Why? It's hard to see what they're good for outside of AI data center workloads. Your average member of the general public arguably doesn't have much use for a "graphics processing unit" that isn't actually capable of processing graphics
This tracks with recent reporting that AI debt issuance fell 50% month over month in September[0]
> Tech companies raised $23 billion in AI-related debt globally in September, less than half the amount raised in August, according to Morgan Stanley data covering public bonds and private placement transactions.
> The decline marks a continued drop from June, when companies raised a record $113 billion for AI infrastructure including data centers and chips. In the US investment-grade market, where blue-chip tech groups borrowed around $306 billion between January and August, AI-related bond issuance stopped completely last month.
To make senes of the drop we need to know how much they expected to raise in September. Were the expecting to raise $50B but only raised $23B? Or were they expecting to raise $25B but only raised $23B? That's an important detail. Further, "bond issuance stopped"... was that intentional or not? It's like we're getting half the story and then have to use our imagination and biases to make the story as bad as possible.
Loans are sold off, which means that they can raise more money to make loans. Each time you do that, you create more money out of thin air, which is profit.
The problem is, it only works if everyone believes.
THe problem with 2008 is that people stopped believing and demanded banks settle immediately. which then meant that they needed to raise cash, which meant looking at their assets, which meant looking at the risks of those assets, and then it went to shit.
That’s true, but before you try to shift the blame onto the homeowners, remember that these mortgages were intentionally issued to extremely risky borrowers, and then that risk was intentionally obfuscated when the loans were collateralized. The fault cannot be laid entirely on the borrowers, which is what I sense you’re trying to do here.
Investors bought a lie packaged by the financial institutions. They shouldn’t have believed in the first place because the risk was junk-level but not sold as such.
No but thats the point, they were always in trouble. sub prime were sub prime.
Everyone thought they were ok because they didn't look at, or were not shown the, default rate of their instrument. The risk profile was deliberately obscured, or the people wilfully ignored it.
yeah, either they discount it enough to look worthwhile to someone, or they keep it and take the risk. Risk assessments are inherently subjective, so both parties can think they're making a deal.
>We lent £1,000 to John Sozzleby at 10%. We quantify the risk level as “iffy”, which is 42.6 qualm points above "dodgy". and are willing to sell you the debt for £700. its a great deal!
> They sell down debt when they are worried about concentration, which is exactly what "approaching internal limits" means in plain English: too much of this bank's balance sheet is now one bet on one sector holding up.
Which is correct, but kind of weird because being confident in a credit and being worried about concentration are not really mutually exclusive things. You can be pretty confident in an investment but not want it to constitute a majority of your portfolio. (Bankers know that even very strong credits are never 100% guaranteed to pay out.)
Let's say I offered you the chance to pay $100 to get $120 back if at least one coin toss in five came up heads. That's a pretty good bet, right? But would you bet your house on it?
Not sure I agree. The credit default swaps have been extremely expensive lately, meaning that someone is already thinking about an AI bubble popping that will take private credit down which will make banks take a serious hit (11% of their outstanding loans are to linked to AI). On top of that we are in a time where the economy is propped up basically by AI spending and delinquencies on all sort of debt are at historically high rates.
Maybe it's because I've lived through 2007, but the time to panic is when ~100 year old institutions disappear overnight (Bear Stearns)
Do we even have one bankruptcy yet?
I'm keeping an eye on Coreweave and Oracle. But they haven't collapsed yet, not as of today anyway.
If it's anything like 2007, it's going to be the one we aren't watching that collapses first. Those in the open about their debts and situation are often more honest and might be more solid than other unknown players.
You absolutely should post about it when you short for good cause - there was a fund that earned it's money that way. Do research, find adverse fundamentals or malfeasance that are underreported, short the security, then publish your findings and let the market do its thing. IIRC, it was them that publicized that Nikola (Electric truck startup) faked their demo video by rolling a non-operative prototype truck down a stretch of slightly inclined highway.
I remember in Summer 2007 feeling that there was an impending implosion coming based on the shady practice of using 2nd mortgages as downpayments - it had become a common practice. Started doing a bit of shorting (not too much), made a little, lost a little and then sort of lost interest in doing it just about the time things fell apart in '08. All that to say that these things can take a lot longer than you'd expect. You can look at the situation (like the circular funding as well as this article) and decide that yes it's not sustainable. But it can sustain longer than you expect.
Yeah. The thing is, we were being told that was a problem in late 2004 (this is when Max Keiser was writing and talking about the risks to Fannie Mae and Freddie Mac; he and others saying this were written off as cranks, even though they laid out a pretty solid scenario that came substantially true.)
So yes, Oracle is likely a canary in the coal mine.
It can only really be seen as an indicator of hidden risks to the banks from exposure to the SPVs.
But how long it will take before the miners start dropping dead from the fumes, who knows.
Funny cause I also feel that a failed Anthropic IPO can trigger the domino effect. And I also t
Feel this is why them and Open ai postpone so much. And shows how smart Musk was with rushing his IPO.
I can’t short anything due to this. I know something smells but I can never predict when the system will fall apart. And especially recently, the markets seem way too “vibes based” as if even normal securities are all meme stocks now.
With these things it's not about spotting the signs but rather predicting the timing. Everyone and their dog can see the signs but that's not enough to be actionable
Who is building the infrastructure to resell the gpus from failed data centers if / when this collapses. It would be a waste for everything to get wasted
The GPUs used in data centers cant really be used outside the same context. They cant connect to standard PCs or servers. They are highly specialized. The resale would need to be at the server or rack level for the same use case.
There have been some examples of folks ripping off the SMX interface and rewiring the card for PCI-E but I'm not sure how efficient that would really be or what kind of market there could be
Yes, they're "just" pcie, though units like H100 and newer (or did it start with pascal? i really don't know) use a proprietary connector called SXM, but it's always just pcie lanes underneath. In some versions there are interposer and mezzanine boards, but again, just pcie with associated bits and bobs like retimers and pcie switches. Not impossible to integrate, just annoying.
You can order the SXM carrier boards on eBay/Alibaba today.
> You wouldn’t repurpose these for home gaming, it would still be for llms
When small businesses, home-labbers and prosumers have cheap access to potent GPUs, there will be a Cambrian explosion of grassroots creativity and innovation in areas that are "beneath" AI labs expansive ambitions; I want to see the OBS-equivalent of LLMs. Hell, I'd love to see what some broke artist can do with OBS, an H100, and an open video model.
Good luck powering a 400kw rack in your basement, if you are looking to do frontier model type stuff. Just a single machine can run in the 150kw range these days for the less exotic stuff.
It can certainly be parted out with proper motivation but it’s going to take a whole new ecosystem to support these in your typical home lab. Even then 20-30kw slices are electric car level energy consumers.
> Good luck powering a 400kw rack in your basement
Wut? I'd gamble on 1 million curious people with 1 SoTA GPU each over one AIaaS company with 1M GPUs when it comes to discovering things that are personally useful to me.
I mean it's not that hard. It's just hard if your frame of reference is "can I download it within one click on my $6000 Macbook".
This whole subthread is just sad. "Oh no, it's not the same form factor. It is useless".
Where's the "Hacker" in "Hacker News" there.
It's a shame that we need to defer to "China" as a shorthand for "actual engineers" just because our own engineers lack imagination and drive that would validate being given that title in the first place.
> It's a shame that we need to defer to "China" as a shorthand for "actual engineers" just because our own engineers lack imagination and drive that would validate being given that title in the first place.
There are a lot of reasons other than skill why China is the way it is. Culture and locality are big ones. I don't know a whole lot but I disagree that any "actual engineer" could easily come up with a sustainable business for this stuff in the US -- for example, harvesting and processing used parts for their components is a hugely laborious process, and China has a huge culture of recycling and reworking that simply does not exist in the US (and would largely be socially unacceptable to introduce!). Take for example the pre-modded 4090s, or the BGA desktop motherboards I mentioned, or any number of other products that are based not merely on sheer engineering. It's not exactly engineering on its own that causes these things to happen, but more the fact that nearly anybody can just throw together an operation like that and have tons of workers available to make it sustainable.
They do not have display outputs, but, at least with my 4090, if I plug the monitor into the iGPU instead, Windows still renders frames on it and just pushes them to the GPU that has a monitor attached.
So the lack of output ports of those DC GPUs might not prevent that?
Or maybe there are other limitations there I am not aware of, of course. Maybe the silicon cannot do the stuff needed for games or something like that.
I think the data centres stay, continue to be used for LLMs but end up not getting the intended returns and possibly ending up in the hands of creditors who don't end up with the sort of money they expected.
I think at least we Europeans will really want these data centres once prices become reasonable though. We really need training infrastructure.
I saw an article a couple of days ago on one of the popular sites (ft or Bloomberg or something), about how now is a great opportunity to purchase private company credit.
Guess this is the submarine https://paulgraham.com/submarine.html
Private credit is far different than the debt of a public ally traded company. It is far riskier as those companies can't easily tap the credit or equity markets for new money if they get into trouble the way a public company can.
But the debt we are talking about is not Private credit debt, its completely different animal with far different risk characteristics
The problem with private credit is that most of the valuations are not publicly tested.
what do I mean by that? oracle is publicly traded, which means they (should) report profit loss and other key metrics in a standardised way. This means that one company's metrics should be comparable to others. Their value is assessed by the second on a public market. This means that anything that affects thier value (real or imagined) is reflected almost instantly. (same goes for debt)
Private credit, and the stuff its invested in, is much murkier. THey do not have the same public reporting rules. There is no testing of value. Its in the interest of the owner of that debt to talk up the value of the assets they hold.
This means that a fund might invest in 10 companies. Those companies might be doing well, or they might not. Because information is limited, and there are no public trades, the value of those companies are only ever tested at sale, or in a crisis. (ie they fold, or do something so public that they can't avoid the consequences.)
THe problem then is that because information is limited (was it a one off, did the owner hide the problem, did they use the wrong investment strategey, is the rest of the portfolio also over valued by 10x?) causes people to panic, and withdraw money.
Because private credit is not that liquid, it means that if people withdraw too much money, they either have to say "no we have no free cash" (lehman brothers insolvancy), they have to take out bridging loans (expensive and very short term) or sell assets (fire sales rarely get good value and shrink the credit supply)
All of which causes a panic, and people look at places like softbank, see the leverage and think "oh fuck thats not worth anything" and the bubble pops.
The CDOs on some loans are getting to the point where it’s too expensive to even bet that the loans will fail as the CDO market seems to have concluded the loans will fail.
> The CDOs on some loans are getting to the point where it’s too expensive to even bet that the loans will fail as the CDO market seems to have concluded the loans will fail.
Did you mean CDS? As your sentence doens't make sense as its written.
Do they normally spend six months trying to offload a single tens-of-billions-of-dollars debt at a discount that is associated with one of the USA's longest standing blue chip tech stocks (whose debt rating has fallen to one level above junk)?
The sirens might not be wailing yet but some of the lights are flashing.
Isn’t this a perfectly normal thing banks do with all large debt? Any bank underwriting any $40b loan was always going to syndicate it.
Not to do so would be reckless.
What is unusual about this is that syndications usually clear in a few weeks but this has been stuck in distribution for six months and still not closed.
What is unusual about this is not that it is a single debt of that size but that it is Oracle that is the poor quality indicator.
And it's just one of them: they know that there is about three trillion dollars of this risky shit out there either in the hyperscaler/FAANG balance sheets or off the books, so they are going to struggle to sell this, and Oracle's position gets worse every day.
We are getting a lot closer to the Margin Call crisis meeting moment where one bank decides it cannot assess the risk of their share of it, are we not?
But this time it is happening in a world where retail investors, lawyers, government officials, elected representatives are much more likely to have seen Margin Call and The Big Short.
Listening to Ed Zitron on various podcasts, there are several issues with Oracle stocks:
1. they are one place above junk;
2. they are being treated as effectively junk;
3. if they get classified as junk, any pension portfolios holding Oracle stock must sell as they can't hold junk stocks.
If that happens then there will be a cascade of effects. Not just in the AI/datacentre sector but elsewhere, as Larry Ellison holds a stake in Paramount/helped finance the Paramount/WB merger.
>3. if they get classified as junk, any pension portfolios holding Oracle stock must sell as they can't hold junk stocks.
Wouldn't hedge funds or PE recognize this arbitrage opportunity and eat this up, making this a non-issue? Vastly simplifying this: suppose "junk bond" means the chance of default is greater or equal to 10%. Of course it's bad if the risk of oracle going bankrupt edges up from 9% to 10.1%, but we shouldn't expect pandemonium from the 10% cliff edge, because there's plenty to be made by selling all the bonds that pension funds are obligated to sell, because now they have higher yields (yields move inverse to bond prices).
I guess, maybe, but perhaps the risk that GP is commenting on is that it could be it's own form of Juicing.
I do sincerely doubt, even with the big shifts that AI enables to have companies 'move away' from Oracle, that Oracle will default.
On the other hand, There's still an elevated chance of decline.
IOW, Those stocks might move from pension type plans to different baskets as a result, however there is a non-negligible risk of some (not total!) loss in the 5-10 year term.
Will add, I would -hope- that the same pension rules around selling questionable stocks, would also have rules limiting what level of exploitation is possible.
I think it's only a matter of time, like the build-up to the financial crash in 2008 or the dot-com bubble. There's currently a huge amount of money being bet on the big AI tech companies without any real sense of return. And the companies like Oracle that are funding data centre build-out are sitting on unused GPU stock.
Recently, the Australian AI firm failed to secure an IPO [1], [2].
OpenAI is looking to go for an IPO despite shaky revenue [3], [4], [5], resulting in a stock drop for companies like Oracle invested in OpenAI [6].
Anthropic is in a similar situation: looking to IPO despite shaky finances [7], [8], [9].
The companies involved are all investing in each other and trading money between them leading to circular investment [10].
Equity shares do not have ratings, bonds have ratings. Everything else is correct, if their bond ratings slip one more level they will be rated junk which may force some holders to liquidate if they’re not allowed to hold junk bonds.
The longer a clearly deteriorating credit like Oracle remains nominally investment grade, the worse the selloff eventually becomes. It’d be a false economy to try to drag it out, especially when investment grade institutions are already pricing Oracle at crossover and preparing to get the hell out of dodge.
The trigger in Margin Call was not that they could not assess the risk, it was that they were improperly assessing the risk, had taken too much risk onboard, and the downside of the risk was more than the value of the firm.
Yes, and no. The film is not really about that one day or that one discovery.
Much of what Margin Call is about has already happened in the few days before the movie (the events that precipitated the firings but are not really discussed).
Sullivan is talking about the music slowing, saying "if the music were to stop…". He's suggesting that it is bad, in the context of those few days, and action is needed.
Tuld saying he fears he "can't hear anything" is about much more fundamental uncertainty — his own belief that no set of numbers could be useful and that the actual situation required acting on his own gut instinct is that it was much, much worse.
In short, the point is they did not have any confidence they could ride out the consequences of the preceding, largely un-discussed few days, because they were beyond their parameters, and that there was no time to assess the real risk.
> Matthew Moniot, co-head of credit risk sharing at Man Group, told the Financial Times the deal sizes involved are "out of scale to anything we've thought about, ever."
The short answer is to free up money to make other investments.
The longer answer is that banks generally have policies about how concentrated they can be in individual borrowers, industry sectors, geographies, etc. That is why you rarely see a single bank take all of a very large corporate loan. They syndicate it out.
They take a cut / fee for financing the debt deal and hold the debt short term. They can't do another deal (simplifying the situation) until they offload the risk. Their business is in their percentage for arranging deals, not in getting paid back over 50 years.
In these times, what other deals are out there. AI has sucked all the air out of the room, difficult to imagine anything getting similar investment over the next decade.
its a different business model? IBs dont generally hold debt to maturity. they underwrite the debt then gradually sell down in the weeks/months following the deal. their business is making money on the spread/underwriting margin.
This is common. Selling debt is normal. If there is money to be made by selling the debt and going back to issuing more loans, they’ll try to find a buyer.
Everyone is so primed to believe that all things related to data centers are bad that they’ll assume this is a special event. The person writing the headline knows that.
To be fair to banks this is what they do with almost all debt their underwrite.
They hold it to help stabilize the market for it while its being sold and then offload it once the market has been established.
Nothing out of hte ordinary here, but to be fair the original article isnt' a financial site so they would have no real idea about how this work behind the scenes.
This doesn't seem like a story at all given how robust this debt is trading.
- I still have a mortgage on a property but don't live there so the mortgage bill gets sent to a different address (b/c rental property)
- Company A originated the mortgage and sells it to Company B
- Company B completely ignores the mailing address and starts sending bills to the original property address
- only through blind luck did I realize that the mortgage bill was being unpaid
This seems like basic, basic bookkeeping but I guess it's also a reminder that "reality has more detail than people realize" and that there are plenty of opportunities for improving systems.
One of those myths too many programmers believe: All street addresses can accept US mail. Not necessarily the issue in this comment, but when somebody gives you both a shipping address and a mailing address and you send mail to the shipping address, the fault is yours if it doesn't arrive.
Another: A particular street address only maps to one person, entity, or family. Counterexample: Giant skyscraper with a zillion tenants, each in a different suite. Or a giant apartment building. Etc.
Still another: One particular suite address only maps to one person, entity, or family. Counterexample: Any of thousands of commercial mailbox-rental / package-shipping stores in US strip malls, which tend to have suite numbers (within the mall) and possibly hundreds of tenants, all at one suite.
One of those facts that mortgage owners refuse to internalize: the party responsible for the payments is not necessarily living on the property. There is absolutely no reason in this century to assume that the mailing address for real estate mortgage related correspondence is the address of the property. Related: the assumption that if there are two names on a mortgage with given names that are commonly understood to be gendered, the parties are legally married.
You should be able to go to the USPS and do a change of address for yourself. That way every time mail for you is sent to your tenant, it will go to you.
I just moved. Moving a business address was hard. For an individual, the only requirement ai saw was that I pay a small fee with a credit card registered to one of the addresses.
This is a great idea. I'll try it but I'm not sure it will work. I believe there is a time limit on the forwarding and it looks like not all mail is forwarded, including financial documents.
Ar some point I had all the bills on my calendar and a event to check if the company confirmed the payment. It saved me at least a handful of times.
I confess that I am not that diligent anymore, but I also have scheduled things in a more reliable way that always notifies me on payment. But the other day the bank skipped one of the payments because some information on the destiny account have changed but nothing was shown.
It is always annoying to stop what you are doing to do this
This is standard. Origination and servicing are typically handled by different companies (or parts of the same company). Usually the Origination side only handles short term loans and doesn’t want stuff hanging around longer than is necessary.
> My mortgage has been sold almost a half dozen times already.
As a Canadian, other than the absolutely insane rate of school shootings, this is probably the next-wildest difference I can think of.
Canadian mortgages are held and stay with the bank that approves the mortgage. The only way the mortgage goes to a different lender is if the owner goes to that other bank to renew the mortgage. It just boggles my mind that mortgages can just be assigned willy-nilly without the owner’s consent.
"This doesn't seem like a story at all given how robust this debt is trading."
That directly contradicts specific points made in the story, about specific Tier 1 banks and specific amounts of debt not finding buyers.
"Nothing out of hte[sic] ordinary here"
Matthew Moniot, Man Group: "the deal sizes involved are 'out of scale to anything we've thought about, ever.'"
Hard to compare your model with his, since he doesn't have one. When quotes from named credit risk guys from named finance organizations sound like angsty Gen Z, it's not ordinary.
In any case, the story is covering something significant. There is an important test occurring, and we're all going to learn the results, sooner or later. It could all turn out fine, but I don't know. I don't believe you really know either.
Well I'd agree the size of these deals are far larger than any debt deals I can remember in the past 10 years.
But that has nothing to do with the scare tactic headline that says banks are trying tot offload debt. As I said this is exactly how banks work and how they treat all debt they underwrite.
I believe the implication is that if AI is really The Second Coming (or First Coming, if you prefer), then banks would want to keep the debt and the potential gains of said debt. Or rather, there should be an overwhelming demand for said debt.
Where it's going to pinch is any incompletely funded AI datacenter build-outs.
If the market is already crammed with sector-concentrated debt (because of the amount already issued), finding buyers for more is going to get increasingly hard.
At some point everyone big already has enough exposure and isn't interested in the risk of more.
The only thing that keeps the party going is being able to drain that debt off bank sheets, but it's hard to see a ton of interest in fixed rate, capped upside exposure to a high risk industry in an environment of rising central bank rates.
If the of fate of the AI boom is so fragile that a digital newsletter site's headlines are a meaningful concern to you, perhaps you have some things to ponder. The story itself names names, quotes verifiable statements and makes reasonable inferences.
And this is supposedly high grade debt we're talking about: JPMC is spending months on end shopping around a few $10's of billions, and having to discount it. Never mind the far greater sums on the books of vastly less credible outfits, complete with meme names that always age poorly when things go pear shaped.
How a bunch of highly speculative moonshot big tech stuff has managed to be financed by Tier 1 financial institutions and high grade debt, as opposed commonplace equities, is a question for the ages, and we may all yet find ourselves suffering the answers.
> This doesn't seem like a story at all given how robust this debt is trading.
FTA: That is exactly why CoreWeave-leased paper prices at 7.00% to 9.875%, against the 5.70% to 7.875% that hyperscaler-leased bonds command, per bond market data compiled by CCIR.
Just so were clear: Core weave paid Micron and NVIDA 80 percent margins to buy equipment at a 9 percent lending rate so it can sell it to providers who are subsidizing tokens at 20x their own API prices (that are likely also subsidized).
Could it all work out? It could. But it does not look like there is the financing to keep building - thats going to impact prices and the moment that happens this debt looks even more risky.
What you've written is the demand for debt. What's the supply like? You can't conclude demand exceeds supply without knowing both sides, and the US has printed more money than any other time in history.
The debt (bonds basically) is sitting on the books of investment banks. The article here is 5 days old - that "9 percent" figure is already out of date.
Here is today: Shares of AI cloud provider CoreWeave (CRWV) dropped over 15% Wednesday, a day after venture capitalist Chamath Palihapitiya posted on X a chart that showed the yield Nvidia-backed firm's 2032 bonds hit 15% yesterday as a sell-off brought prices into distressed territory. SOURCE: https://finance.yahoo.com/markets/stocks/articles/nvidia-bac...
I can conclude that supply is far exceeding demand because of the rates attached.
Bonds are a weird product. As long as the issuer honors them (remains in business) they are a guaranteed return.
In this case CoreWeave has sold ~4billion in 5 year bonds at a 9% rate. If you buy one for 1000 bucks you will get 2x45$ payments a year for 5 years. Then at the end a 1000 cash payment.
For the sake of argument let's say you bought this earlier this year. You now have an asset worth 1000 dollars. If it is a good bond (A US treasury or Municipal Bond) you can borrow against it from lots of place - you have something worth 1000 dollars on your "books" (accounting). For the earliest sales (you) core weave has already made a 45 dollar payment.
You now have a 1000 dollar asset and 45 bucks in cash. IF you hold to maturity and keep all the money you get 1000 bucks back and 450 bucks in interest.
What happens if you want to sell it? Maybe rates went down a lot, so that 9 percent is huge, and the AI boom is going gang busters so that 9percent looks like a good deal. I offer you 1100 for the bond. You now have 1145 dollars in cash.
What happens if you need money, and rates are the same and CoreWeave looks the same? I will offer you 980 for it... You now have 1025 bucks in cash... and I get the rest of the payments (45$x9) and 1000 bucks back at the end. (This has tax implications).
What happens if you want to sell cause you need money AND, you think CoreWeave is gonna burn, or this "ai thing" is a bust. Well it's likely that I think that too. the bond is only ever going to pay 45X9 more payments but I want a higher rate. I say to you "I'll give you 800 for it" (a 15% rate), cause CoreWeave is fucked. You need the cash so you sell it for that much and have 845 bucks. Thats a loss of $155.
Let's say you bought a billion of those bonds, and expected to walk away 1.45 billion in 5 years - but you sold them all off today and LOST 155 million.
Someone owned a LOT of 9% Core Weave bonds and either found a BETTER investment -OR- they wanted out of CoreWeave and no one had the money or desire to own them. The bond buying market is announcing that "We don't want any more CoreWeave Debt".
There are a lot of problems with CoreWeave: those 5 year bonds it sold, well the contracts to support them (15 billion a year in revenue on 60 billion of debt) only run for 3 years. At current rental rates for long term commitment NVIDA hardware had a 3 year pay off term... But thats not power, the building, staff, the business was (is) marginal.
Who was selling the debt? Well my guess would be a chunk of it was JP Morgan (who set up the bonds to start with). Dumping it right after you announce your numbers (they were a minor miss) and lower your target for CoreWeave's stock would be an ideal time. But that is PURE speculation.
So 9% is the yield and people demanding 15% for new bonds instead of the previous 9% means they lack confidence you'll be around to pay back the original principle?
Are you saying there are existing bonds at 9% being sold for some amount of money (possibly even less than the original principle?)
How does selling the $1k bond at $800 translate into 15%? The bond was originally worth $1k+$45*10=$1450, now $45*9 ($1405?), so if you buy it for $800 that's approximately 85% of its original value ($800+(9*$45)/$1405) and thus a 15% decrease? How does that relate to the original 9%?
Is this about what existing bonds are being sold for or new bonds?
Presumably a us treasury bond yield is basically the same as inflation?
> Are you saying there are existing bonds at 9% being sold for some amount of money (possibly even less than the original principle?) ... Is this about what existing bonds are being sold for or new bonds?
The bonds were issued earlier this year. Their sale is a loss for someone, and NOT a small one.
> How does selling the $1k bond at $800 translate into 15%?
1000 5 year at 9 percent 5x90(450) + 1000 back.
If you need to dump it for 800, the bond still pays out 1000 in the end, The issuer still pays out 5x90(450). The 200 bucks I pick up at the end gets spread out over the 5 years so it looks like 5x130 (650+800).
That would be 16 percent, but all the numbers for the sake of sanity got rounded from their real ones. 9.xx rate, 8xx resale price, and 14.9 (closing on 15) rate.
It gets calculated that way because if you need to sell more new debt it only makes sense to know how to price it.
> Presumably a us treasury bond yield is basically the same as inflation?
Sometimes, often not. But Treasury Bonds have quite a few advantages. No state tax on the interest payments. They are almost as good as cash, so you can borrow against them at a pretty good rate (Better than a HELOC and way better than a credit card). Lots of institutions will give you a line of credit for up to 90 percent of the value of the bond. You can also borrow against them for margin trading (60percent? better than stocks) -
1. Bank gives you $X to buy an asset and burdens you with debt of $Y, which is much more than $X as it includes interest.
2. Bank sells the debt at $Z because holding the debt is risky and the bank would rather have cash than debt. Its counterparty would rather have more cash over time and is fine with the risk but does not want to bother with the whole process of closing the loan. $X < $Z < $Y.
3. Profit! Bank spent $X to get $Z, which is greater than $X.
If it were trivial to find buyers, wouldn't that be a sign that they priced the debt too low? It should be just hard enough to find buyers that your lost sales are less than the extra price you are able to extract to the sales you make.
Most backs don’t want to take the loss and legally they don’t have to book it until the asset is sold. So banks often get stuck with bad loans because selling at market price requires them to take a write down
Sorry if this appears obtuse, but what exactly is an “AI Data Center”?
Is it a data center with nothing but GPUs, that does nothing but inference all day and all night?
Is it a data center custom-built for only AI companies to use?
Do these “AI Data Centers” have data storage? VPS? …
Or, as I suspect, is “AI Data Center” simply a fake media buzzword that really means “Normal Data Centers” but with political baggage and stigma built-in.
It's a data centre being build specifically to run LLMs (and possibly other ML models). As such, it will be focused on NVIDIA servers with large numbers of inter-connected GPUs (so they can run the 2T-4T parameter models from OpenAI, Anthropic, etc.) [1].
It's a different build-out to a data centre that can host video/YouTube content (fast CPU, no GPU, large number/size of hard drives), or that can run web servers (fast CPUs, low to average hard drive capacity).
> Or, as I suspect, is “AI Data Center” simply a fake media buzzword that really means “Normal Data Centers” but with political baggage and stigma built-in.
No you're absolutely right. The only difference that I can tell is power requirements are legitimately higher because GPUs consume more electricity. More cooling is also needed because there's more heat generated. Aside from those two things, it's a plain old datacenter that companies have been building at break neck speeds since like 1996. Now that I think about it, how come there's no panic about additional Internet backbone capacity needed to support these "AI datacenters"?
This sounds like fearmongering. Banks aren't in the business of collecting on debt. They are in the business of writing loans they can sell to people who collect the debt. This is pretty standard business in the 2020s. Not defending data centers or AI or whatever, not denying that there is an AI bubble. Just saying that this is how business is done now.
185 comments
[ 0.19 ms ] story [ 26.7 ms ] thread> Tech companies raised $23 billion in AI-related debt globally in September, less than half the amount raised in August, according to Morgan Stanley data covering public bonds and private placement transactions.
> The decline marks a continued drop from June, when companies raised a record $113 billion for AI infrastructure including data centers and chips. In the US investment-grade market, where blue-chip tech groups borrowed around $306 billion between January and August, AI-related bond issuance stopped completely last month.
0: https://www.investing.com/news/stock-market-news/tech-compan...
Sure. Sometimes they cause global financial crises as part of that process.
https://en.wikipedia.org/wiki/2008_financial_crisis
[0] https://en.wikipedia.org/wiki/Tulip_mania [1] https://nvidianews.nvidia.com/news/nvidia-partners-with-apol...
Loans are sold off, which means that they can raise more money to make loans. Each time you do that, you create more money out of thin air, which is profit.
The problem is, it only works if everyone believes.
THe problem with 2008 is that people stopped believing and demanded banks settle immediately. which then meant that they needed to raise cash, which meant looking at their assets, which meant looking at the risks of those assets, and then it went to shit.
Investors bought a lie packaged by the financial institutions. They shouldn’t have believed in the first place because the risk was junk-level but not sold as such.
Everyone thought they were ok because they didn't look at, or were not shown the, default rate of their instrument. The risk profile was deliberately obscured, or the people wilfully ignored it.
they were valuable until it wasnt.
Lots of names to obfuscate the same financial yada yada.
/s
>We lent £1,000 to John Sozzleby at 10%. We quantify the risk level as “iffy”, which is 42.6 qualm points above "dodgy". and are willing to sell you the debt for £700. its a great deal!
... and from the linked piece ...
| Banks do not sell down debt they are confident in
... I have no claim to expertise but both of these can't be true.
I wouldn't worry about it anyway, you can be sure when the day comes "they'll be to big to fail".
> They sell down debt when they are worried about concentration, which is exactly what "approaching internal limits" means in plain English: too much of this bank's balance sheet is now one bet on one sector holding up.
Which is correct, but kind of weird because being confident in a credit and being worried about concentration are not really mutually exclusive things. You can be pretty confident in an investment but not want it to constitute a majority of your portfolio. (Bankers know that even very strong credits are never 100% guaranteed to pay out.)
Let's say I offered you the chance to pay $100 to get $120 back if at least one coin toss in five came up heads. That's a pretty good bet, right? But would you bet your house on it?
When the debt in 2007 sucked, it crashed the economy. But when the price was good in 2006, everyone celebrated.
Note: prices are getting worse for bonds today but nothing panic worthy yet.
I am panicking a bit TBH
Do we even have one bankruptcy yet?
I'm keeping an eye on Coreweave and Oracle. But they haven't collapsed yet, not as of today anyway.
If it's anything like 2007, it's going to be the one we aren't watching that collapses first. Those in the open about their debts and situation are often more honest and might be more solid than other unknown players.
Which is why I hate when people criticize some obviously overvalued shit, and the response is "why don't you short it".
Sometimes the best way to be safe of something about to explode is to just get away from it.
So yes, Oracle is likely a canary in the coal mine.
It can only really be seen as an indicator of hidden risks to the banks from exposure to the SPVs.
But how long it will take before the miners start dropping dead from the fumes, who knows.
https://ebay.io/m/imDgiL
Cooling is usually what kills you with server gear at home. I don't have a cold aisle, or a case that flows that much air.
No video outputs would probably be my biggest annoyance.
Yes they use pcie electrically, but its not the same slot shape and there is usually carrier boards
You wouldn’t repurpose these for home gaming, it would still be for llms
This is already happening.
You can order the SXM carrier boards on eBay/Alibaba today.
> You wouldn’t repurpose these for home gaming, it would still be for llms
When small businesses, home-labbers and prosumers have cheap access to potent GPUs, there will be a Cambrian explosion of grassroots creativity and innovation in areas that are "beneath" AI labs expansive ambitions; I want to see the OBS-equivalent of LLMs. Hell, I'd love to see what some broke artist can do with OBS, an H100, and an open video model.
It can certainly be parted out with proper motivation but it’s going to take a whole new ecosystem to support these in your typical home lab. Even then 20-30kw slices are electric car level energy consumers.
> Good luck powering a 400kw rack in your basement
Wut? I'd gamble on 1 million curious people with 1 SoTA GPU each over one AIaaS company with 1M GPUs when it comes to discovering things that are personally useful to me.
This whole subthread is just sad. "Oh no, it's not the same form factor. It is useless".
Where's the "Hacker" in "Hacker News" there.
It's a shame that we need to defer to "China" as a shorthand for "actual engineers" just because our own engineers lack imagination and drive that would validate being given that title in the first place.
There are a lot of reasons other than skill why China is the way it is. Culture and locality are big ones. I don't know a whole lot but I disagree that any "actual engineer" could easily come up with a sustainable business for this stuff in the US -- for example, harvesting and processing used parts for their components is a hugely laborious process, and China has a huge culture of recycling and reworking that simply does not exist in the US (and would largely be socially unacceptable to introduce!). Take for example the pre-modded 4090s, or the BGA desktop motherboards I mentioned, or any number of other products that are based not merely on sheer engineering. It's not exactly engineering on its own that causes these things to happen, but more the fact that nearly anybody can just throw together an operation like that and have tons of workers available to make it sustainable.
They do not have display outputs, but, at least with my 4090, if I plug the monitor into the iGPU instead, Windows still renders frames on it and just pushes them to the GPU that has a monitor attached.
So the lack of output ports of those DC GPUs might not prevent that?
Or maybe there are other limitations there I am not aware of, of course. Maybe the silicon cannot do the stuff needed for games or something like that.
I'd much rather be in the recovery/asset remarketing market rather than the primary/reseller one when this thing pops, for sure.
I think at least we Europeans will really want these data centres once prices become reasonable though. We really need training infrastructure.
But the debt we are talking about is not Private credit debt, its completely different animal with far different risk characteristics
I agree with JPM that now is a good time to invest in public debt.
what do I mean by that? oracle is publicly traded, which means they (should) report profit loss and other key metrics in a standardised way. This means that one company's metrics should be comparable to others. Their value is assessed by the second on a public market. This means that anything that affects thier value (real or imagined) is reflected almost instantly. (same goes for debt)
Private credit, and the stuff its invested in, is much murkier. THey do not have the same public reporting rules. There is no testing of value. Its in the interest of the owner of that debt to talk up the value of the assets they hold.
This means that a fund might invest in 10 companies. Those companies might be doing well, or they might not. Because information is limited, and there are no public trades, the value of those companies are only ever tested at sale, or in a crisis. (ie they fold, or do something so public that they can't avoid the consequences.)
THe problem then is that because information is limited (was it a one off, did the owner hide the problem, did they use the wrong investment strategey, is the rest of the portfolio also over valued by 10x?) causes people to panic, and withdraw money.
Because private credit is not that liquid, it means that if people withdraw too much money, they either have to say "no we have no free cash" (lehman brothers insolvancy), they have to take out bridging loans (expensive and very short term) or sell assets (fire sales rarely get good value and shrink the credit supply)
All of which causes a panic, and people look at places like softbank, see the leverage and think "oh fuck thats not worth anything" and the bubble pops.
The world is a napkin
/s
'nuff said.
Did you mean CDS? As your sentence doens't make sense as its written.
The sirens might not be wailing yet but some of the lights are flashing.
Sullivan: We have to hold these assets on our books longer than we might ideally like to.
Tuld: Yes.
JARED: Sell it all. Today.
JOHN: Is that even possible, Sam?
SAM: Yes. But at what cost?
Not to do so would be reckless.
What is unusual about this is that syndications usually clear in a few weeks but this has been stuck in distribution for six months and still not closed.
And it's just one of them: they know that there is about three trillion dollars of this risky shit out there either in the hyperscaler/FAANG balance sheets or off the books, so they are going to struggle to sell this, and Oracle's position gets worse every day.
We are getting a lot closer to the Margin Call crisis meeting moment where one bank decides it cannot assess the risk of their share of it, are we not?
But this time it is happening in a world where retail investors, lawyers, government officials, elected representatives are much more likely to have seen Margin Call and The Big Short.
1. they are one place above junk;
2. they are being treated as effectively junk;
3. if they get classified as junk, any pension portfolios holding Oracle stock must sell as they can't hold junk stocks.
If that happens then there will be a cascade of effects. Not just in the AI/datacentre sector but elsewhere, as Larry Ellison holds a stake in Paramount/helped finance the Paramount/WB merger.
[1] https://www.youtube.com/watch?v=rHaPfdnRoZ4 -- Mostly Economics Podcast #55: Who pays when the AI bubble bursts with Ed Zitron (12:36 - 23:53)
Wouldn't hedge funds or PE recognize this arbitrage opportunity and eat this up, making this a non-issue? Vastly simplifying this: suppose "junk bond" means the chance of default is greater or equal to 10%. Of course it's bad if the risk of oracle going bankrupt edges up from 9% to 10.1%, but we shouldn't expect pandemonium from the 10% cliff edge, because there's plenty to be made by selling all the bonds that pension funds are obligated to sell, because now they have higher yields (yields move inverse to bond prices).
I guess, maybe, but perhaps the risk that GP is commenting on is that it could be it's own form of Juicing.
I do sincerely doubt, even with the big shifts that AI enables to have companies 'move away' from Oracle, that Oracle will default.
On the other hand, There's still an elevated chance of decline.
IOW, Those stocks might move from pension type plans to different baskets as a result, however there is a non-negligible risk of some (not total!) loss in the 5-10 year term.
Will add, I would -hope- that the same pension rules around selling questionable stocks, would also have rules limiting what level of exploitation is possible.
Of course past performance is no guarantee of future results, as they say.
Recently, the Australian AI firm failed to secure an IPO [1], [2].
OpenAI is looking to go for an IPO despite shaky revenue [3], [4], [5], resulting in a stock drop for companies like Oracle invested in OpenAI [6].
Anthropic is in a similar situation: looking to IPO despite shaky finances [7], [8], [9].
The companies involved are all investing in each other and trading money between them leading to circular investment [10].
[1] https://www.wsj.com/finance/firmus-ipo-ai-trade-f490c240
[2] https://www.abc.net.au/news/2026-10-10/how-firmus-the-bigges...
[3] https://techcrunch.com/2026/10/08/openais-revenue-is-reporte...
[4] https://www.inc.com/chris-morris/what-openais-leaked-financi...
[5] https://futurism.com/future-society/ai-financial-bubble-open...
[6] https://www.cnbc.com/2026/10/08/open-ai-revenue-nvidia-oracl...
[7] https://www.investors.com/news/technology/anthropic-ipo-fili...
[8] https://www.newconstructs.com/wp-content/uploads/2026/10/Ant...
[9] https://www.tradingkey.com/analysis/stocks/us-stocks/2622154...
[10] https://graceblakeley.substack.com/p/the-ai-circular-economy
Equity shares do not have ratings, bonds have ratings. Everything else is correct, if their bond ratings slip one more level they will be rated junk which may force some holders to liquidate if they’re not allowed to hold junk bonds.
Much of what Margin Call is about has already happened in the few days before the movie (the events that precipitated the firings but are not really discussed).
Sullivan is talking about the music slowing, saying "if the music were to stop…". He's suggesting that it is bad, in the context of those few days, and action is needed.
Tuld saying he fears he "can't hear anything" is about much more fundamental uncertainty — his own belief that no set of numbers could be useful and that the actual situation required acting on his own gut instinct is that it was much, much worse.
In short, the point is they did not have any confidence they could ride out the consequences of the preceding, largely un-discussed few days, because they were beyond their parameters, and that there was no time to assess the real risk.
The longer answer is that banks generally have policies about how concentrated they can be in individual borrowers, industry sectors, geographies, etc. That is why you rarely see a single bank take all of a very large corporate loan. They syndicate it out.
This is common. Selling debt is normal. If there is money to be made by selling the debt and going back to issuing more loans, they’ll try to find a buyer.
Everyone is so primed to believe that all things related to data centers are bad that they’ll assume this is a special event. The person writing the headline knows that.
Stop trying to be someone you ain’t.
They hold it to help stabilize the market for it while its being sold and then offload it once the market has been established.
Nothing out of hte ordinary here, but to be fair the original article isnt' a financial site so they would have no real idea about how this work behind the scenes.
This doesn't seem like a story at all given how robust this debt is trading.
Where have I heard this before.
Sure that could be true, but that has nothing to do with the public debt we're talking about.
Completely different markets.
That's like me saying the condo Market in Miami is hot and you rebutting that single family homes in HOA's in Phoenix are not hot right now.
Both can be true at the same time as they are completely different market that are only very tangentially relate.
- I still have a mortgage on a property but don't live there so the mortgage bill gets sent to a different address (b/c rental property)
- Company A originated the mortgage and sells it to Company B
- Company B completely ignores the mailing address and starts sending bills to the original property address
- only through blind luck did I realize that the mortgage bill was being unpaid
This seems like basic, basic bookkeeping but I guess it's also a reminder that "reality has more detail than people realize" and that there are plenty of opportunities for improving systems.
Still another: One particular suite address only maps to one person, entity, or family. Counterexample: Any of thousands of commercial mailbox-rental / package-shipping stores in US strip malls, which tend to have suite numbers (within the mall) and possibly hundreds of tenants, all at one suite.
https://moversguide.usps.com/mgo/move-combined-mobile will let you do it online.
It's a pain that you have to do this because of their incompetence. But it will save you the embarrassment.
One of the changes USPS made in recent years is that they are securing mail forwarding. That type of arrangement may not be possible.
https://www.usps.com/manage/forward.htm
(They'd stop fucking this sort of thing up pretty quick if they faced financial consequences for it.)
As a Canadian, other than the absolutely insane rate of school shootings, this is probably the next-wildest difference I can think of.
Canadian mortgages are held and stay with the bank that approves the mortgage. The only way the mortgage goes to a different lender is if the owner goes to that other bank to renew the mortgage. It just boggles my mind that mortgages can just be assigned willy-nilly without the owner’s consent.
That directly contradicts specific points made in the story, about specific Tier 1 banks and specific amounts of debt not finding buyers.
"Nothing out of hte[sic] ordinary here"
Matthew Moniot, Man Group: "the deal sizes involved are 'out of scale to anything we've thought about, ever.'"
Hard to compare your model with his, since he doesn't have one. When quotes from named credit risk guys from named finance organizations sound like angsty Gen Z, it's not ordinary.
In any case, the story is covering something significant. There is an important test occurring, and we're all going to learn the results, sooner or later. It could all turn out fine, but I don't know. I don't believe you really know either.
But that has nothing to do with the scare tactic headline that says banks are trying tot offload debt. As I said this is exactly how banks work and how they treat all debt they underwrite.
It’s not necessarily a statement about AI, it’s just finance and accounting.
If the market is already crammed with sector-concentrated debt (because of the amount already issued), finding buyers for more is going to get increasingly hard.
At some point everyone big already has enough exposure and isn't interested in the risk of more.
The only thing that keeps the party going is being able to drain that debt off bank sheets, but it's hard to see a ton of interest in fixed rate, capped upside exposure to a high risk industry in an environment of rising central bank rates.
And this is supposedly high grade debt we're talking about: JPMC is spending months on end shopping around a few $10's of billions, and having to discount it. Never mind the far greater sums on the books of vastly less credible outfits, complete with meme names that always age poorly when things go pear shaped.
How a bunch of highly speculative moonshot big tech stuff has managed to be financed by Tier 1 financial institutions and high grade debt, as opposed commonplace equities, is a question for the ages, and we may all yet find ourselves suffering the answers.
FTA: That is exactly why CoreWeave-leased paper prices at 7.00% to 9.875%, against the 5.70% to 7.875% that hyperscaler-leased bonds command, per bond market data compiled by CCIR.
Just so were clear: Core weave paid Micron and NVIDA 80 percent margins to buy equipment at a 9 percent lending rate so it can sell it to providers who are subsidizing tokens at 20x their own API prices (that are likely also subsidized).
Could it all work out? It could. But it does not look like there is the financing to keep building - thats going to impact prices and the moment that happens this debt looks even more risky.
Here is today: Shares of AI cloud provider CoreWeave (CRWV) dropped over 15% Wednesday, a day after venture capitalist Chamath Palihapitiya posted on X a chart that showed the yield Nvidia-backed firm's 2032 bonds hit 15% yesterday as a sell-off brought prices into distressed territory. SOURCE: https://finance.yahoo.com/markets/stocks/articles/nvidia-bac...
I can conclude that supply is far exceeding demand because of the rates attached.
Consider this an invitation to expound at great length on the subject, if you are in to that sort of thing.
In this case CoreWeave has sold ~4billion in 5 year bonds at a 9% rate. If you buy one for 1000 bucks you will get 2x45$ payments a year for 5 years. Then at the end a 1000 cash payment.
For the sake of argument let's say you bought this earlier this year. You now have an asset worth 1000 dollars. If it is a good bond (A US treasury or Municipal Bond) you can borrow against it from lots of place - you have something worth 1000 dollars on your "books" (accounting). For the earliest sales (you) core weave has already made a 45 dollar payment.
You now have a 1000 dollar asset and 45 bucks in cash. IF you hold to maturity and keep all the money you get 1000 bucks back and 450 bucks in interest.
What happens if you want to sell it? Maybe rates went down a lot, so that 9 percent is huge, and the AI boom is going gang busters so that 9percent looks like a good deal. I offer you 1100 for the bond. You now have 1145 dollars in cash.
What happens if you need money, and rates are the same and CoreWeave looks the same? I will offer you 980 for it... You now have 1025 bucks in cash... and I get the rest of the payments (45$x9) and 1000 bucks back at the end. (This has tax implications).
What happens if you want to sell cause you need money AND, you think CoreWeave is gonna burn, or this "ai thing" is a bust. Well it's likely that I think that too. the bond is only ever going to pay 45X9 more payments but I want a higher rate. I say to you "I'll give you 800 for it" (a 15% rate), cause CoreWeave is fucked. You need the cash so you sell it for that much and have 845 bucks. Thats a loss of $155.
Let's say you bought a billion of those bonds, and expected to walk away 1.45 billion in 5 years - but you sold them all off today and LOST 155 million.
Someone owned a LOT of 9% Core Weave bonds and either found a BETTER investment -OR- they wanted out of CoreWeave and no one had the money or desire to own them. The bond buying market is announcing that "We don't want any more CoreWeave Debt".
There are a lot of problems with CoreWeave: those 5 year bonds it sold, well the contracts to support them (15 billion a year in revenue on 60 billion of debt) only run for 3 years. At current rental rates for long term commitment NVIDA hardware had a 3 year pay off term... But thats not power, the building, staff, the business was (is) marginal.
Who was selling the debt? Well my guess would be a chunk of it was JP Morgan (who set up the bonds to start with). Dumping it right after you announce your numbers (they were a minor miss) and lower your target for CoreWeave's stock would be an ideal time. But that is PURE speculation.
Are you saying there are existing bonds at 9% being sold for some amount of money (possibly even less than the original principle?)
How does selling the $1k bond at $800 translate into 15%? The bond was originally worth $1k+$45*10=$1450, now $45*9 ($1405?), so if you buy it for $800 that's approximately 85% of its original value ($800+(9*$45)/$1405) and thus a 15% decrease? How does that relate to the original 9%?
Is this about what existing bonds are being sold for or new bonds?
Presumably a us treasury bond yield is basically the same as inflation?
The bonds were issued earlier this year. Their sale is a loss for someone, and NOT a small one.
> How does selling the $1k bond at $800 translate into 15%? 1000 5 year at 9 percent 5x90(450) + 1000 back.
If you need to dump it for 800, the bond still pays out 1000 in the end, The issuer still pays out 5x90(450). The 200 bucks I pick up at the end gets spread out over the 5 years so it looks like 5x130 (650+800).
That would be 16 percent, but all the numbers for the sake of sanity got rounded from their real ones. 9.xx rate, 8xx resale price, and 14.9 (closing on 15) rate.
It gets calculated that way because if you need to sell more new debt it only makes sense to know how to price it.
> Presumably a us treasury bond yield is basically the same as inflation?
Sometimes, often not. But Treasury Bonds have quite a few advantages. No state tax on the interest payments. They are almost as good as cash, so you can borrow against them at a pretty good rate (Better than a HELOC and way better than a credit card). Lots of institutions will give you a line of credit for up to 90 percent of the value of the bond. You can also borrow against them for margin trading (60percent? better than stocks) -
2. Bank sells the debt at $Z because holding the debt is risky and the bank would rather have cash than debt. Its counterparty would rather have more cash over time and is fine with the risk but does not want to bother with the whole process of closing the loan. $X < $Z < $Y.
3. Profit! Bank spent $X to get $Z, which is greater than $X.
stop trying to normalize that oracle got away with shady deals with those banks.
could you get 20mi on your 1mi house? yeah, oracle could, and much more. how on earth is this "normal" and "what banks do"?
Also just in: water is wet!
This might play out just fine, or crash like the housing market.
https://www.ft.com/content/678bcd31-6c2a-447b-a68c-ce6dec220...
Is it a data center with nothing but GPUs, that does nothing but inference all day and all night?
Is it a data center custom-built for only AI companies to use?
Do these “AI Data Centers” have data storage? VPS? …
Or, as I suspect, is “AI Data Center” simply a fake media buzzword that really means “Normal Data Centers” but with political baggage and stigma built-in.
It's a different build-out to a data centre that can host video/YouTube content (fast CPU, no GPU, large number/size of hard drives), or that can run web servers (fast CPUs, low to average hard drive capacity).
[1] https://www.nvidia.com/en-gb/data-center/
No you're absolutely right. The only difference that I can tell is power requirements are legitimately higher because GPUs consume more electricity. More cooling is also needed because there's more heat generated. Aside from those two things, it's a plain old datacenter that companies have been building at break neck speeds since like 1996. Now that I think about it, how come there's no panic about additional Internet backbone capacity needed to support these "AI datacenters"?
https://news.ycombinator.com/item?id=50050553
Is this our chance to 'get rich' by buying a negative valued asset? Like that time barrels were trading at 0$ or negatives?