletseat now even has AI features. =/
Re: (Score:2)
When the Internet bubble burst.
Any moment now, keep the faith strong my young padawan.
I don't believe ... (Score:2)
... that the people investing in the AI bubble would be investing in nothing otherwise.
Nor do I believe that they would just be letting their money sit in a Scrooge McDuck money bin.
Re: (Score:2)
Exactly. Gdp growth doesn't happen in a vacuum. Some other sector would be absorbing some or all of that spending.
Left out a few words (Score:2)
"Renaissance Macro Research estimated in August that the dollar value contributed to GDP growth by AI data-center buildout had surpassed U.S. consumer spending for the first time. Consumer spending makes up two-thirds of GDP. Tech giants including Microsoft, Google, Amazon, Meta and Nvidia poured tens of billions of dollars into building and upgrading data centers."
The US GDP is around $30 trillion. Consumer spending is two-thirds of the $30 trillion, or around $20 trillion. AI data-center spending is in
I forsee... (Score:2)
Re: (Score:2)
Lots and lots of cheap VPSes offering cloud servers and storage from all of these soon-to-be future useless AI datacenters.
That may be true to a certain extent, but I submit two counterbalances to this:
1. VPS workloads aren't usually helped by GPU acceleration; even if DigitalOcean or Hetzner got these datacenters in a fire sale, they'd still be sitting on a pile of GPUs they would either have to leverage or sell.
2. While VPS companies will undoubtedly continue to have their workload quantities increase over time, if the AI tech bubble bursts, they are likely to end up being on the working end of that at some level, too. I'm su
Harvard economist Jason Furman (Score:2)
Is going to be on Trump's hit list for pointing this out.