I realize there’s a lot of hate in this thread, but I think there’s another side to this.
Force majeure is basically there for situations where something outside a company’s control prevents them from delivering on a contract.
I can see Oracle’s argument here. A number of states and local governments that were originally very supportive of new data centers, including places like Texas and Pennsylvania, have since started blocking or restricting new construction.
If Oracle signed these contracts and the government later changed the rules and have stopped those projects from moving forward, it’s hard to argue they should still be held to the same delivery commitments.
(That seems pretty close to the type of situation force majeure is meant for.)
Did they, or was it found that they had ignored them and are now being forced to follow the law? Those state and local governments were so supportive that the processes for approvals and reviews was all done in secret and without necessary inputs. You know, the things the laws in those areas said you had to do.
In this particular project, it looks like Oracle and Blue Owl Capital singed the contract before state regulators reviewed and provided necessary approvals.
Maybe some are not as familiar with this, but I'll use Texas as an example.
In 2025, Texas state government announces a $40B infrastructure investment specifically calling out building data centers for AI as a key part of that package.
In that case, would the people caught unemployed and with upside down loans in the 2008 financial crisis would have a reasonable argument for "force majeure" to cancel their mortgages? ... After all, they had a reasonable expectation that the economy would continue to support their employment, either with their current employer or by finding a new job that paid just as well?
at some point the phrases loses its meaning if it becomes applicable to pretty much anything.
No, but the economic situation is banning you from getting a job, which is even more powerful than the government, especially when the government does not support a rising economy which will create jobs, people can build wealth with.
A reasonably-written contract would have weasel words in it - something like "subject to zoning approval, and approval of the needed power and water and etcetera". Oracle has very expensive lawyers; they should have had those clauses in there.
Still, you write those for the known possible obstacles. Having the state government change the ground rules may not have been one of your known possibles. So, yeah, I could see force majeure being at least possibly reasonable.
"No one could have predicted that permits and infrastructure necessary for this project's completion would be delayed". Except for every project that's needed permits and infrastructure ever.
Force majeure can be defined within contracts. So if the contract said that force majeure included changes in law or failure to receive permits, then it's force majeure. There can also be a separate "change in law" provision.
One can't say whether it's valid or not without reading the actual contract and looking at the state's case law.
I really don't get who keeps using Oracle outside of Java, and how they make their money, must be some legacy contracts or something, but Oracle has always been customer hostile, from all the things I've heard, they sound like a ransomware company (idk if they still do it, but they'd sell you the database, then "audit you" and if you used anything that wasn't in compliance, fine you, until you converted that to Oracle DB too or whatever wild stuff). I just don't get it. I get even less how its execs are okay running what could be a much more successful company so poorly. You can change Oracle for the better and become even more profitable if you change how you do business.
Take Java as a solid example, they removed JavaFX from the core of Java, which made everyone think JavaFX was dead, recently they started providing enterprise support for JavaFX, which signals its production ready, but, who wants to trust Oracle? I sure don't. I can use C# for free without any of that baggage, and if I get all sorts of stable options for a GUI, and with Avalonia (and Uno?) I can run my GUI on any OS.
I think all their recent layoffs show that nobody uses oracle, and the data centers for open ai are their last best hope for future revenue.
Nobody picks oracle anymore, and everybody who's using oracle wishes they were off it and is working on projects to detach themselves from paying oracle heaps of money for software that's no better than open source alternatives.
"Enterprise" is still very much in the camp of "using Oracle". Which is likely how they survive despite being so customer unfriendly. "Enterprise" users are simply using what "the enterprise" purchased and have little choice in the matter.
It's only relatively recently (in the lifespan of large companies) that OSS alternatives have come along that can do much more of what Oracle can do. Oracle was miraculous for a long time. It was also a huge pain to operate (source: I know people who've operated it), particularly in its RAC configuration, but if you could operate it you could run a bank and rely pretty well on Oracle being rock solid.
I’ve been on some teams where some SQL devs can do some non intuitive things with SQL (particularly some of the enterprise features from Oracle). Not saying they were the right decisions but I do think some processes get knitted to these enterprise features and it’s painful to walk back later
There's likely an awful lot of tricky logic in the queries, stored procedures, etc. Dealing with taxes (very location and situation depending), details of any given customer's plan, equipment charges, etc etc for lines going back for decades... this stuff is hard to migrate correctly. It's not necessarily about qps.
Oracle Database is still the gold standard for data in an enterprise setting. Postgres is great, but it doesn't have the clustering or replication capabilities, or support, that Oracle does.
When you're talking about business, the software itself doesn't really matter at all. Accountability is everything. There has to be a throat you can choke, an ass you can drag onto the carpet, when it fails. That is the reason why Oracle persists as an organization: they assume accountability for the data storage layer, freeing their clients to worry about other things.
This also has implications for developers grousing about "being an accountability sink" for AI-generated code. Being an accountability sink is the business and as Larry has demonstrated, quite lucrative indeed. Developing the code is a side effect that can now be done close to 100% by machine.
You don't have experience from corporate environment. Oracle is very much present, alive and kicking there, in 1000 and 1 little or big applications, sometimes even out of convenience and cost (is you have just 1 DBA team supporting oracle, adding even 1 postgres to production system could be a cost nightmare to support and manage.
And SQL DBs are not going anywhere in next few decades.
Zipcar started in 2000/2001 on Oracle and relied heavily on stored procedures, particularly in billing (as SaaS didn't exist back when it started and was a bad fit for a long time). It took an entire rewrite of the system to get off of Oracle.
Will be interesting to see what excuses these companies will come up with to justify stopping investment in AI without spelling it out. The first one was already announced: it's going to kill everyone so we need to proceed more slowly.
Interesting. I had been interpreting "it's too dangerous; make us slow down" as hype to get more investment and/or customers. "We don't have the money to build out" is an interesting alternate interpretation.
It's easier to pay back the loans if they don't spend the money in the first place. This sounds like they can't spend the money, so paying it back should be easy, not hard.
> Why would Anthropic pay through the nose and make such desperate moves for compute?
That's pretty consistent with them having not been able to get the physical data centre capacity they have supposedly had built actually online, I think?
More to the point, Anthropic appears to mostly be capable of serving its customers, doesn't it? It's not so resource-starved that people aren't able to use it at all.
So what does it tell you that only 300 megawatts was enough to make a very meaningful difference to that?
If Anthropic's fortunes can be meaningfully changed by only 300MW which allowed them to uncap their customers, what is the case — even including OpenAI and Google -- for the demand for the 23GW of unbuilt AI data centre build commitments out there?
Satya Nadella himself said there will be some oversupply.
What if there is actually lots of oversupply?
My instinct is that the internet/software/IT industry sees its own demand for Anthropic and OpenAI and is misleading itself into thinking that this is representative of customer demand in the rest of the industry, let alone others, when what possible other human industry tasks are there that match the compute demands of writing software?
Another way to look at it: if the general demand for AI compute is so insatiable, how in this world of existing insatiable demand did xAI have spare capacity to lease? Why does Meta have spare capacity to lease?
The issue here is very evidently one of allocation. But allocation tends to get solved over time. That could leave quite a lot of very big data centres relatively empty.
I mean I do think there’s some likely extra demand, though I think it’s likely some demand will fall away from data centres to the desktop.
(And what you imply is also true, isn’t it - OpenRouter and equivalents will address this demand while also resolving some underlying underutilisation)
But the claims of awful shortages feel wrong when there’s little evidence of radical undersupply for the two big firms who are offering heavily subsidised plans.
It all seems pinned on the idea that some compute-guzzling application users clearly don’t have right now will emerge.
Due to the difficulty of building and the time lead, there will be a multiple of applications for every data center that actually comes to fruition.
But if you're talking about the capital investment that's being talked about as too high, I would agree that it's way too high, as all these big tech breakthroughs result in too much building at first. Which hurts the builder's profitability but helps consumers.
When our spot instances go away and don't come back, that's a sign that there's not enough capacity.
There is more demand for compute than there is supply, and that demand is going up.
Look at all of the things that have been and are being automated and realize that we're year 6 into a hundred year journey.
Look around you at all the things that aren't automated. They will be.
The world you grew up in won't be here for much longer. We're going through a transformation even more dramatic than the industrial revolution. There were no airplanes and then we landed on the moon.
Edit: downvotes as expected, but I'm shocked that a bunch of smart people at HN aren't taking the bet where this leads to a complete upending of the status quo. The holdout crowd needs to dream bigger.
The models are solving Millennium Prize problems, finding novel biological pathways our scientists can't see, putting every single software engineer out of the job of "coding by hand", and doing the same to designers, marketers, filmmakers, and 3D artists who used to have to operate pixel by pixel / photon by photon.
There's a huge gap between what's being described in that comment and an "extinction level event" is that's what ELE means.
What's being talked about is a massive extension of human capability, into the intellectual sphere rather than the physical environment. The atom bomb had a clear way to get to extinction, a sudden chain reaction that leads to collapse of the entire biome.
The only way AI gets there is something like control of the nuclear weapons on all sides and that control becoming suicidal. Every other major event leads to at most civilization collapse, after which I assume we will have some sort of Butlerian Jihad restrictions on getting into the same situation.
But, I would love to know what I'm missing here, if ELE has clearer routes!!
> The only way AI gets there is something like control of the nuclear weapons on all sides and that control becoming suicidal.
Not the way I was thinking off; I was thinking more along the lines of brain atrophy until humans are helpless (maybe 2 complete generations of humans).
And that's a short extinction; the KT extinction took place over thousands of years. If, 1000 years from now, we are finally at the place described in "The Machine Stops", only then will an observer be able to decide that we don't have the necessary critical mass of intelligent humans anymore.
Humans got to be the apex predator simply on intelligence alone. It's possible that without thinking ability we'd go extinct.
> Not the way I was thinking off; I was thinking more along the lines of brain atrophy until humans are helpless (maybe 2 complete generations of humans).
That's not how humans work. We developed math, art, literature, etc., because people love the act of doing it. We will continue to do that.
We're not going to stop thinking because we don't have to. We stop thinking because we subject ourselves to political propaganda, or religious belief systems, or other group control methods that turns off critical thought. But even then, it's only a subset of the population, it won't get us all.
That's the usual bait and switch though. Very few people are saying AI the statistical software is going away. What people are dubious of is the commercialisation model of building a butt ton of data centres to serve massive general models by two heavily subsidised VC companies even as industry gets more mature and fickle as they get their heads around using it to solve problems and it goes onto Opex budgets. I think this is more like the early days of cloud which finished in "software built around clouds" rather than "everyone spins up their own fleet of servers for everything". Obviously the current investment levels need everyone to run their own fleet of Ralph looping agents on frontier models so that's what's marketed as the smart thing to do.
On HN there are a lot of people who truly believe that this whole thing is hype and the bubble will burst and things will go back to how it was before.
The real answer to “how many data centers do we need” is hard. Anyone giving confident answers to this should be suspicious.
The easy type of answer you get from those kind of people will always be “we are building too much”. They’d have said that 3 years back and have no shame in admitting it.
The thing is, you can overestimate or underestimate. Both have cons. The people sincerely believing this is all hype will always underestimate because all their epistemic muscle haven’t been used.
Finding out that Oracle has a bond rating of BBB- [1] (one step above junk bonds) feels very appropriate for such a company with such a sordid history.
I can't reconcile that with https://news.ycombinator.com/item?id=49831204 . Everyone knows about Oracle's incompetence already and should reasonably be able to predict it, especially themselves.
Right, Oracle is at the bottom of the heap (they borrowed the most with the least initial credibility— who the fuck seriously thinks Oracle was ever going to be a major player in AI), so weakness in the market will appear with them first. But I think it will bubble its way up.
> Oracle was ever going to be a major player in AI
Stupidity on the behalf of management to get into AI at any cost even if it means becoming a land-lord of AI data-centers. They have good company with allbirds and other crypto-miner-turned datacenter operators. This is what a mania looks like. Hustlers of all shapes and sizes line up for dumb money which wont ask any questions. At the backend of this, these operators will be cleaned out, and sold to the highest bidders and that explains why microsoft is renting capacity, because they know at some point in the future there will be foreclosures, and they can buy on the cheap.
Oracle's specialty is extracting money from the government for everything IT-related. Why wouldn't they be a major player? Investors might not have foreseen DOGE and Grok.
Everyone is interpreting this as a sign the bubble is bursting, when the actual story is costs for materials and labor are going up and delays are happening because so many data centers are getting built.
If Oracle was able to build the data center under-cost and faster than expected, that would be a better sign the bubble was bursting!
What does it change exactly? It's not like gas or materials are going to get cheaper any time soon. It can burst for a whole lot of reasons, including the 2nd coming of Jesus or resources price... it's not like people have been calling AI ressource intensive for years right ?
It changes because the usual story behind bubble is that all demand is fake.
Stock market valuations measure future revenues.
So if the reason is that the demand is so high that we are facing new bottlenecks, it doesn’t imply that the underlying capabilities are fake or that the demand is fake. Just that it’s another hurdle to cross.
Right, so the presumably the revenue projections aren't increasing at the same pace and the financials don't make sense anymore. Which is potentially a sign of the bubble bursting.
>How profitable a datacenter is, is kind of irrelevant to the convo, if you are unable to power it due to material shortages.
It isn't irrelevant, because they have to make money to be built. The debt for these projects was already issued, so you have to pay the coupons. If Oracle doesn't pay the developer is gonna suffer on their end.
Wut? Bubble bursting doesn’t mean bottlenecks can’t exist lol. Bubble can burst if projected revenues can’t be met. Unless you think revenues projected 15 years in future are hampered due to this news, it has nothing to do with bubbles.
I realize there’s a lot of hate in this thread, but I think there’s another side to this.
Force majeure is basically there for situations where something outside a company’s control prevents them from delivering on a contract.
I can see Oracle’s argument here. A number of states and local governments that were originally very supportive of new data centers, including places like Texas and Pennsylvania, have since started blocking or restricting new construction.
If Oracle signed these contracts and the government later changed the rules and have stopped those projects from moving forward, it’s hard to argue they should still be held to the same delivery commitments.
(That seems pretty close to the type of situation force majeure is meant for.)
> the government later changed the rules
Did they, or was it found that they had ignored them and are now being forced to follow the law? Those state and local governments were so supportive that the processes for approvals and reviews was all done in secret and without necessary inputs. You know, the things the laws in those areas said you had to do.
In this particular project, it looks like Oracle and Blue Owl Capital singed the contract before state regulators reviewed and provided necessary approvals.
https://www.datacenterdynamics.com/en/news/new-mexico-regula...
And not seeking necessary regulatory approval doesn't really seem like things 'beyond their control.'
Maybe some are not as familiar with this, but I'll use Texas as an example.
In 2025, Texas state government announces a $40B infrastructure investment specifically calling out building data centers for AI as a key part of that package.
https://gov.texas.gov/news/post/governor-abbott-google-annou...
Then, just this week, Texas government reverse their position and halts permits for all data centers:
https://gov.texas.gov/news/post/governor-abbott-directs-tceq...
I'm not trying to defend Oracle, but I can understand their perspective on this.
If the state government is only now blocking what their contract commits them to deliver, it's unreasonable.
Seems like Oracle didn't have anything in writing. Looks more like a risk of doing business and over promising.
In that case, would the people caught unemployed and with upside down loans in the 2008 financial crisis would have a reasonable argument for "force majeure" to cancel their mortgages? ... After all, they had a reasonable expectation that the economy would continue to support their employment, either with their current employer or by finding a new job that paid just as well?
at some point the phrases loses its meaning if it becomes applicable to pretty much anything.
Cancel their mortgages and keep their houses?
I don’t think Oracle is proposing they get a data center they didn’t finish paying for.
In your example, governments we not enacting new regulation that banned you from getting a new job.
In this situation, governments are now enacting new regulation that ban companies like Oracle from building new data centers.
No, but the economic situation is banning you from getting a job, which is even more powerful than the government, especially when the government does not support a rising economy which will create jobs, people can build wealth with.
so the argument is even stronger than for oracle.
A hurricane wipes out your data center and someone sues you for it not being finished on time. Thats what force majeure is for.
You made a series of bad business decisions != force majeure.
As the AI bubble starts to pop we’re going to see a lot of creative claims as everyone rushes for the exits.
A reasonably-written contract would have weasel words in it - something like "subject to zoning approval, and approval of the needed power and water and etcetera". Oracle has very expensive lawyers; they should have had those clauses in there.
Still, you write those for the known possible obstacles. Having the state government change the ground rules may not have been one of your known possibles. So, yeah, I could see force majeure being at least possibly reasonable.
If developer's lawyers were better than Oracle's lawyers and stopped those weasel-words in the contract - that would be poetic justice for Oracle.
Doesn't 'Force Majeure' mean exceptional events outside ones control? Pretty sure everything involved here is Oracles own making lol
"No one could have predicted that permits and infrastructure necessary for this project's completion would be delayed". Except for every project that's needed permits and infrastructure ever.
I hope whoever buys Oracle fires Larry by email.
> I hope whoever buys Oracle fires Larry by email.
I certainly hope not; you can't fire[1] someone via email.
============================
[1] Out of a canon. Into the sun.
> Out of a canon. Into the sun.
Takes to much delta-V. Like... a lot of delta-V. Shooting out of the solar system takes less energy. #begreen
I chucked rather harshly, how dare you
Rumor is this feature will be in Opus 5.6
Honestly, I do have a list of people I can volunteer to test that new and exciting deliverable.
Force majeure can be defined within contracts. So if the contract said that force majeure included changes in law or failure to receive permits, then it's force majeure. There can also be a separate "change in law" provision.
One can't say whether it's valid or not without reading the actual contract and looking at the state's case law.
"Well a wave hit it"
> "Well a wave hit it"
I'm assuming this is a reference to the classic Clarke and Dawe skit:
* https://www.youtube.com/watch?v=3m5qxZm_JqM
* https://en.wikipedia.org/wiki/Clarke_and_Dawe
A wave hit the Oracle?
I really don't get who keeps using Oracle outside of Java, and how they make their money, must be some legacy contracts or something, but Oracle has always been customer hostile, from all the things I've heard, they sound like a ransomware company (idk if they still do it, but they'd sell you the database, then "audit you" and if you used anything that wasn't in compliance, fine you, until you converted that to Oracle DB too or whatever wild stuff). I just don't get it. I get even less how its execs are okay running what could be a much more successful company so poorly. You can change Oracle for the better and become even more profitable if you change how you do business.
Take Java as a solid example, they removed JavaFX from the core of Java, which made everyone think JavaFX was dead, recently they started providing enterprise support for JavaFX, which signals its production ready, but, who wants to trust Oracle? I sure don't. I can use C# for free without any of that baggage, and if I get all sorts of stable options for a GUI, and with Avalonia (and Uno?) I can run my GUI on any OS.
I think all their recent layoffs show that nobody uses oracle, and the data centers for open ai are their last best hope for future revenue.
Nobody picks oracle anymore, and everybody who's using oracle wishes they were off it and is working on projects to detach themselves from paying oracle heaps of money for software that's no better than open source alternatives.
> I really don't get who keeps using Oracle
"Enterprise" is still very much in the camp of "using Oracle". Which is likely how they survive despite being so customer unfriendly. "Enterprise" users are simply using what "the enterprise" purchased and have little choice in the matter.
Is there a list somewhere of major "enterprises" in this position? I've always kinda wondered how Java keeps its edge over C# in the TIOBE rankings.
TIOBE rankings are meaningless.
What do you suggest instead, then?
Not worrying about language rankings.
Oracle is more than Java and an RDBMS. You cannot throw a rock without hitting a Fortune 500 company running PeopleSoft and/or the Oracle ERP.
FBI was breached via PeopleSoft. Do you think more companies would leave it if they were breached too?
https://www.404media.co/we-hacked-the-fbi-hackers-say-they-h...
https://techcrunch.com/2026/09/22/hacking-group-shinyhunters...
It's only relatively recently (in the lifespan of large companies) that OSS alternatives have come along that can do much more of what Oracle can do. Oracle was miraculous for a long time. It was also a huge pain to operate (source: I know people who've operated it), particularly in its RAC configuration, but if you could operate it you could run a bank and rely pretty well on Oracle being rock solid.
They get enterprise and government contracts because of kick-backs and revolving doors.
I’ve been on some teams where some SQL devs can do some non intuitive things with SQL (particularly some of the enterprise features from Oracle). Not saying they were the right decisions but I do think some processes get knitted to these enterprise features and it’s painful to walk back later
My company uses Oracle BRM to bill for millions of phone calls each month. It would be such a crazy upheaval project to move from that.
Millions per month is tens per minute. That doesn't seem like all that much to me?
There's likely an awful lot of tricky logic in the queries, stored procedures, etc. Dealing with taxes (very location and situation depending), details of any given customer's plan, equipment charges, etc etc for lines going back for decades... this stuff is hard to migrate correctly. It's not necessarily about qps.
> they removed JavaFX from the core of Java
You can still use OpenJFX with OpenJDK, can't you?
Use, yes. Or with pretty much any JDK. It just isn't included as part of the JDK like AWT & Swing anymore.
Oracle Database is still the gold standard for data in an enterprise setting. Postgres is great, but it doesn't have the clustering or replication capabilities, or support, that Oracle does.
When you're talking about business, the software itself doesn't really matter at all. Accountability is everything. There has to be a throat you can choke, an ass you can drag onto the carpet, when it fails. That is the reason why Oracle persists as an organization: they assume accountability for the data storage layer, freeing their clients to worry about other things.
This also has implications for developers grousing about "being an accountability sink" for AI-generated code. Being an accountability sink is the business and as Larry has demonstrated, quite lucrative indeed. Developing the code is a side effect that can now be done close to 100% by machine.
You don't have experience from corporate environment. Oracle is very much present, alive and kicking there, in 1000 and 1 little or big applications, sometimes even out of convenience and cost (is you have just 1 DBA team supporting oracle, adding even 1 postgres to production system could be a cost nightmare to support and manage.
And SQL DBs are not going anywhere in next few decades.
Zipcar started in 2000/2001 on Oracle and relied heavily on stored procedures, particularly in billing (as SaaS didn't exist back when it started and was a bad fit for a long time). It took an entire rewrite of the system to get off of Oracle.
My friend you don’t understand the world of business..
Btw I hate oracle but they’ve got their business model down to a T.
I'm surprised to hear that solid-oxide fuel cells are apparently a non-gimmick thing now.
> "Oracle signed a deal with solid oxide fuel cell (SOFC) developer Bloom last year for 1.8GW of power, expanding it to 2.8GW of capacity in April."
https://www.datacenterdynamics.com/en/news/new-mexico-regula...
Apparently this one would have been the largest SOFC installation, by far.
Bloom has always been gimmicky.
https://hntrbrk.com/investigations/bloom
Will be interesting to see what excuses these companies will come up with to justify stopping investment in AI without spelling it out. The first one was already announced: it's going to kill everyone so we need to proceed more slowly.
Interesting. I had been interpreting "it's too dangerous; make us slow down" as hype to get more investment and/or customers. "We don't have the money to build out" is an interesting alternate interpretation.
It's hype to get Trump to regulate the import of competitive Chinese models. A regulatory capture play.
Surely this will assuage the creditors that Oracle is going to pay back its $100B+ loans.
It's easier to pay back the loans if they don't spend the money in the first place. This sounds like they can't spend the money, so paying it back should be easy, not hard.
Stock price is contingent on the idea of returns from that debt being spent. If they pay it back, stock price goes down.
The number of datacenters needing to come online doesn't add up to me. It's suspicious
How so? It's clear the hyperscalers can't keep up with demand.
Is it clear they need anything like the unbuilt capacity that is in the pipeline?
They are using tiny fractions of the data centres they claim to have "built".
There's a significant risk of a Chinese real-estate bubble here.
What source do you have for "using tiny fractions of the data centres"?
The money doesn't seem to back this up. Why would Anthropic pay through the nose and make such desperate moves for compute?
> Why would Anthropic pay through the nose and make such desperate moves for compute?
That's pretty consistent with them having not been able to get the physical data centre capacity they have supposedly had built actually online, I think?
More to the point, Anthropic appears to mostly be capable of serving its customers, doesn't it? It's not so resource-starved that people aren't able to use it at all.
So what does it tell you that only 300 megawatts was enough to make a very meaningful difference to that?
If Anthropic's fortunes can be meaningfully changed by only 300MW which allowed them to uncap their customers, what is the case — even including OpenAI and Google -- for the demand for the 23GW of unbuilt AI data centre build commitments out there?
Satya Nadella himself said there will be some oversupply.
What if there is actually lots of oversupply?
My instinct is that the internet/software/IT industry sees its own demand for Anthropic and OpenAI and is misleading itself into thinking that this is representative of customer demand in the rest of the industry, let alone others, when what possible other human industry tasks are there that match the compute demands of writing software?
Another way to look at it: if the general demand for AI compute is so insatiable, how in this world of existing insatiable demand did xAI have spare capacity to lease? Why does Meta have spare capacity to lease?
The issue here is very evidently one of allocation. But allocation tends to get solved over time. That could leave quite a lot of very big data centres relatively empty.
Open alternatives / routers stand ready to serve routed requests.
That’s why Altman is burning the excess on those research projects.
Your take is spot on. Inb4 boosters like echelon come in.
Oversupply is always going to happen- humanity over invests. When has it ever underinvested? Lmao.
I mean I do think there’s some likely extra demand, though I think it’s likely some demand will fall away from data centres to the desktop.
(And what you imply is also true, isn’t it - OpenRouter and equivalents will address this demand while also resolving some underlying underutilisation)
But the claims of awful shortages feel wrong when there’s little evidence of radical undersupply for the two big firms who are offering heavily subsidised plans.
It all seems pinned on the idea that some compute-guzzling application users clearly don’t have right now will emerge.
Due to the difficulty of building and the time lead, there will be a multiple of applications for every data center that actually comes to fruition.
But if you're talking about the capital investment that's being talked about as too high, I would agree that it's way too high, as all these big tech breakthroughs result in too much building at first. Which hurts the builder's profitability but helps consumers.
When our spot instances go away and don't come back, that's a sign that there's not enough capacity.
There is more demand for compute than there is supply, and that demand is going up.
Look at all of the things that have been and are being automated and realize that we're year 6 into a hundred year journey.
Look around you at all the things that aren't automated. They will be.
The world you grew up in won't be here for much longer. We're going through a transformation even more dramatic than the industrial revolution. There were no airplanes and then we landed on the moon.
Edit: downvotes as expected, but I'm shocked that a bunch of smart people at HN aren't taking the bet where this leads to a complete upending of the status quo. The holdout crowd needs to dream bigger.
The models are solving Millennium Prize problems, finding novel biological pathways our scientists can't see, putting every single software engineer out of the job of "coding by hand", and doing the same to designers, marketers, filmmakers, and 3D artists who used to have to operate pixel by pixel / photon by photon.
Everything is going to be AI soon.
> I'm shocked that a bunch of smart people at HN aren't taking the bet where this leads to a complete upending of the status quo.
If your take on the abilities of AI is correct, we aren't talking about upending the status quo, we're talking about an ELE for the human race.
There's a huge gap between what's being described in that comment and an "extinction level event" is that's what ELE means.
What's being talked about is a massive extension of human capability, into the intellectual sphere rather than the physical environment. The atom bomb had a clear way to get to extinction, a sudden chain reaction that leads to collapse of the entire biome.
The only way AI gets there is something like control of the nuclear weapons on all sides and that control becoming suicidal. Every other major event leads to at most civilization collapse, after which I assume we will have some sort of Butlerian Jihad restrictions on getting into the same situation.
But, I would love to know what I'm missing here, if ELE has clearer routes!!
> The only way AI gets there is something like control of the nuclear weapons on all sides and that control becoming suicidal.
Not the way I was thinking off; I was thinking more along the lines of brain atrophy until humans are helpless (maybe 2 complete generations of humans).
And that's a short extinction; the KT extinction took place over thousands of years. If, 1000 years from now, we are finally at the place described in "The Machine Stops", only then will an observer be able to decide that we don't have the necessary critical mass of intelligent humans anymore.
Humans got to be the apex predator simply on intelligence alone. It's possible that without thinking ability we'd go extinct.
> Not the way I was thinking off; I was thinking more along the lines of brain atrophy until humans are helpless (maybe 2 complete generations of humans).
That's not how humans work. We developed math, art, literature, etc., because people love the act of doing it. We will continue to do that.
We're not going to stop thinking because we don't have to. We stop thinking because we subject ourselves to political propaganda, or religious belief systems, or other group control methods that turns off critical thought. But even then, it's only a subset of the population, it won't get us all.
AI sucks or it's an "extinction level event".
Crazy times.
That's the usual bait and switch though. Very few people are saying AI the statistical software is going away. What people are dubious of is the commercialisation model of building a butt ton of data centres to serve massive general models by two heavily subsidised VC companies even as industry gets more mature and fickle as they get their heads around using it to solve problems and it goes onto Opex budgets. I think this is more like the early days of cloud which finished in "software built around clouds" rather than "everyone spins up their own fleet of servers for everything". Obviously the current investment levels need everyone to run their own fleet of Ralph looping agents on frontier models so that's what's marketed as the smart thing to do.
I think firms will eventually move toward sovereign protected compute. It’s inevitable.
Govt’s are waiting for this stuff to mature and shake out. USA can’t be trusted anymore.
On HN there are a lot of people who truly believe that this whole thing is hype and the bubble will burst and things will go back to how it was before.
The real answer to “how many data centers do we need” is hard. Anyone giving confident answers to this should be suspicious.
The easy type of answer you get from those kind of people will always be “we are building too much”. They’d have said that 3 years back and have no shame in admitting it.
The thing is, you can overestimate or underestimate. Both have cons. The people sincerely believing this is all hype will always underestimate because all their epistemic muscle haven’t been used.
[flagged]
A better overview of the situation with more context exists on Yahoo Finance [0], also there is no paywall.
[0] https://finance.yahoo.com/technology/ai/articles/oracle-cite...
Finding out that Oracle has a bond rating of BBB- [1] (one step above junk bonds) feels very appropriate for such a company with such a sordid history.
[1] https://www.spglobal.com/ratings/en/regulatory/article/-/vie...
'Force Majeure' is a good way to describe their incompetence.
I can't reconcile that with https://news.ycombinator.com/item?id=49831204 . Everyone knows about Oracle's incompetence already and should reasonably be able to predict it, especially themselves.
Right, Oracle is at the bottom of the heap (they borrowed the most with the least initial credibility— who the fuck seriously thinks Oracle was ever going to be a major player in AI), so weakness in the market will appear with them first. But I think it will bubble its way up.
> Oracle was ever going to be a major player in AI
Stupidity on the behalf of management to get into AI at any cost even if it means becoming a land-lord of AI data-centers. They have good company with allbirds and other crypto-miner-turned datacenter operators. This is what a mania looks like. Hustlers of all shapes and sizes line up for dumb money which wont ask any questions. At the backend of this, these operators will be cleaned out, and sold to the highest bidders and that explains why microsoft is renting capacity, because they know at some point in the future there will be foreclosures, and they can buy on the cheap.
Oracle's specialty is extracting money from the government for everything IT-related. Why wouldn't they be a major player? Investors might not have foreseen DOGE and Grok.
Still too high of a rating.
It's based solely on thejr expected ability to replay, not their ethics.
Reuters has a non-paywalled (short) report with a bit of detail:
https://www.reuters.com/business/oracle-cites-force-majeure-...
Reuters is paywalled now - sometimes. They use what they call a “dynamic paywall”, discussed here: https://digiday.com/media/how-reuters-dynamic-paywall-is-boo...
[dead]
Oracle's $165B New Mexico Data Center Plan Hits a Gas Pipeline Delay - https://news.ycombinator.com/item?id=49321142 - August 2026
[dead]
[dead]
[flagged]
Everyone is interpreting this as a sign the bubble is bursting, when the actual story is costs for materials and labor are going up and delays are happening because so many data centers are getting built.
If Oracle was able to build the data center under-cost and faster than expected, that would be a better sign the bubble was bursting!
What does it change exactly? It's not like gas or materials are going to get cheaper any time soon. It can burst for a whole lot of reasons, including the 2nd coming of Jesus or resources price... it's not like people have been calling AI ressource intensive for years right ?
It changes because the usual story behind bubble is that all demand is fake.
Stock market valuations measure future revenues. So if the reason is that the demand is so high that we are facing new bottlenecks, it doesn’t imply that the underlying capabilities are fake or that the demand is fake. Just that it’s another hurdle to cross.
Right, so the presumably the revenue projections aren't increasing at the same pace and the financials don't make sense anymore. Which is potentially a sign of the bubble bursting.
I mean the literal guidance is that Oracle may not make money on the datacenter by 2028 because they won't be able to turn it on.
How profitable a datacenter is, is kind of irrelevant to the convo, if you are unable to power it due to material shortages.
>How profitable a datacenter is, is kind of irrelevant to the convo, if you are unable to power it due to material shortages.
It isn't irrelevant, because they have to make money to be built. The debt for these projects was already issued, so you have to pay the coupons. If Oracle doesn't pay the developer is gonna suffer on their end.
Wut? Bubble bursting doesn’t mean bottlenecks can’t exist lol. Bubble can burst if projected revenues can’t be met. Unless you think revenues projected 15 years in future are hampered due to this news, it has nothing to do with bubbles.
[flagged]