Sounds like there’s a market for a hugging face comeptitor.
Nvidia is seeing the writing on the wall that local models are the only viable future for AI. They have to try and squash that now before it’s too late. I mean, it is already is too late, as they’ve hooked there wagons to these AI companies, and that shit’s going to end and nividia is going to be a bag holder. I can’t wait.
Nvidia is seeing the writing on the wall that local models are the only viable future for AI
Not necessarily. HuggingFace already lets you run models in their servers. Nvidea might be buying HF so that they can sell cloud-compute to even more people.
Some do, others just charge for everything or charge for something that’s released for free elsewhere… Cults3d and Yeggi are bad for that from recent memory.
It feels like one of those technologies that can’t truly be contained. It’s like how the government keeps trying to end piracy. It’s like a hydration hydra. Cut off one head and two more show up. Local models will be the dagger to premiere models.
Unfortunately, AMD’s presence in that market is only enough for Nvidia to say ‘see, we’re not a monopoly.’
AMD has repeatedly tried to fund CUDA translation, on the down-low, and then instantly walked away whenever that came to light. They know Nvidia would pull a Sony and pile on losing court cases until the justified party is broke.
Nvidia must be shattered over CUDA. We cannot allow computing to remain proprietary, for your own code.
Maybe, but CPUs are coming out with LLM tuned chips, and I can run a basic model on an i5, 8gb ram, and no dedicated card. It’s not super powerful, but for most users, it’s more than enough for what they use the big models for. Also, if it does take a discrete card to get that needed boost in performance, then at least consumers would be able to get GPUs again.
I think as hardware improve and is further designed around LLM efficiency, and local models are tuned for specific uses and being able to run on lesser hardware, it will make Nvidia obsolete for large swaths of the population. A good GPU will still be necessary for high performance, graphic/physics intense gaming, but that’s a really small subset of all users.
Hopefully Nvidia just shits and has to grovel back to the consumer to get there marketshare back when all of the DCs go tits up.
I can run models on my 3yo midrange smartphone. Gemma-4-E2B totally runs on there. Bonsai-8B too. But neither is really good for most tasks.
Heck, “a basic model” even runs locally on the 8MB RAM of an ESP32-S3, but it’s utterly worthless at anything.
If you get a bit more into self-hosting AI it quickly becomes obvious that for any actually useful real world tasks you need at least 24GB you can dedicate to the LLM alone, and if this is fully GPU vRAM, the performance is way, way higher than on CPU, even with an NPU.
any actually useful real world tasks you need at least 24GB you can dedicate to the LLM alone, and if this is fully GPU vRAM
Agreed. For anyone who doesn’t have a 24GB card, the AMD 7900 XTX is basically the only sanely priced 24GB GPU left in 2026 and can be had for under a grand new, whereas even a used 3090 is $1200 and $2000 new.
Qwen 3.8 27B runs nicely on my XTX. I have to use Claude at work, and Qwen 3.8 in llama.cpp with the OpenCode desktop app leaves nothing to be desired in comparison. 3.8
27B is about on par with Sonnet 5 according to benchmarks, though I even prefer it over Opus because Qwen writes in clear, understandable language. We are at the point now where anyone with a 24GB GPU has little reason to use data center LLMs.
Sounds like there’s a market for a hugging face comeptitor.
Nvidia is seeing the writing on the wall that local models are the only viable future for AI. They have to try and squash that now before it’s too late. I mean, it is already is too late, as they’ve hooked there wagons to these AI companies, and that shit’s going to end and nividia is going to be a bag holder. I can’t wait.
Nvidia releases open models too
Not necessarily. HuggingFace already lets you run models in their servers. Nvidea might be buying HF so that they can sell cloud-compute to even more people.
or to lock the models behind a paywall… look what’s happened with 3d printing, so many sites now charge for STL files.
From what I’ve seen, the sitse that charge for STLs are priced based on the sculptor, unless you mean something else
Some do, others just charge for everything or charge for something that’s released for free elsewhere… Cults3d and Yeggi are bad for that from recent memory.
We need to get models into a distributed distribution. Huggingface is way too central.
Yeah, would love to see more of a torrent type website that simply hosts torrent files to models.
some sort of arr stack for models tbh
It feels like one of those technologies that can’t truly be contained. It’s like how the government keeps trying to end piracy. It’s like a
hydrationhydra. Cut off one head and two more show up. Local models will be the dagger to premiere models.drink water!
Damn hydration. Always growing too many heads.
Hail hydration
yeah fully agree. They may go more underground, thinking if govs try to block them, but they’ll still be there.
There will be a new generation that learns about VPNs and torrents. Life is cyclical.
NVidia wins if people run local models on their hardware too. Hugging face is not a competitor, it is an enabler.
What about graphics card not made by Nvidia? I run local models with AMD chipset and GPU. Works great.
I doubt I’m the only one.
Unfortunately, AMD’s presence in that market is only enough for Nvidia to say ‘see, we’re not a monopoly.’
AMD has repeatedly tried to fund CUDA translation, on the down-low, and then instantly walked away whenever that came to light. They know Nvidia would pull a Sony and pile on losing court cases until the justified party is broke.
Nvidia must be shattered over CUDA. We cannot allow computing to remain proprietary, for your own code.
Maybe, but CPUs are coming out with LLM tuned chips, and I can run a basic model on an i5, 8gb ram, and no dedicated card. It’s not super powerful, but for most users, it’s more than enough for what they use the big models for. Also, if it does take a discrete card to get that needed boost in performance, then at least consumers would be able to get GPUs again.
I think as hardware improve and is further designed around LLM efficiency, and local models are tuned for specific uses and being able to run on lesser hardware, it will make Nvidia obsolete for large swaths of the population. A good GPU will still be necessary for high performance, graphic/physics intense gaming, but that’s a really small subset of all users.
Hopefully Nvidia just shits and has to grovel back to the consumer to get there marketshare back when all of the DCs go tits up.
Bitcoin went through a similar trajectory. Graphic cards to specualty chips.
I can run models on my 3yo midrange smartphone. Gemma-4-E2B totally runs on there. Bonsai-8B too. But neither is really good for most tasks.
Heck, “a basic model” even runs locally on the 8MB RAM of an ESP32-S3, but it’s utterly worthless at anything.
If you get a bit more into self-hosting AI it quickly becomes obvious that for any actually useful real world tasks you need at least 24GB you can dedicate to the LLM alone, and if this is fully GPU vRAM, the performance is way, way higher than on CPU, even with an NPU.
Agreed. For anyone who doesn’t have a 24GB card, the AMD 7900 XTX is basically the only sanely priced 24GB GPU left in 2026 and can be had for under a grand new, whereas even a used 3090 is $1200 and $2000 new.
Qwen 3.8 27B runs nicely on my XTX. I have to use Claude at work, and Qwen 3.8 in llama.cpp with the OpenCode desktop app leaves nothing to be desired in comparison. 3.8 27B is about on par with Sonnet 5 according to benchmarks, though I even prefer it over Opus because Qwen writes in clear, understandable language. We are at the point now where anyone with a 24GB GPU has little reason to use data center LLMs.
except the chips are bought out by AI companies years ahead of time. it will take a long time for nvidia to switch back.
Doesn’t matter to Nvidia. It’s just about having a plan B if all those future orders start to collapse.