718 comments

[ 0.18 ms ] story [ 56.4 ms ] thread
tldr: the marketplace for opensource models became wide open
What is the business model of hugging face? My understanding was they were basically just a file hosting platform.
Isn't this pretty much the model for any tech company? Rush to get MVP released, stoke the hype, cash in on FOMO.
I guess we may have to start paying to download models.

Or perhaps they will start throttling downloads for free users.

I don't know what they business case is, it might be to shut them down: I suspect good free models on local hardware is a threat to Nvidia's investments in OpenAI/Anthropic.

They have various paid services that relate to AI—paid inference hosting; paid accounts aimed at AI development with GPU rentals [credits plus paid overages, I believe], more private storage and public storage than free accounts; and many of the community a and some other benefits on individual/team/enterprise tiers; additional paid storage above the base quotas for the paid account tiers; on demand rentals of HF managed containers on GCP and AWS, and some other things.
Didn’t we just witness it? Become a critical ecosystem player then get acquired based on the value of that position.
Possibly, but they were profitable already.
This. They are buying a brick in the xkcd #2347 diagram
You missed the part about get hacked by one of the cornerstones of the ecosystem so that you could get infinite bad press for them if you wanted.
They don't need a business model to have value here.

They sell a few cloud services, but the most valuable thing in their portfolio is live, engaged accounts from pretty much every person directly interested in running AI models.

If you were an org positioned to do sell better cloud services on a wildly better margin, wanted to promote consumer-binned AI hardware from your manufacturing chain, or wanted to promote new AI enterprise or research-support initiatives, those accounts would seem to be an absurdly fantastic collection of warm leads that you could gradually mine and upsell (and critically: that your competitors could no longer buy/use themselves).

(comment deleted)
If GLM-5.3 was truly trained and ran on Chinese chips, we'll have to ask what the business model of Nvidia itself is soon :)
it's literally the GitHub for AI
Github was acquired for $7B. It's hard for me to see how HF is worth almost double Github.
GitHub was acquired for 10B in 2026 dollars. And software on HF requires quite a bit more hardware expenditure to run.
AI is the current thing, HF does the current thing. Hype cycles in computing anyone?
Here is an employee of HF at a time answering how do they make money, supposedly [1]

> We make money via compute credits + Enterprise Hub + HF Pro subs [...]

I guess this plus custom inference deployments, external inference providers...

[1] https://x.com/reach_vb/status/1928050126498713706

an FTP server delivering tarballs.
Having an account of practically every software engineer remotely interested in AI....Profit
Hold on, these say no deal has been reached yet.
the title of this post is misleading… “in talks” != not “agrees”
It’s not misleading according to the The Information piece (second link)

  > Nvidia Agrees to Buy Open Source Model Repository Hugging Face For $12.9 Billion
The story is in an intermediate state right now, but as semiquaver points out, The Information is reporting it as fact. Since their reporting tends to be as reliable as it is hardwalled, we went with their claim in the title.
I hope they're good stewards. This reminds me of the Microsoft GitHub acquisition in terms of the importance to the broader community.
As someone who've played video games and obsessed over graphics cards long before this all started, don't get your hopes up.
as gamers, we got burnt by nvidia way too many times. I'm still holding grudge over shutting down 3dfx.
3dfx had their own API which locked in game developers. This was far worse than anything NVidia has tried
have you heard of PhysX, Hairworks, or the CUDA moat?
3dfx hardware didn’t run anything but Glide very well. It did run some OpenGL after iD twisted their arm, but it was a really crummy implementation

While ATI and Nvidia never tried to implement a Glide emulation layer, several hobbyists did and 3dfx shut them down with legal threats

Were there any legal barriers to ATI/AMD or Intel implementing CUDA or PhysX on their systems?

You Jews just make shit up lol

No gamer had a problem with nvidia.

Many bought their mobos and video cards.

Mac was pretty hated, Steam was very hated, we didn’t bow to any software company.

Nvidia actually made hardware.

Your propaganda is so awful because it’s always obvious you have no experience

Its becoming impossible to track all the acquisitions.
Anyone old enough to remember anti trust laws or even just laws?
sometimes it is comments like these that make me wish hn had a block button
The AI judge doesn’t have that feature yet
Laws are for poor people
Please don't post unsubstantive comments to Hacker News.

No doubt you have the nucleus of a substantive comment here, but that's not enough. If you only post the shallowest top stratum of what you're thinking, other people do the same, and then we get "Laws are for poor people" and endless descending repetition. The whole point of this site is to try for something other than that.

https://news.ycombinator.com/newsguidelines.html

p.s. Also, please don't be snarky on HN. That's also in the guidelines.

You are right but its the initial shock on bad news, hopefully this will spawn alternative competing services.
I thin HF holds huge power in how they’ve consolidated all the open models and data, but monopoly feels like a stretch here. They are in reality a fairly small startup with lots of alternatives (all of which are much worse at this point granted) and are mostly a loss leading part of the open source ecosystem. I think it would be very hard to make a case against this on monopoly grounds, though I wouldn’t be opposed to someone trying if they thought it was possible.
The companies have aligned incentives so I don't think this is a terrible acquisition for customers.
Vertically integrated monopolies naturally have aligned in incentives for themselves, which allows more capture which turns out bad for customers. Whether it is a monopoly is an open question.
This isn't vertical integration. This is a commoditize your complements play and that does benefit customers. Nvidia benefits from supporting the open model ecosystem, if models are a commodity the gpus become what's scarce.
Someone with financial / decision making weight their is going to ask why they are commoditizing models that run on "others" GPUs .... Since doing that makes their gpus less of the scarce resource..
Oh, I totally agree, this alone is not vertical integration, but can be a path too it, long term being able to vertically integrate undercutting others that would be on top of the chain then retracting the open model.
>Vertically integrated monopolies naturally have aligned in incentives for themselves, which allows more capture which turns out bad for customers.

Apple's devices have resulted in a great increase in quality of life for myself and family members who used to waste a lot of time with Windows drivers/BSODs/random updates/malware/etc.

It's been almost 20 years, and it's turned out OK.

of course it is

works fine for now

whats nvidia gonna do except make it worse?

usually buyouts go something like this:

1 buy company

2 fire various people

3 enshittify

hugging face even said they didnt want to accept a 500 million dollar investment from nvidia, because they didnt want nvidia to run the ship.

instead they sell it... guess who'll run the ship?

https://techcrunch.com/2026/08/24/hugging-face-reportedly-in...

I’ve asked ChatGPT etc a few times how Hugging Face makes money. Never quite got to understanding . Good for them.
Gemini did a good job of explaining it. fwiw, it may have helped that i accused Gemini of leading me to a "scam site" like "mega dot com"
You could just go to their website and look at the pricing pages? Why are you asking an LLM?
Because I never ever see anyone talking about using it, except for occasional reference to weights.
I really wonder why Nvidia would do this.
Owning the roots, the trunk, the branches and the tip of the leafs.
Keep alive the on-prem hardware sales? Not as big part likely but still something good to have. On-prem hardware needs models to be available to be useful. So pushing those could sell more on-prem hardware.
Ai companies are making their own chips, Apple unified and other unified ram solutions also a threat

Nvidia is making tons of money by selling GPUs at a very high profit margin. They will do anything to remain dominant

Why would the most rich and evil company on the field want to own a practical monopoly?
I'm puzzled by the reactions. This is not a good thing. Yes, they're releasing more open source code and models, but they do it BECAUSE OF pressure from the community and other open source projects.

But Nvidia is a terrible open source and consumer company. They gatekeep a lot and oftentimes it's only open source in name. Outside contributions are often slow-walked or rejected if they don't align with business incentives , and leadership is retained 100% in a couple of people from a certain country.

It also means that finding models that aren't "approved" will be dealt by other websites.
Potentially horrible for monopoly reasons, but if other big acquisitions in the AI era teach us anything, developers are about to get a whole lot of free and discounted trial credits.

That’s at least a plus. I will happily burn through as much VC money as they will give me to tinker with my projects.

Yes. Great for devs, for a while. Awful for overall competition and industry's health.

I think the federal antitrust regulators are asleep.

Edit: antitrust regulators' job has just begun -- we'll see how this deal gets adjudicated by the FTC (if at all).

You can torrent and seed these models via torrenting. They are not piracy so there's no problem hosting them besides maybe not having enough bandwidth
Interesting ideas, maybe the real open alternative will indeed be torrents.
... You wouldn't steal a car ... you wouldn't steal a bag ... piracy is a crime. (Girl cancels the download and walks away from the PC. Remember? :-))

That is what will happen to torrents. Followed by VPNs. China, Russia and Iran had paved the way already.

Can't you see the puzzle coming toghether?

Ai models are not piracy, because you never had to pay for them and simply seeding them and downloading them does not break their licenses
> I think the federal antitrust regulators are asleep.

Have been for a long time. All of the big techs should have been broken up long ago.

Current US government is pay to play. As in buy my cryptocurrency and you get a pardon or case dismissal, no backroom deal required. Payment directly. Blatant corruption sucks.

If the House or Senate turn that might actually stop or slow down. The market falling by 30% when the AI bubble eventually pops would also trigger it.

> That’s at least a plus. I will happily burn through as much VC money as they will give me to tinker with my projects.

The era of VC-funded home-delivery recipe boxes is still my favourite.

Yeah this is the "early netflix" era. Take advantage while you can and plan for the future.
Counter point is China can afford to keep pumping out free models for long enough to crash the AI industry, As well as spin up memory production.

Its likely in the near future we will be able to buy 128gb mac minis and run local AI for free.

Why would China want to crash the AI industry? Their goal is not to damage the United States, necessarily. Their goal is to have the best domestic AI on the planet. They intend to do that through the way that they dominate every other industry: good enough quality at a much greater scale.
Surely crashing the Ai industry (in terms of reducing commercial revenue) makes it harder for the leaders to keep raising money and so keep investing in r&d to the same levels. Thus allowing competitors without access to latest American hardware to keep up.
When did apple stop charging more for their memory upgrades then gold by the ounce costs?
They didn't, but their price increases haven't kept pace with general memory price increases (yet) so currently their pricing seems reasonable.
That means we will torrent these ai models. They are not pirated so there's no problem doing so
True, but better nvidia than any other big tech contender. Nvidia has a very strong business incentive to make huggingface thrive, whereas pretty much everyone else has the opposite incentive.
Especially when all major labs are building their custom inference chips.
Yeah I think this is probably one of the best outcomes you could hope for if you want to use open models.
I think this is good.. They have incentives to keep things free to keep people using their GPUs. I think that's one of the least enshittifying outcomes possible
The Government plans to regulate open source ai. Nvidea is dependent on the Government to back loans, provide property, and make laws in their favor. Now they have ownership of what the Government wants to control. A perfect wedge they can use to get more from the Government.
100x this: "The Government plans to regulate open source ai." and ... s/regulate/restrict
I have never seen an acquisition be good for the users. I struggle to believe this one will be different
Acquisition of GitHub let to increased downtimes, but also to unlimited private repos for the free tier.
... Which gitlab already had before github. And still does.
The unlimited private repos for free tier would have happened anyways. You have to remember that this was around the time people were jumping ship from GitHub exactly because they didn't offer private repos for free. It was something many wanted, and they were actively seeking out other places that would give it to them.
I'm not convinced that's the case. GitHub was like that forever, they were well established and the change happened not long after Microsoft bought them. There was no global shift to gitlab instances or Atlassian as far as I am aware.

You can dislike an entity or an action and still acknowledge it also had some positive effects.

Overall GitHub still got significantly worse since the acquisition. Downtime was essentially zero before.
Usage was 5% of what it is now before.
I only ever used GitHub for private (and a few public) repos, so for me personally it got significantly better, being able to have private repos for free.

Pretty bad tradeoff of course, improving the situation for non paying people while making it bad for paying customers.

Truer words were never spoken. This is painfully true.
The fit is theoretically good (nvidia is model agnostic) but I can see then pushing the ecosystem increasingly towards cuda and some custom nvidia software packages for inference etc etc
Yeah I'm afraid they will only keep cuda support for their software. Good thing it's all Open source
I'm sure it'll be fine. As long as you have the latest 90-class card you'll probably continue have nearly full functionality from the HF libraries. (Upgrade to enterprise gear for full functionality)
Apple buying Workflow (became Shortcuts). Its users have a better, faster tool and got it soon after acquisition. And it has many more users now.
Was Workflow available on every platform running on every hardware provider?
Is huggingface profitable?

I'm not a fan of big-tech acquisition results either, but one benefit can be that a product continues to exist when it would otherwise become insolvent.

Nokia buying Qt in 2008 brought an LGPL option.
What’s with all the big acquisitions recently? Is it advantageous for some reason to announce all of this stuff now?
The VC capital for AI is probably drying up. Now the question is: who will buy OpenAI and Anthropic?
No, its just the final stage of the bubble before it all comes crashing down. :)
What does Linus’s hatred of NVidia drivers imply for this situation?
The other hater is Apple, Linus’s isn’t alone…
The open weights AI ecosystem is way too concentrated with HF. That said, it was clear they were going to either get bought or aggressively monetized at some point, Nvidia seems like it could be one of the more aligned acquirers.

I do worry about any sort of crowding out or downplaying non Nvidia-relevant quants, and about changing rules to crack down on models or datasets that for one reason or another “don’t align with their corporate values” - uncensored etc. Someone mentioned Microsoft and GitHub, they appear to me anyway to have been very hands off, I hope it’s the same model.

We also have modelscope.cn, I'm sure that one won't get acquired
It's already part of Alibaba, which has its own incentives. So does kaggle with Google. Civitai is neutral but it is mostly nsfw even though they don't show it outright. Torrenting is the Best way
There are always torrents, which would seem to be a fine way to distribute models.
Big Wolf buying the sheep farm. Nice
(comment deleted)
Well congrats to Clem and the team. I remember when huggingface was doing things like coreference resolution models on spacy.

I hope nvidia does right by the community.

The $13b doesn't go to hugging face's bank acct to pay for anything like S3 egress fees... It go to hugging face's owners' Bank accounts for them to do anything else. New cohorts of billionaires and centi millionaires getting minted.
> billionaires and centi millionaires

I think you mean hecto-millionaires.

I think you mean Dodecahedrillionaires
I think you mean people whose net worth, expressed in millinaires, exceeds 10^11
I think he was telling something called a joke. It's funny because S3 is expensive and they just got some money.
Dont the American investors (Salesforce, GV, IBM, Nvidia) get any of that?
They need to switch to egrees free cloudflare R2!
I wonder if Cloudflare would find a cause to exclude them because I can't imagine the transfer amounts this would generate.
Nothing is free on cloudflare enterprise plans and they will make you move to enterprise once you start using enough services.
I'm using Cloudflare Enterprise and the R2 add-on is provisioned where we only pay for storage, not egress.
Cloudflare R2 has free egress only until Cloudflare’s enterprise sales team sets its sights on your wallet :)
Totally, and as long as they don't turn on the AWS NAT Gateway, they'll be just fine.
Ah yes Spacy, good times :)
nvidia has a reputation for abandoning communities
oh for F's sake. come on.