Nvidia to acquire Hugging Face

303 points by tosh 15 hours ago on hackernews | 91 comments

unddoch | 15 hours ago

Theoretically, can Nvidia direct Hugging Face to sue OpenAI now for the hack? Surely they could find some interesting stuff in discovery :)

hmokiguess | 14 hours ago

I was also theorizing that part of their acquisition was to get more intel on that, though I feel like in practice that whole thing right now is just theatre for inflating the cybersec market

oh_no | 14 hours ago

the last thing nvidia would want is to make ai look more legally risky

to say nothing of their massive investments direct and indirect in openai

jackb4040 | 14 hours ago

Yep. In fact, eliminating that risk on OpenAI's behalf was probably a small factor in the acquisition

Grombobulous | 12 hours ago

The OpenAI web page on the Hugging Face incident is so slick that it basically confirms that the entire incident was a marketing exercise.

It’s obvious that designers and other content specialists spent multiple days crafting this page.

https://openai.com/index/hugging-face-incident-and-the-road-...

pantalaimon | 11 hours ago

There was this summery about it, not sure how you'd need more intel - the technical details of the attack are ephemeral

https://www.dwarkesh.com/p/openai-huggingface

LtWorf | 14 hours ago

Why would they sue one of their most important customers?

deaton | 14 hours ago

Not to mention that they own a 10% stake in OpenAI

aslkalska | 14 hours ago

why would they sue their client ?! what does that say to all their other clients ?

BloodyIron | 13 hours ago

Clients sue other business partners all the time, that doesn't mean they stop working together. There's lots of examples of this.

mcintyre1994 | 13 hours ago

I suspect there are very few examples of it when they are as entangled and invested in them as Nvidia is in OpenAI.

aslkalska | 12 hours ago

no they don't - not when the lawsuit would affect the demand for the thing they're selling

pixl97 | 9 hours ago

There is a particular problem with this view. Yes, you learn how many lawsuits occur per year, but you are clueless on the base rate of actions that could potentially lead to lawsuits.

Is it a 1:10 ration, 1:1000, 1:100000?

You just don't know because companies don't tell you this information. If for example there were a magic oracle that forced every suitable action into a suit, would the entire legal system of the US instantly explode?

acd10j | 12 hours ago

Why would they sue OpenAI, openai is their one of biggest customer.

senshan | 8 hours ago

Or, alternatively, they can now direct Hugging Face not to sue...

cmiles8 | 14 hours ago

Not sure why this is news again this week when it was already announced previously. NYTimes was also reporting this like it was news today.

retinaros | 14 hours ago

Wasn’t official

cmiles8 | 14 hours ago

Sure, but that’s the case for basically all acquisitions.
Unfortunately not all deals go through. Like Wiz + Google, or Figma + Adobe. A lot can go wrong between “in talks to acquire” and “acquired.” The official announcement is less surprising but still newsworthy.

tantalor | 13 hours ago

Um, Google did actually acquire Wiz in March.
I meant the first attempt in 2024, which fell apart: https://www.nytimes.com/2024/07/23/technology/google-wiz-dea...

the__alchemist | 14 hours ago

In related news, Voyager has left the solar system!

Bluestein | 13 hours ago

Further, they launched a second probe. First!

- https://www.youtube.com/watch?v=2WoDQBhJCVQ

Because it's a new blog post with new information

retinaros | 14 hours ago

Life changing for engineers having joined a few year ago. Congrats

arbuge | 14 hours ago

They approached Nvidia to start the conversation. I guess sometimes great companies are sold, not just bought.
Happy for the hugging face team. Many great folks there who’ve waited patiently for HF to find a business model, and finally they get something even better.

abhayaya | 14 hours ago

Hugging Face has an identity problem. It needs a gestalt therapist to take it back to where it was meant to grow up.

dfdydx | 13 hours ago

Where was it meant to grow up?

abhayaya | 3 hours ago

Maybe Wisteria Lane.

iworshipfaangs2 | 14 hours ago

Another investing W for Kevin Durant, who is likely producing better returns than most VCs while being a full-time athlete.

https://finance.yahoo.com/news/kevin-durant-investment-portf...

Grombobulous | 12 hours ago

Let’s not forget to give credit to his family office. High net worth individuals don’t just personally do things like make investments themselves. They hire experts to do that.

Sure, at some point Kevin Durant made the high-level decision that he wanted to invest in high-growth startup companies. But there’s a very low chance that he’s working the day-to-day of that process.

ishtanbul | 12 hours ago

I’d put it close to zero. He was likely presented with the deal once it was sourced and reviewed by his investment team, who likely did 100% of the work. It doesn’t mean he didn’t make the final decision, but pro athletes dont have time to do much investing work beyond high level networking and approving deals.

bdcravens | 10 hours ago

Durant is more than just an athlete who gets suggestions from those who handle his investments; he co-founded a venture firm ten years ago that has invested in over 100 companies.

https://www.hustlefund.vc/post/angel-squad-kevin-durant-inve...

While it's true he may not "do the work" and relies on those he's hired, the same is true of the Mark Cubans of the world.

Going to Golden State was fruitful for quite a few players. Great timing.

Grombobulous | 2 hours ago

This is exactly what I was saying (grandparent comment).

I’m not trying to detract from his success or call him stupid, not at all.

But it’s worth pointing out that the only qualification for starting a successful venture firm is having money.

philipwhiuk | 13 hours ago

I don't really understand why the AI equivalent of Maven Central is worth 12.9 billion dollars but maybe I'm an idiot.

retinaros | 13 hours ago

Its a s3 wrapper. Can be replaced by a wordpress with torrents too. But they built an ecosystem around it and some libs/content. i guess thats what nvidia value

tantalor | 13 hours ago

Network effect.

Same reason why Github (fancy file storage website) was worth $7B+

unethical_ban | 10 hours ago

What data can an be extracted from hf though? With GitHub you get a lot of user behavior analysis and can offer enterprise features and see private repositories.

tantalor | 9 hours ago

The users are the product.

unethical_ban | 5 hours ago

I get that concept, but what part of the user is the product at HF? Even openrouter I get because at least it's a broker of all the user's information. HF to me is a glorified file host with no private user data access, no complex user interaction, etc.

It feels like nvidia has money to burn and just wants to make sure they control the portal to open models in case it could benefit them at an unknown point in the future.

trilogic | 13 hours ago

Huggingface, llama.cpp and all European frontier. Congrats to Huggingface Team, that amount is tempting but nowhere near your real evaluation. You were about to blow up and there is no AI bubble (just not enough infrastructure like energy and hardware to process current and exponential inevitable growing tokens). Huggingface is bringing the real value to AI community, all the rest is more like scalping IMHO. Nvidia is a great company but not opensource friendly by historical doings. Today is a sad day for the community. Put a timer 2 year from now and mark this words: Huggingface is the free foundation, get what you can today.

philkuz | 13 hours ago

Outdated take. Nvidia now open sources whatever will make them sell more chips. Keeping the backbone of Oss model delivery afloat seems like it would sell more chips.

tomalbrc | 11 hours ago

> Nvidia now open sources whatever will make them sell more chips

Oh what did they release?

pixl97 | 9 hours ago

NVIDIA Nemotron is one example. I think they have some others related to robotics.

homarp | 8 hours ago

philkuz | 2 hours ago

* Tensorrt-LLM * Dynamo

barrenko | 12 hours ago

Yes, also feel this is a massive underpricing.

trilogic | 9 hours ago

Yes, for sure.

forestry | 13 hours ago

Dupe.

sealthedeal | 13 hours ago

These Hugging Face guys made out like BANDITS!!!! Good for them. My favorite part is they approached Jensen for the buyout. They knew, very smart and mature on the founders/board.

chews | 13 hours ago

I don’t think many realize how hard it is to raise for a no revenue company that isn’t buying out large GPU contracts and racking machines.

They have a winning product for sure… the worlds models have chosen a hugging face workflow, but honestly where would they be able to find the revenue required to justify another couple rounds at ever higher valuations.

To me, I think these folks see that the music is getting harder to hear and might stop entirely, might as well cash out. I do hope their employees were well taken care of, and 13b of NVDA is likely to be worth a lot more in the coming years.

willy_k | 12 hours ago

> To me, I think these folks see that the music is getting harder to hear and might stop entirely

At the frontier companies, but not for hugging face or nvidia. Their business are going to be around for a while considering how local models are performing now.

slowin | 11 hours ago

What is hugging face's business? I thought they were just free model hosting.

retinaros | 10 hours ago

Selling pro accounts for storage. Its at best a few ten millions arr maybe 100-150

fennecbutt | 9 hours ago

Data. Loooads of data for Nvidia. And an audience.

slowin | 7 hours ago

I don't think they have much data, they host the (open) models but I don't think they're getting any logs from users or anything like that. I don't believe they monetize data either. I also suspect Nvidia has a much larger audience than Hugging Face.

dragandj | 10 hours ago

I don't think local models would require hardware at nearly the scale that the datacenter buildout for OpenAI/Anthropic/etc. required. Nvidia got as high as the Icarus, it will be hard not to melt, despite they being a real company with real product that is so in demand that you can't buy their hardware even when you're willing to pay above the recommended price. They are addicted to billions, and the cold turkey clean-up will be painful as hell.

sealthedeal | 12 hours ago

Agreed. Take a ~$13bn buyout im sure cash and stock is the smartest thing to do for a company their size and age. Take the profits now and keep building securely. I imagine their employees very much gotten taken care of. Jensen mentioned an additional $1bn in cash compensation etc to keep employees at HF.

alexpadula | 13 hours ago

Maybe they approached, tried to work together but then got bought out? Very common.

sealthedeal | 12 hours ago

I watched their CNBC interview this morning with Hugging Face CEO and Jensen. HF CEO specifically said they went to Jensen and said they wanted to join forces. Obviously more elegant lol.

irthomasthomas | 10 hours ago

Overheard: "It would be a shame if Hugging Face had to press charges for hacking against your best customer. Perhaps if I was busy counting my money, it might slip my mind."

himata4113 | 10 hours ago

I doubt that would go anywhere, but this is a very funny joke regardless!

ChrisArchitect | 12 hours ago

$12,930,300,000?

Hugging face emoji = HTML entity 🤗

(https://x.com/calebfahlgren/status/2095516397879152851)

malwrar | 12 hours ago

Is the fact that Nvidia is also massively financing the closed-models companies, which stand to lose when open model project(s) take off, a conflict of interest here? I can’t tell if this is nvidia hedging or planning to embrace & extinguish.

pphysch | 12 hours ago

Nvidia is among the best positioned if open models win.

gscott | 10 hours ago

Most of nvidia's revenue comes from data center scale out. They will loose their ass if local models win.

c0rruptbytes | 10 hours ago

open models don’t imply local models… it just allows more people to run them, most likely with nvidia GPUs

pixl97 | 9 hours ago

Any model ran takes compute. Everybody, datacenters and local need compute.

And I don't see a future where datacenters lose much to open models. I've ran local server rooms before clouds were a thing. Cooling and power is a very expensive pain in the ass. You're not dropping a rack of local models in most buildings without breaking the local power grid.

malwrar | an hour ago

What about when the open models get to the point that you don’t need these fancy GPUs? It seems inevitable this will happen over time, we see cracks right now within even just GPT w/ expert streaming and flash attention. ASICs are coming soon too.

ijidak | 9 hours ago

AWS and Azure would love to cut out OpenAI and Anthropic and sell inference on open models directly to businesses with no cut to share, just as AWS serves other open source software.

Open models are great for everyone except anthropic and OpenAI.

notatoad | 11 hours ago

no, it's not a conflict of interest to invest in multiple companies who compete with each other.

liquicity | 11 hours ago

Openrouter and huggingface both sold to multi billion dollar entities in the same few weeks - perhaps this is great for the founders, but as a consumer I hate this acquisition loop that happens with every company of adequate size these days.
Or maybe they know that this AI bubble will not last.

oefrha | 11 hours ago

Feels like acquiring Docker Hub (just the hub, not even Docker Inc that controls Docker the software) in ~2018. AI craze valuations are truly out there.

deweywsu | 11 hours ago

There goes the neighborhood.
> Together, we will make AI more open, more capable and more accessible to people and institutions around the world.

Agreed. Then maybe Nvidia should open source nvcc and CUDA the next software that should be open sourced.

Is it feasible for someone else to just start something almost identical by hosting a bunch of local models?

doublerabbit | 10 hours ago

Not without lots of money or decent network peering. A 30GB model downloaded a lot of times is a lot of bandwidth.

There's more to than just the website, network bandwidth costs. Security, storage and infrastructure costs to name a few.

You can't just spin up a VPS and host a brain jar shop.

Capricorn2481 | 9 hours ago

> There's more to than just the website, network bandwidth costs. Security...

Well they may have saved some money on one of those

corysama | 10 hours ago

My understanding of Hugging Face is limited to "File Host with Model Cards". Can someone with more understanding explain what the $12 billion value comes from?

eitally | 10 hours ago

They offer inference, too. I don't know anyone who uses this service, but presumably it must be reasonably popular.

fennecbutt | 9 hours ago

They are the primary service for this. That comes with an audience and a lot of data.

hmartin | 8 hours ago

This is the most compelling case I've come across:

https://www.decodingdiscontinuity.com/p/nvidia-hugging-face-...

tldr; Acquisition is an option on Hugging Face becoming the distribution and data-creation infrastructure for open physical AI.

SillyUsername | 10 hours ago

Somebody better prep archive.org to buy a few more petabytes of storage.

Danox | 10 hours ago

The goal of the deal is to speed up open source according to Nvidia what a laugh ha ha ha …

Where is the EU in all this? Is this not similar to letting another European tech company disappear from Europe ie.. Arms Holding which was sold to the Japanese, how do you compete going into the future in the tech world if you let European companies be bought out by foreign companies. Declaring them gatekeepers down the road and issuing a fine will never make you a tech powerhouse.

Is this just another symptom of the EU being run by politicians and lawyers who don’t seem to grasp/understand tech.

porridgeraisin | 9 hours ago

HF is an american company.

bionhoward | 9 hours ago

I personally think this acquisition is genius anybody who could afford it would be wise to own HuggingFace because they are incredible

Psillisp | 8 hours ago

Not A Bubble

Not A Bubble

Not A Bubble

Not A Bubble

Not A Bubble

Not A Bubble