That's Kami Kaze for US frontier labs. There's no way from here for them towards trillion dollar valuations.
That's strategic sabotage to say the least.
Ban all open weight models! Cried the VCs.
That's Kami Kaze for US frontier labs. There's no way from here for them towards trillion dollar valuations.
That's strategic sabotage to say the least.
Ban all open weight models! Cried the VCs.
> That's Kami Kaze for US frontier labs. There's no way from here for them towards trillion dollar valuations.
Those valuations are a scam, they have always been a scam, like an updated version of the collateralised debt obligations (CDOs) from 2008 [0]
[0] https://www.theguardian.com/business/2010/apr/18/goldman-sac...
I rewatched The Big Short recently, and all I could think about was AI labs.
Too bad they aren't public so we can't short them
shorting is not really viable anyway because you need the timing as much as the direction
it worked for the big short guys because the knew exactly when it was going to go bust
Michael Burry burned through tens of millions of dollars on holding his position. Others enter later but still ahead of the collapse. They knew and they didn’t. And even if Michael Burry new exactly when, he had to convince his investors that the millions they were losing leading up to it would get recouped.
You can buy CDS on Oracle, as many did already.
Through my broker online for a couple thou or do I need big money for a sit-down agreement? Apologies, this is an instrument I didn't think existed to the retail market.
I saw the same kind of comments when the first Deepseek model was initially released. However the relevance that investors assign to something like this does not seem to match the excitement expressed here.
I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
> I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
They'd be wrong. Especially about Europe. Many companies are in a bind about this: they can't upload their confidential files to a service provider but they do want AI. So right now they end up with a choice: do nothing, breach their contracts, use a service provider that pinky promises not to share their data (which we have already seen they will violate) and run local models.
The latter is the reason you can't buy powerful GPUs for reasonable prices because a local model (GLM 5.3, DS 4, etc) + some hardware is all you need to solve the problem in a way that avoids the legal issues.
You can’t buy gpus at reasonable prices because US frontier labs are spending nearly a trillion dollars this year on data center buildout, not because some mid tier European insurance co needs to serve deepseek
Yes, well, you're wrong about that. They're buying GPUs by the pallet here.
Mid tier insurance co is working with local inference providers and some of these now have thousands of GPUs up and running. Don't drink the HN cool aid that the EU is a backwater where nothing happens too much or you'll end up surprised one day.
The one thing where Europe really does lag is in training, that would be fair. But inference providers are popping out of the ground like mushrooms and there are some billion euro+ projects underway. We -unfortunately- also have some scammers that are bilking (or attempting to bilk) investors pretending they have more than they do, this is annoying but I would expect that other countries have similar players as well.
Wow, whole pallets!?
There are roughly 70 GW of datacenters under construction in the US right now, and a GW is roughly 500k GPUs, so 35 million GPUs.
The 'by the pallet' buys are for small local projects, not data center size projects (of which we have more than a couple as well), as you probably already knew. This isn't some silly dick measuring contest.
The point was: they're buying them from the retail channels in relatively high volume because there is a massive shortage of GPUs in general, but that does not seem to stop parties with smaller projects who do not qualify for buying directly from NV (assuming NV even has stock to sell).
5090's now sell for the same amount (or more) that you could buy an RTX 6000 for a few months ago, and RTX 6000's are ~19K and up.
I found a stash somewhere, bought two, then two more and when I was ready to buy the remainder they were all gone. Serves me right for being too careful.
No, the point is that US demand is what's driving up GPU prices, as GP asserted. Europe's purchases are a rounding error. And even then, most of the data centers under construction in the EU are from American companies.
The US has never cancelled as many data center projects as now.
this is a fun comparison, so I ran the numbers - in the US the cancelations alone are as big as Europe's entire planned buildout (~11 GW), or almost 3x of what is currently under construction (~3.8 GW)
Not nearly as many as you’d think from news headlines: https://newsletter.semianalysis.com/p/datacenters-2027-the-y...
Your link doesn't speak of cancellations only about a certain GW worth of data centres being built but that are not operational.
There’s a link in the first paragraph that addresses what you’re looking for
A rounding error is taken to be sub 1%. Anyway, I shouldn't engage, good luck.
You can’t buy gpus at reasonable prices because US frontier labs are spending nearly a trillion dollars this year on data center build
Its not just the US ... Where do you think all those 5090's end up? Its just crypto mining all over again. Most people do not realize that China was the biggest crypto mining operation because of its cheap electricity and people willingness to take big risks.
The same happens now also, with people buying up 5090's (and lower class cards) in bulk, to strap into AI rigs. With Nvidia discontinuing the 32GB 5090s, we already see pressure on the lower end cards that had a better price / performance for a while now.
China will do China things *lol* ... People there are just more willing to take bigger gambles (and the cheap elec helps a ton). But that is also a factor why a lot of consumer level hardware gets under pressure. Its not just the supply chain to the enterprise market that is eating a lot for AI.
That does not matter so much anymore, when Huawei starts selling those NPUs for a fraction of the NVIDIA price. It's all downhill from there on for OpenAI, Anthropic and NVIDIA.
It's all downhill for OpenAI and Anthropic investors, because they will never recoup what was invested. NVDA is safe(r) - Huawei first needs to catch it, and as AMD learned, it's a fast moving target.
Indeed NVIDIA has an actual moat. The other two, not so much.
> US frontier labs are spending nearly a trillion dollars this year
They are not, not even close. I wonder if you're aware how money is being "spent" between Nvidia and US AI corporations!
The other choice is AWS Bedrock, if it happens to have the model you're interested in. They do have GLM-5.3. They don't have DS-4 or 4.1 yet, and updates are intermittent at best, but they're already on the trusted provider list for anywhere large enough to be interesting.
That may not be acceptable for many companies outside of the USA.
Depends on the parameters of "acceptable", but they do offer inference in non-US zones. If the company is already on AWS for meaningful workloads with sensitive data (which is not uncommon), they've already been through the hard part.
That being said it's not quite as simple as would be ideal, you do need to check whether there's an additional EULA on your specific model choice.
Acceptable as in 'does not break the law'.
For data residency reasons? You can meet data residency requirements on AWS, I'm pretty sure. Mind sharing which condition blocked it in your experience? Mostly curious because in the US there isn't that much interest in on-prem deployments because it's easier to scale up and upgrade using the cloud and it's easier to use AWS because you don't have to run a new compliance process for each part of the stack.
You could until the Trump admin decided to (attempt to) overrule the deal that was in place (the DPF).
The perceived lack of independence of the FTC is a major driver behind the exodus on this side of the Atlantic and it isn't just the smaller players that are moving. Airbus for instance.
Now the datacenters supposedly operating under at-arms-length management are no longer sufficient for compliance.
There are calls to repeal the whole thing with respect to the US.
https://noyb.eu/en/us-supreme-court-just-blew-eu-us-data-tra...
Yes, biased source and title, but the contents are pretty good.
I see. So there's a risk that the EC or CJEU make a determination against the existing system because of the non-independence ruling of the FTC. So the movement away to regional providers is because no one wants to be caught surprised if the EC/CJEU make such a determination. Interesting. It must be early stage because from what I can see the most spend is still on the big two, but this does bode well for Mistral and EU-based open model hosts. Lots of room in the 20%.
It's early stage mostly due to inertia but there is a groundswell of movement. Airbus really turned some heads.
Yeah, about that... My kid works for a very small company. Chatted the other day about in-house inference vs ZDR providers and all. His team is trying to get access to GLM on AWS Bedrock. It is a second week they are trying to schedule a simple meeting to discuss what they need. Random silence, and back and forth from a new person each time. 2 WEEKS to just schedule a meeting... And as far as I know, no luck so far.
Update: they got back to us! Scheduled meeting early next week :facepalm:
You do realize there are hundreds of datacenters in Europe and the US that rent GPUs to European companies, right? You don't have to buy a GPU or even figure out how to use one. Pay for SageMaker.
Europe isn't in a bind. They can, and have been, running open weights on inference providers around the world. And they pay a premium for US models and accept its risk for the same reason they remain US allies. The cost is worth it.
Yes, I do. Not sure what in my comment prompted you to write all this.
> I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
Quite failure by someone to assume that EU businesses would use US over Chinese, based on the current political atmosphere. US causes even bigger concerns than Chinese labs.
i am amazed they expect everyone else to see the chinese as these fiction-like evil corrupt dictators
It is flabbergasting, indeed. I'm no friend of China, because it is an evil dictatorship, but as a business man I prefer dealing with China over the USA nowadays. At least China is predictable, even in its corruption.
It's just a single dictator. China has one dictator with full unilateral uncontested power. No courts shooting down orders or elections changing tides. One guy, all the power, for life. Every official, citizen, and business owner take his orders, uncontested.
If you think that is the choice basket to put your eggs in, best of luck.
Exactly. Imagine China's dictator would kidnap some other foreign president just to steal their country's natural resources. Or if he just starts a war against some Middle East country to take the focus away from some pedophile scandal.
Thanks God nothing like that could happen in the US, thanks to its constitution, courts and armed informed citizens. That's why I'm putting all my eggs in that basket.
The irony here is the blessing that in the US, all the dirty laundry gets put on blast for everyone to see.
And the airing of said laundry achieves what exactly? I don't see any knock on effect from that dirty laundry being aired, other than entertaining the masses with new scandal and signaling to the rest of the world that the US is not to be trusted.
> the airing of said laundry achieves what exactly?
No midterm routs in China.
For Europe, the answer is using homegrown metal, even if it is severely underpowered, and accepting that it will be running American and Chinese open-source models for the near future. (The French lobbyists looked likely to turn Mistral into another Arianespace, but they’ll be resisting implosion for at least another year.)
It's the feedback loop of democracy.
If you don't think it's important or needed, well, there is still a lot for you to learn.
You are delusional and indoctrinated.
US per capita spending on healthcare: $13,473 per person
China per capita spending on healthcare: ~$763 per person
Both have average life expectancy of 79 years. LITERALLY China has better healthcare, better housing, literally better in all the ways that count.
What does our "democracy" get us exactly???
Oh really? What about the Epstein files? What about the State protecting billionaire pedophiles? You have got to be naive beyond wit or willfully malicious to spew this nonsense. Did you vote for Trump? I think you did.
I think Xi is pretty benevolent as dictators go. Bad things you could point to are the continued threats against Taiwan, repressive policy in Xinjiang, and strategic support of Russia (whose dictator I think is far more amoral than Xi). Good things are Xi's apparent preference for advancing China's interests rather than self-gratification, his hostility to corruption, and the transformation of the country from an industrial to a technological superpower with a paternalistic but not suffocating style of governance.
Did you hear about the belt and roads initiative? Like this is the stuff that gives soft power in geopolitics. When was the last time the US did something comparable?
That is not entirely true. The central committee is extremely powerful, but so are the local governments that actually send the representative, the reason the central committee always votes in unanimity is because the discussion happen at the local government level, a local Chinese government may be ruling over hundreds of millions of people.
I'm not saying this is good, I'm merely stating that the central committee power and Xi's power is still limited by the sheer size of the Chinese state. If you compare that dictatorship to the Stalin dictatorship or Gheddafi dictatorship you'll find that Xi does not have that kind of unfettered power through sheer violence, he still needs to make a high number of cogs to work in unison to exercise it.
lol that's so false and uninformed it's hilarious
Please, corrections are welcome.
A single court ruling overturning an order from Xi will shut me up immediately.
Our pedophile convicted Felon president literally tried to take control by force, caused an insurrection and said I quote:
> In four years, you don't have to vote again. We'll have it fixed so good, you're not going to have to vote.
At least the Chinese president isn't a pedophile enriching himself at the expense of the Chinese people. China has steadily increased the quality of life for its people where the quality of life for the average American has steadily decreased.
> they do not believe that big corporations in the US or Europe would actually use chinese models
This is exactly right. The only AI that matters from an investor perspective is the AI that is built into your phone, your car, ships with Windows and Mac, and built into legal and business tools like Lexus Nexus and Salesforce.
Everything else is hobby.
That's why OpenAI hard pivoted from consumer tech into enterprise ?
The only AI that makes money is enterprise. Being open weights/self-hostable is a huge win before the pricing even enters the conversation, but pricing is also adding a ceiling on what US labs get to charge. And stuff like powering other software also works better with fine-tunes, hence legal models, etc.
European companies are more and more considering and planning to deploy Sovereign local models.
Currently the price sovereignty is still not outweighing the cost of US dependency. But big providers are already working on finding ways to ensure customers independence of the cloud act.
No worries, our corrupt politicians will soon force them to give all of our data and money to the USA, as usual.
Also the concern that open weights models could contain dormant adversarial behavior (i.e. they are sleeper agents) is more popular amongst decision makers than one would think.
In that regard it was a clever marketing move by Anthropic to release their paper "Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training" ....
Well, they do. So you review the output. Models should be treated as juniors, not a line of code or legalese they write should be used without review by a human, not another AI.
> In that regard it was a clever marketing move by Anthropic to release their paper "Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training" ....
I find such claims a bit like if 3M/DuPont released a paper on the carcinogen dangers of PFAS and someone said this was "a clever marketing move".
Like, great, this can have hidden dangers we only learn about years after it's irreversible… how does this make "our"* companies the good guys?
A lot of us were already concerned about backdoors mandated by the US government, but that was just encryption and this is everything that AI ends up being able to do.
* in the sense of "the G7" until the start of this year, because why would other nations prefer us over China in the first place, and thanks to Trump vs. Canada, Greenland, Iceland also now the rest of the G7 has to treat seriously the possibility that the US is more of an expansionist threat than China is.
I love how Anthropic has created Frankensteins monster security wise and then pointing at it and yelling "DANGEROUS" haha
Tech never got burned by China like manufacturing did, so their is a lot of naive ignorance around China's "goodwill".
burned? your business leaders willingly sold out your manufacturing industries to make just a little bit more money, knowing itd devastate many. It was your own who sold out your working class time and time again, not "big boogeyman china" pulling the strings.
...and here are all the tech boys running to use Chinese models to save a little themselves. I'd say that is a pretty clear parallel, no?
Same trick, just too naive to see it.
You realise that the grand plan of the USA is to make everyone unemployed right?
I don’t know about investors, but EU businesses with regulatory requirements are favoring EU-hosted Chinese open-weight model over US closed-weight services.
to be honest anything built into my phone, car, ships with mac is a hobby
what's important is where big enterprise point their budgets. and there we agree, it's not to chinese models (as hopeful as I am that we can avoid a duopoly and great local models)
>I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
Correct. And they're likely very correct in the belief.
When people on the internet claim that frontier labs are done for because Chinese open-weight models exist, they really don't appreciate that the people spending the money do not want to hear "we can just run it cheaper ourselves!" Any software engineer with even a few years of experience knows that phrase is absurd, and executives are simply not interested in the risks associated with such an endeavour when they can just buy the solution outright.
Consumers will have access to open-weight models just like how consumers have access to Linux. But you can just go look at the market share of operating systems to see why something being free and openly available doesn't matter to executives who just wants to get things done and is willing to pay a premium to a trusted business to do the work for them.
And none of that even touches the fact that these frontier labs have only ever been in-front of open-weight alternatives. Again, if you're an executive, and you're questioning if you should pay a premium for a better version of a tool which you presume is important for your company's ability to compete, then the answer is immediately obvious: pay the premium or risk falling behind competition who is willing to pay the premium. And when you're talking about global companies who can pay these bills without it being significant enough to even show up on a quarterly financial statement, then it's an obvious sell.
>But you can just go look at the market share of operating systems to see why something being free and openly available doesn't matter to executives who just wants to get things done and is willing to pay a premium to a trusted business to do the work for them.
That's why most of the internet runs on Windows IIS rather than Linux or BSD. That's why everyone here is running a VM on their Mac that hosts Windows containers to do development in.
> executives are simply not interested in the risks associated with such an endeavour when they can just buy the solution outright
I think you're missing two things here.
1. The margin by which a hosted product is operationally superior to self hosting an open source alternative is precisely calibrated. The SaaS provider has no reason to invest beyond the point were they can reliably make the sale.
2. Self hosted open source solutions come with a big bonus re: the ability to read and modify their code. But realistically ain't nobody got time for that.
AI upsets the apple cart in both cases. It's now far easier to manage locally hosted open source solutions, which means that in order to maintain the favorable effort ratio for making a sale, products must all get significantly better than their open source alternatives. That's gonna be a slow process involving innovative product owners. AI is getting better faster than it can happen.
Also, since AI is trained on the open source code, it's very good at integrating open source components. Previously nobody was bothering to gain that kind of mastery over their tools but now you can leverage it right out of the gate--but only if you're using projects that are in your assistant's training data. Maintaining a moat around a software product has always been orthogonal to making your product useful, but as we lean on AI more, it'll get worse. The more you protect your IP, the more its usability degrades. This was always true, but now it doubly true.
Executives may want to pay for a solution to their problems. But the companies that might provide those solutions no longer have a sustainable path.
The future of software is as a tool, not as a product or a service.
Linux is actually a pretty poignant example here because Linux is not a "hobby" or "homegrown" operating system. Every server everywhere just about is running linux. When these software companies, who by the way are the ones who will be using AI, build software, they target Linux. They build their software to run on Linux containers then deploy them to Linux servers. If not their own, then servers running on the biggest cloud providers which all use... Linux.
The piece I think we're missing is that, currently, AI is for software development. It theoretically has other uses, but nobody has found much of a way to capitalize on them. For the time being, most of the money being made is in models and harnesses for software development. There's already precedent here to use more open/cheaper solutions.
Not to mention, if you make data security guarantees or have specific contracts, you can't even use US providers anyway.
There are times I wonder if the "just use China!" sentiment is astroturfing. That'd certainly explain why there's a lot of calls for that in the face of China's long and storied history of hacking Western networks.
It's generally understood that for better or for worse, if you're working on anything (not just IT) that is security-sensitive, you're going to have restrictions on what can and cannot be done that may seem absurd at first, but hold up to scrutiny once more thoroughly examined. I would say that you don't even need to thoroughly examine this particular restriction. I have prod data at work that I can't give to humans outside of the United States. Not in India, not in the UK, not even in Canada. Clients insist on these clauses in contracts. This has been a known practice across many fields in this industry for decades.
Now we're suddenly seeing a chorus of posts saying that we should loosen up and send data to a non-human intelligence that was developed by a known geopolitical rival with a history of malicious intrusion on Western networks. Where does that come from?
There were pretty specific rumors about IPOs happening early this year, up to apparent insider bets on prediction platforms. Don't you think the apparent delays in IPOs could have something to do with the open weights/open source competition? The monetary conditions (interest rates,...) were better early in the year, arguably predictably so, so what would be your alternative explanation why they chose to postpone?
Because none of them make financial sense yet and as long as the music doesn't stop they're happy to delay and take investor money.
> the relevance that investors assign
I think there is also a perhaps the need to assign air-quotes to "investors" when talking about the US markets.
There's too much fluff and noise there. Too much (over)leveraged trading. Too much speculation.
What the real investors think is hidden behind the actions of the speculative traders.
> I think they do not believe that big corporations in the US or Europe would actually use chinese models
The Chinese models are open-weight. You can download them and run them on your own servers. Or you can outsource to Mistral, Privatemode or any of the many other US/EU inference hosters.
So the "security, regulatory and/or business continuity concerns" arguments are effectively non-existent. And whatever argument does possibly remain, also applies equally to Anthropic, OpenAI & co.
Many big investors may be too deep into this, so they prefer to extend and pretend rather than cut their (potentially massive) losses.
So, they choose to believe, lacking a better strategy. They will also try to lobby their belief into existence.
> I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
European business users (+ consumers) can access GLM-5.3 via Mistral.ai hosting. (with GDPR)
https://docs.mistral.ai/models/zai-glm-5-3
https://legal.mistral.ai/terms/get-started/
Yes, but that does not preclude the model from containing adversarial payloads so you still have to review everything.
Of course the same goes for any other model that you haven't trained up yourself.
This isn't true, especially not with the inference engine becoming public, now you can build your on models on the same or similar architecture and ship massive gains on top of these stacks.
I think people don't realize the Salesforce SAP Palantir would love to ship these models, and they have/are shipping them.
The threat of exfilteration of data via these models is high but it's nowhere near as high as the margins.
For instance if you could cut your ai input bill by over 90% would you do it? Can you put some of that extra margin say 10-20% to make it more secure.
I know of atleast 10 fortune 500 companies that are entirely planning to or have moved to Chinese models for a large portion of their work.
It's insane people in finance are delusional enough to think that won't happen.
Maybe Govt will try to de-risk with bans on using open models but I don't think that's going to work at scale, people will literally spawn labs to launder DeepSeek/GLM models into a US model by some novel training etc approach or something.
I still don't think that it's as much a kami-kaze rather it's an attempt to normalize the world around the Chinese stack.
If DeepInfer is supported by everyone, it will be much easier to export stuff aka hardware that only supports it for now or is optimized for it.
Assume china starts selling hyper optimized Asics for inference at some point.
This feels like a race to the bottom as soon as possible because they have much lower costs, if there is a Wipeout in AI Labs in US chinese labs are largely insulated, and they can hopefully buy Nvidia chips for cheap if that does happen.
Some activist VCs in Washington lobbying for dumb shit again
Because they can and there’s no consequence to drumming up fear, same shit since the days of Napoleon and Nelson…
If US govt bans it, I will move my business out of US lmao. I say that as non us resident who has business there for ease of doing business...
If you believe those investor, I assume you mortgage and bet all your wealth on it. SpaceX? Anthropic? Chinese has talents, electricity and Huawei. I really believe those investors are tulip blind to pile money into those western AI. China even now world number 3 ssd+ram ahead of SK Hynix. So armageddon of RAM SSD LLM model at dirt cheap price. And as time progress, large ram and small gpu will be built into handphone (which again dominate by Chinese).
If they are open sourced, it means any random data center could host it domestically and AI becomes a utility.
Europe is going to be just as worried about using US models for large parts of government related functions and they're going to feel alot better about open source self-hosted chinese models than US models which can literally get turned off on a whim by Trump. They'd prefere home grown models if they can, but if they can't they'll go self-hosted and that'll be Chinese at this rate.
Exactly and that is precisely what is happening.
The US will move to ban it for most major businesses in some capacity.
You're entire company would become ineligible for gov contracts if you have use DeepSeek. Something like that.
Seems a toss up whether Europe trusts US or China more.
Hardly a toss up based on this. Kind of stunning to see how much damage has been done in less than two years.
https://www.pewresearch.org/global/2026/07/15/people-in-many...
> I think they do not believe that big corporations in the US or Europe would actually use chinese models in their business operations due to security, regulatory and/or business continuity concerns.
Well they are wrong, I'm at a large EU startup and they are currently considering Deepseek as their default model, Opus/Fable would still be available for complex tasks.
lol big corporations are 100% using Chinese models in their business operations. They dgaf about security, speak to anyone in cybersecurity or cyber insurance
The labs pushing this cyber scare is an attempt to cry wolf. The corps don’t give a fuck
There's very little understanding of business operations on HN, especially outside the tech echo chamber. The "software will eat the world" mindset.
Based on HN comments, every company would be running Gentoo and developing line of business applications in Haskell or OCaml.
I concur. Everything is such a hard left turn.
The reality is what ever better and cheaper is where the money flows. Especially from an international corporation perspective.
[dead]
That's exactly what people said when Deepseek V3 was released. Two years later the valuations roughly have quadrupled.
"valuations" are irrational and make dot com _valuations_ look like a miniature caricatures
What dot com valuation would you compare Anthropic or OpenAI to?
The classic dotcom example is pets.com. Would you liken Anthropic to pets.com?
More like boo.com - I know, that's old folks.
Except the revenues have moved in lock-step with valuations
Valuations mean nothing to me, I'm not in the investor class that gets to buy and trade before IPOs, nor do I plan on putting a big share of my wealth in tech behemoths that everyone else is hyped by.
Rather than caring about whether it's over for western providers, I look at what competitors for the western providers have been doing since their first moment in the spotlight. And what I see is that their offerings keep getting better and better, and that at the same time they're opening not just their weights but their architecture and experience.
As a customer (rather than a pundit), I just like their tools and the user experience more than I do those of hte US giants. I admire the US offerings in many respects, but also find the user experience has a lot of friction and annoyances. Deepseek et al don't always deliver the same immediate 'wow factor' but when I sit down to build something, I enjoy using their tools. After 40 years of using computers for fun and profit, I will always always bet on a product that's fun to use.
no one has had to pay back any of the loans yet. the valuations right now mean precisely nothing.
if these hold a year after an IPO then we can talk.
"the market can stay irrational longer than you can remain solvent"
Whenever I see this, all I hear internally is "I'm not willing to put my money where my mouth is"
A better interpretation is that there is a window within which you can make money on an unconventional view. You can't be too close to the market's conventional wisdom, but you can't be too far away from it either.
My mouth is at: I don't want to be exposed to anything related to Musk.
My mind is at: Even though it's overvalued, it will likely keep climbing and won't crash.
Where is the market at?
> Even though it's overvalued, it will likely keep climbing and won't crash.
Then...it's not overvalued?
Unless you mean "it will crash relative to inflation over 20 years, but not over 5 years" or something like that.
Because we are still at the first stage of enshittification
> That's strategic sabotage to say the least.
It is clear from the beginning that the Chinese Open Model are a proxy for economic warfare. Anyway their singular goal is to dethrone US monopoly on SOTA model and burn all the economic hype and investments. If this strat works, which we will see after Anthropic/OAI IPO, we'll be in for an interesting ride
I am of course eager to pay needlessly more to support poor Anthropic/OpenAI.
> It is clear from the beginning that the Chinese Open Model are a proxy for economic warfare.
I mean, what would you expect? Like USA isn't openly planning to weaponize their current AI primacy [1, 2].
[1] https://www.whitehouse.gov/presidential-actions/2025/07/prom...
[2] https://www.whitehouse.gov/wp-content/uploads/2025/07/Americ...
Well, what can I say? If you play dirty, then so can everyone else. The way I see it, consumers are the real winners here. Just another variation of classic competition. If such open models didn't exist, then make no mistake, OpenAI and Anthropic would be making sure you are bleeding dry with your wallet without much innovation.
Imagine calling competition economic warfare
Way too many monopolies these days
It's the USA, you can just ban it.
https://public-inspection.federalregister.gov/2026-21447.pdf
No, that isn't going to happen. The US can't afford to lose the AI race and US companies that can't compete will get thrown under the bus and replaced with those that can.
The DHS order only bans buying DeepSeek products directly (e.g. doing inference on their China hosted servers). It does not apply to open weights at all. And DeepSeek just gave their response to the DHS order.
No it doesn't. It bans:
"third-party products and services developed using DeepSeek implementations".
So gov can't use western inference of DeepSeek either.
"DeepSeek implementation" in this context means a third-party product that uses a DeepSeek product, e.g. the DeepSeek API. This does not apply to open data running on a non-affiliated third-party inference provider.
footnote 10 says it includes 3rd party implementations of DeepSeek
I agree with you - but there is also an ambiguity, the text of the the order also seems to ban any products coded with DeepSeek
No, it is being interpreted that way by people who either don't understand basic legal writing and the legal standards for an order of this nature, or by people [who] are purposefully FUD'ing in order to take advantage of that lack of specialist knowledge.
I don't understand basic legal writing, you you may very well be right. I would be nervous about fulfilling a contract or building a company with, for example, my code base, which is heavily built with deepseek
I understand basic legal writing very well (I went to school for it, though I don't practice) and I think you're dead wrong. 'third-party products and services developed using DeepSeek implementations' (used consistently in the text of the rule) is deliberately broad and indicate a refusal of all derivative products, not just offerings of the DS models provisioned on OpenRouter or suchlike.
Now, most software products developed with DS can pass an audit, whether human or from a US AI provider, that they do what they say and aren't secretly funneling telemetry to Beijing. When I say 'most..can pass an audit' I mean that most products are simple enough to undergo such an audit, rather than implying that some have exfiltration mines buried within; it's possible but I have have no reason to think it's probable. I believe the purpose of the broad wording is to facilitate the waiver requirement, which costs money to apply for and will require payment to an accredited or government approved audit provider, thus generating DSP on the US balance sheet even if the original prototyping/buildout was done with Deepseek.
Yeah, but they can't ban the rest of the world from using it :)
You could claim you're the more ethical of the two countries and hope people boycott it
ahahahah
it would take a regime that doesn't bomb girl schools for that to be effective
That works. But make sure you don't say it outright because that might make people stop and think about it for a moment.
That's right, they could
Exactly, the EU only deals with ethical countries that support Ukraine
ahahahah
Especially not after spending the last two years alienating it.
Isn’t this a ban on it being used by the state, rather than people/civilians?
For now
From government and regulated places perhaps but a large portion will not care
Thats like sticking your head in the sand :)
It's a project based on vLLM, explain why you consider it a total game changer?
Can you explain how this removes their moat? Genuine question, I'm not sure I understand enough about LLMs to get how an open source inference engine undermines them.
Today the closed-source giants maintain a substantial cost-per-token advantage due to proprietary inference tricks like speculative decoding kernels, ultra-optimized FlashAttention derivatives, multi-GPU pipeline orchestration and dynamic prefix-caching algorithms.
I'm not sure whether the DeepSeek's open-sourced inference engine does all of these, but it appears that plenty of inference tricks have just been revealed for the open source community to learn from and start improving upon.
So... this makes it easier for someone in the position to run chatgpt class models as a service to get similar returns to openAI/anthropic, if I got you.
And I imagine this engine is by default mostly just deepseek compatible and inference engines are not universally applicable to all models?
It sounds significant for competition but I'm not sure how it sinks Anthropic/openAI. Presumably they still have their frontier models, the training data they're vacuuming from current users and access to hardware to serve at a ridiculous scale, right?
DeepSeek can accomplish over 90% of tasks of equivalent value at one-tenth of the fully loaded cost of O/A, and this capability can further spread through open source. In my view, the IPO risk for O/A is no longer just a valuation discount, but a substantive challenge to its long-term business model as a high-margin foundation model provider.
In that scenario, even if global Token usage increases 100-fold in the future, that alone would not be enough to justify these foundation model companies being worth one or two trillion dollars.
I think you are underestimating the value of that last 10%. Returns to intelligence are non-linear. The 80/20 has been true for over a year now, and it has not impacted the labs revenue.
What is the outlook for the hyperscalers?
I think for me it's more the possibility of the ideas here being incorporated in other Opensource projects (vllm, llama.cpp, or in the architecture choices of the models themselves)
Deepseek manages to be ridiculously cheap and efficient with prompt caching, i spent 550M~ (1.9b input cache miss, 1.7b output) tokens and it cost me 5~ bucks, that would be ~150 bucks on Opus 5.5, if i ran that workload in parallel that would be $50 vs $1500, e.i. i wouldn't want to subagent heavily with anthropic models.
now imagine Kimi K4 or GLM 5.5 with those cost savings
I think you’re correct, although suggesting that open AI is a trickle drown of AI companies is a bit naive.
But the key point is that something good enough always attracts more interest and contributions because a productive outcome is very realistic and tangible.
Open Weight models and competing open source inferences engines massively undermine monopolies; in theory at least.
See how cloud is easy to divest and “nationalise”.
All of those are in open-source engines. I'm not convinced they have that much of an advantage.
[dead]
What moat?
It is, and always has been, data contracts. Which is why the hyperscalars are a more likely candidate to take American frontier business than Chinese vendors.
I can click a button in Azure foundry and run open weight models that are approved through my compliance and procurement teams.
I don’t know how open sourcing this stuff helps deepseek but it means I’m getting ever better and cheaper models so I’m fine with it.
Well, whatever real or perceived moat Openai/anthropic had. OP claims this open source inference engine reduces their chances of high valuation, I'm assuming this is due to their in house engine being something they could use as differentiation before?
That is probably what has been the most fascinating part of this AI/LLM boom, the moat was apparently never the technology. It has been very much like nuclear energi (and weapons), once you know how to do it, it's not that difficult. It's still fairly hard, but you can do it. The moat has been the funding for the infrastructure and if that dries up you're in trouble.
Yeah what moat is the right question to ask.
I've been trying to find an answer to that, too. The best I've got right now is this:
Moat-like things going for them:
- access to the US government (for huge inflated government contracts as a source of free capital, having the competition banned, having their products legally mandated, getting other preferential treatment, having foreign governments bullied into granting similar advantages)
Too unsustainable, too fragile, too small:
- Strong brands: Brand loyalty seems like a pretty insignificant factor in AI right now
- Semi-exclusive hardware access, RAM buyouts: Unsustainable even in the short term
- Lock-in effects: Their models aren't very stable across versions, or very unique, and switching providers is about as easy as it can be in a highly competitive market.
- Exclusive technology that is important, advanced and hard to replicate to a degree that would materially set them apart and justify potentially much higher cost. Maybe I'm missing something here, but a lot of the claims they make to that end fall apart even on light scrutiny (due to things like test sets that are small, hand-picked and clearly overfitted on, or broad general claims that cannot be substantiated). I'm getting the feeling they don't have anything that materially sets them apart at all, but that's seems impossible to measure in any meaningful way, so who knows, they might have something here.
In theory, this can allow a small US lab to compete with O/A/G and take marketshare from them with small open models deployed centrally or (and this is think is the real moat) on-prem for enterprises.
When the model is fungible and the switching cost is near zero (business logic lies in the harness), then there really isn't a moat for models.
If logic was in vogue then this would be true and SpaceX wouldn't be worth 2T dollars
Considering how absurdly far ahead of everyone else SpaceX is, 2T seems low, and is more of a testament to how immature the space market itself is at this time.
If SpaceX was reset to 2016, they would still be ahead of everyone else even today. Their deployed capabilities today are ahead of the plans of most of their competitors.
There are only a few industries where being absurdly far ahead of everyone means you get a 13 figure valuation though.
Global space launch market is 6.5B in 2026. Even the total global satellite market is only 23B in 2026.
You have to assume some pretty large numbers for YoY growth, market share, and profit margin to get to 2T valuation.
even if they are far ahead, where does the revenue come from? current market size is like 20-40 billion annually.. even if it grows to 2x it's still a long time until that would justify the valuation.
but perhaps you are thinking that starlink is the key? sure probably a bigger market - but any market it accesses is bound to eventually prefer ground lines.
so the only way to justify the valuation is if the AI side of the business becomes the most dominant AI business
China is catching up there too.
Anyway, where's the revenue from space going to come from? Obviously not data centres now that Deepseek is commoditising inference.
The telecom business? Amazon is catching up there and Europe/the rest of the world is racing to build capabilities to not be dependent on Musk.
China isn’t even close to where SpaceX was 10 years ago.
They may be catching up, but not anytime soon.
It doesn't make sense by any standard valuation metric. Yes SpaceX is the best company in the world at getting things into space. The valuation doesn't reflect the actual size of that market at all. It's assuming AI in space will be world-changing, which is a huge leap to make.
VC and valuation (alone) backed success has never been dependent on a business or product being good. All that is needed is finding the bag holder(s), and the way our economy and financial systems are designed that's the core competancy of this system.
That’s why we will probably see Fable 5.5 this evening.
US companies are throwing money at the problem, and considering how good Opus 5.5 is I’m not too concerned about them.
Interference optimization is a solved problem with no moat. Anybody can achieve this result, optimize for a particular hardware, optimize for a particular problem. In general, software is no longer a moat.
The moat is custom inference hardware, lower power, higher throughput, etc.
Yup. Hyperscalers are going to eat them alive. AWS/Azure/GCP hosted instances of cheaper-but-almost-as-good models like Deepseek are going to eat OAI and Anthropic alive.
They'll need to release cheaper, better models to compete. Which may not be possible.
I wouldn't say yet its Kami Kaze.
The overall market will still be big, very big, if it replaces real people or augments every single worker.
Some contracts will go to your local provider (US Gov, etc.).
And there is still a moat between them and DeepSeek.
As long as i can afford it, i use Opus.
I'm not bound by it and switching is very fast
It's crazy that a ~400 person side project by a Chinese firm can put up such a fight.
All other models are also either by <1k labs or side hustles by tech companies who are not particularly known for LLMs
Their success is subsidized heavily by distillation, and even then, the Big Labs are no more than an OOM larger.
[dead]
Isn’t the Muse team just 200 people? I guess you don’t need that much work done.
Not really. Anthropic’s headcount isn’t huge, and only a subset of their employees are working on the frontier. It’s also easier when you’re a follower.
I don't gamble, but if you really believe this, there's apparently some serious money to be made.
https://polymarket.com/event/anthropics-valuation-end-of-oct...
These have not really budged (yet) on this news.
Aren't they betting on the valuations like a week from now?
It's possible to be correct that the AI companies have no future, but not be able to time the exact month they collapse.
Present valuations are an aggregate of what investors (and speculators, in the case of the secondary prediction market) believe about their future.
If by 'collapse' you really mean a 50%+ fall, timing is not going to make or break you by any means.
You linked to a betting market that ends in a week and then said that timing isn't important. I don't understand you.
Let me spell it out.
If Anthropic is going to crash 50% or more at some unknown time in the next two years, and you have foreknowledge of this, you can easily make a massive profit by placing a bet weekly. The odds are VERY long per week, so on each losing bet you only lose a tiny amount, but on your winning week you will more than make up for it.
Timing is, as a general rule, important. In the particular case we're discussing, the odds of a 50% swing are so low that you can tolerate being VERY wrong on the timing if you are correct on the swing.
This assumes all other information remains exactly the same as it is currently. That is not how the real world works, however. As time goes on, and it becomes more obvious that the company is floundering, the "long shot" will no longer be such a long shot, and will not have as high of a return.
idk, if "prediction markets" remain a real thing, there should probably be some requirement for previous experience in options-adjacent markets to ensure people actually understand the economics of time-valued positions.
there's hardly any volume there, so I wouldn't take it seriously. The last transaction was 4 days ago
No, that's open source and it will be an accelerator for everyone, instead of a small advantage kept by a few. Even the frontier labs benefit. I'd argue (like I did when the initial DeepSeek was released), that the AI market *needs* to those inference breakthrough.
Can you say more? Why would the inference engine be the secret sauce that would protect trillion-dollar valuations? Isn't the inference side somewhat of a known-quantity, while the training is more of the lab's trade secret?
Where are the competitor to the big frontier labs? The Chinese providers are demonstrating everything can be done cheaper and practically identical. I'm surprised upstarts aren't showing up.
Deepseek came out of a hedge fund, right? Do they just bet against US stock and release for free a pile of stuff that basically shatters them? I never seen anyone saying that, but it would be an awesome business model.
I hope more hedge funds start doing that. Basicaly funding open development to profit from tanking uptight super serious American stocks, to fund even more open development.
If you can't beat them, give the common folk means to skip them and profit from their fall.
I'm betting that China is betting their government can better handle the massive unemployment these will create within the white collar labor force.
This don't need to be Chinese hedge funds. Western ones should be doing that as well.
Drop a few mil on tokens to make enterprise grade database to put out for free and bet against Oracle. Or train a model, or something.
>That's Kami Kaze for US frontier labs.
Well deserved too! The chutzpah of trying to peddle the AI equivalent to knives which refuse to cut flesh because it might be human flesh... astounding, and utterly disrespectful to their customers. Claude's guardrailing was so bad during OpenAI's HF attack that HF had to go to Beijing!!
I'm hoping Anthropic and OpenAI drop like stones on IPO. Never thought that liberation in this space would come from the Chinese!
This is not to rebut your point, but I think for a lot of people/teams the silicon valley "lose money until we make money" pricing from OpenAI and Anthropic will continue to win (assuming the numbers below aren't completely off).
I'm out of my league on knowing if these ChatGPT projections are close, a little off, or way off, but we currently pay about $5k per month for Claude for our whole team. The projection for "run Deepseek on AWS to replace Claude" is coming in at $30k+ for GPU hours.
> That's Kami Kaze for US frontier labs. There's no way from here for them towards trillion dollar valuations.
I don't understand the reference to Kamikaze in this context. Is Deepseek destroying itself in order to deep-six US frontier labs?
0. https://en.wikipedia.org/wiki/KamikazeDo not under estimate the retardation of investors
There will always be special demand for the "best possible" and "fasted around" inference and they will pay. "Good enough" models are useful for many tasks but cannot replace the whatever frontier is at the time. That's why the top 2 labs are rake in ~80% of total inference income.
I don't see how this proposition holds. It's just part of an inference engine for serving DS4.1 Flash. Which vLLM and SGlang can already do. Less efficiently, I'm sure, but on much broader hardware, and with other models in the mix.
People will take the novel bits here and port them to the above. The community as a whole will benefit. Including infrastructure engineers inside OpenAI and Anthropic.
But a lot of what's in here is already in published papers out of DeepSeek already.
Sure, any day now DeepSeek is going to destroy the valuation of OpenAI and Anthropic. I've been hearing this since DeepSeek first launched in Jan 2025, 21 months ago. And with every new model generation, both open and closed, I hear a chorus of "this is now good enough!". And then I hear it again for the next generation.
Meanwhile OpenAI has grown their revenue like 10x and Anthropic has grown theirs like 65x in that same time period.
There's no cap on what you can do with frontier intelligence, nor how much of it you can use. Yes, DS and other open weight models will continue to grow like crazy (as long as they can source compute), but so will the frontier labs. I ran like 20 agents overnight last night on Opus, Fable, Sol, and Astra last night, trying to use up quota before my weekly reset, and I'd gladly have done 10x that much and used agents 10x more intelligent if I could have. And in six months I probably will.
how? openai can copy some of the ideas on top of their own. Smaller labs gain this tech without much development, but is isn't like they will magically beat frontier labs. Is a smaller deal than you think vs kami kaze.
That’s a bingo
Has anyone figured out why Chinese billionaires are releasing these models? Do they hold short positions on US markets? Why would Xiaomi, et al. release these models in the first place? What is the strategic purpose here? I cannot believe for a second that they are doing it for human good.
One very simple reason could be that the US has been very vocal about China being its geopolitical adversary (they may be right).
People only use Chinese models because it is opensource.
I think people in China (huge market and many people) would use them even if they are closed.
Best leveraged debt on the US economy ever.
This really not that hard.
Deepseek started as Liang's side project funded with his hedgefund couch change compute. That set the economics for everyone else in PRC. If deepseek releases for free because Liang doesn't give a shit about making $$$ on his hobby, every other PRC player has to follow. And serendipitously, due to US export controls, PRC tech companies can't dig themselves into ruinous seppuku debt buying compute even if they wanted to - what they can buy, i.e. everything around DC infra including power, they get for fraction of the cost expediently.
TLDR Liang said fuck profits... which means every other PRC competitor had to fuck profits, which ends up fucking profits of western labs so deep in debt due to rampant financialization and can not afford to say fuck profits.
Sometimes a billionaire is fine just burning $$$ on a hobby while remaining a billionaire instead of trillionaire, and it just so happens the side effect might mean preventing others from being trillionaires as well. Bonus others is geopolitical adversary who is hedging economy on trillion $$$ returns.
I'm still not understanding why one hobby projecing being open source means others should also be open source. So you're saying that PRC is making a bet against US companies and the US markets to crash them in order to stifle the US economy? That is an interesting take, I would think the US market is far more resilient than just the tech stocks. Speaking with my communist friends though, it does seem like this is the ultimate goal: Topple the US hegemony and US foreign control by crashing it's economy while China makes a bet on free/cheaper AI models, with the ultimate goal of agitating the people enough to cause a revolution to turn the US into a communist state (really delulu stuff if you ask me)
Seems kind of pointless? Wouldn't anyone in the US/western world simply use the chinese models in their own inference? That would not strengthen the Chinese position because the models got downloaded for free off HuggingFace, and then cloud inference bills would just put money into US companies.
Which of my words implied any of this?
One person's hobby started the industry in PRC and set the trend and inspired talent in nascent field with no established business model. PRC competitors didn't want to FOMO and just knew they had to compete with opensource and everyone fast followed. They bought what limited compute they could, hired what talent they could and just told them get to work. There's a reason even Chinese Uber / DiDi was releasing open source models. There wasn't some grand coordinated posture to compromise US markets, it's incidental to PRC industry being PRC industry. Very few in PRC thought they could close gap this fast with compute deficit. US chose to continuously dump venture $$$ and company warchests quarter after quarter on their own to increase exposure. There's no grand conspiracy to turn US into communist state (wut), there is not "bet", just typical PRC domestic involution business shenanigans clashing with US financialized VC shenanigans. In alternate world if PRC models not competitive, open source wouldn't even matter to US valuations. In alternate world if PRC models competitive and US AI isn't leveraged to shit to multiple % of gpd, i.e. systemic contagion risk, it also wouldn't matter.
It doesn't matter if US/west makes $ off PRC models, PRC doesn't have enough indigenous compute let alone spare to meaningfully compete globally. In so much it matters economically (again happy happenstance for PRC) is denying US AI making $$$$$ while spending $$$$ vs US AI making $ while spending $$$$. Extra happy happenstance that $$$$ US spending on is basically all the VC money and FAANG warchest stockpiled by US tech over decades... i.e. US spent so much that it meaningfully constrains private money to spend more in future. Meanwhile PRC AI spending $ to make $ (or less), but still have $$$ in reserve for other shit. It also doesn't help that half of business model depends on building infra on tight schedule that US is... not great at.
Ignoring all the slop out there, we are seeing AI being sent to reverse engineer, accelerate projects, etc.
According to Greek economist and former finance minister Yanis Varoufakis, an open source AI strategy is one way a capitalist economy can avoid the rent trap (platforms being created, monopolizing resources, and extracting ever higher rents from everyone else).
https://www.yanisvaroufakis.eu/2026/08/31/ai-is-reviving-chi...
[dead]
TARIFFS! LOL
[dead]