THE NVIDIA AI GPU BLACK MARKET | Investigating Smuggling, Corruption, & Governments
THE NVIDIA AI GPU BLACK MARKET | Investigating Smuggling, Corruption, & Governments


cross-posted from: https://piefed.ca/post/152656
THE NVIDIA AI GPU BLACK MARKET | Investigating Smuggling, Corruption, & Governments
cross-posted from: https://piefed.ca/post/152656
Sweet Tech Jesus! It's 3 and 1/2 hours long.
It’s a really good watch. I’ve put it on while playing some Factorio, so time went by quite quickly
that's an amazing idea, i'll do that.
any other factorio-playing videos for my second monitor? :p
yeah even I'm not that autistic and I watch videos of people soldering
I dunno, if it were Hbomberguy talking about YouTube plagiarism for four hours I’d watch it.
I usually put it on when I'm playing Alan Wake 2 so I wouldn't shit my pants too hard.
Pace has a really nice series on YouTube ;-)
I really respect Gamers Nexus for actually doing the shit Linus Tech Tips just haws and haws over.
I don't think LTT ever did or claimed to do investigative journalism (or really journalism at all)?
They make entertainment videos that sometimes have mildly informative value.
I just got sick of there always being a new "controversy" with him and his inability to seem like he gives a shit about it. Each time he gets accused of something, whether he did it or not, he just goes on the WAN show to sigh as big as possible and say "Guys..."
I don’t think LTT ever did or claimed to do investigative journalism (or really journalism at all)?
But thats how it's presented, and that's how many viewers interpet it.
First it's not investigative journalism, it's hardware reviews that may or may not reveal problems, and he fakes and cheat with his, and takes money illegally under the table for making good reviews.
Difference between him and for instance Gamers Nexus, is that Gamers Nexus reveals if he detects something is off when he tests. While LTT prefer to get companies to pay for good reviews that gloss over the problems.
People are idiots if they believe your claim is true, and apparently most do?!
LTT doesnt make deep investigative hours-long videos anyway
Oh I almost forgot about him, I blocked his channels years ago. He is an idiot and a shill, and did illegal hidden advertising.
Why anyone would still watch his channels is very strange IMO.
Illegal hidden advertising?
It’s kind of a guilty pleasure for me. I like it in the background sometimes. It’s like watching code blue cam. Cops are occasionally useful bullies but I don’t watch it for the cops. I watch it for the white trash judge judy bullshit. It’s a methed out freak show sometimes. It’s spectacle.
Thanks Steve.
Papa's here
awkward pause See what I did ther... video cut
But now that Nvidia agreed to pay a bribe to the government it's no longer black market
Good ol' orange market.
Orange is the new black.
Funny thing is 'Local ML' tinkerers largely can't afford GPUs in the US either.
The 5090 is ludicrously expensive for its VRAM pool. So is the 4090, which is all but OOS. Nvidia will only sell you a decent-sized pool for $10K. Hence non-techbros here have either been building used RTX 3090 boxes (the last affordable compute GPU Nvidia ever sold), EPYC homelabs for CPU offloading, or have been trying to buy those modded 48GB 4090s back.
The insane supply chain is something like this:
All because Nvidia is playing VRAM cartel and AMD, inexplicably, is uninterested in competing with it when they could sell 48GB 7900s basically for free.
You could also buy the Apple Studio with its large amount of unified ram for a similar price of a 5090. Of course it’s not as fast but it could run a model that needs more ram.
The pricing for memory is still pretty bad. $4K for 96GB, $5.6K for 256GB, $10K for 512GB. One can get 128GB on the M4 Max for $3.5K, at the cost of a narrower bus so it's even slower, but generally, EPYC + a 3090 or 4090 makes a lot more sense.
SOTA quantization for these are mostly DIY. There aren't many MLX DWQs or trellis-quantized GGUFs floating around.
But if you want to finetune or tinker instead of just run, you're at an enormous disadvantage there. AMD's Strix Halo boards are way more compatible, but not standalone yet and kinda rare at this point.
barebones 780m amd mini pc. 128gb ddr5600. About $700. Under $800. Vulkan is good, non AMD supported version of ROCm works for some, because AMD says fuck you to its users. $2000 higher end mini pc with 8600s gpu is other option for high ram with under double the performance, but closer for cheaper than apple solution.
I live in Taiwan and the GPU is more expensive than America.
The issue is that nVidia are increasingly marketing their consumer grade GPUs to "prosumer" users. Whether that is small research groups working with "AI" or people farming the latest memecoin or... the other things you would need REALLY REALLY high bandwidth linear algebra from and let's move on.
Whereas AMD are actually still targeting that consumer market. I think it was the nvidia 40x generation where their consumer cards had like no memory at all and AMD were pumping out 16 GB on their cheap(-ish) models? My brain can only remember card generations while I am actively shopping and... yeah.
And yeah. I would LOVE an AMD card with 32 or even 64 GB of even slower memory. But games are still going to target nvidia because people keep buying it and that means that you just won't have much use beyond the 8 (or apparently now 16) GB that nVidia are going to let you buy. At which point... why waste money?
As for the prosumer and enterprise space? nVidia... have a long history of being assholes and previous GN videos have talked about the behind the scenes pressure they allegedly apply to system integrators and the like. And I will leave that there for Reasons.
But yes, many mid-tier and even high-tier companies could benefit from just buying AMD cards and there is very much a market for "high end" AMD cards... it is just that they have so few customers to make it worthwhile.
I dunno what your talking, but all AMD has to is this:
...That's it.
They'd make seperate SKUs with double the VRAM. AMD doesn't have to waste a cent.
It's not explicable why AMD is not breaking rank on VRAM and vGPU, same reason as failing Intel.
But the reason is not mentionnable in polite company and it relates to why AMD exists at all.
By all accounts, AMD should have gone under decades ago, instead they're one of the only x86 platform licensee and they got that basically because of a fluke in history.
But here's the real deal, because of the regulatory environment, monopolies are technically illegal. Of course since 1980s enforcement of that has been a total joke as proven with the failure of the Microsoft anti-trust case. Anti-trust is currently neutered and even back then it wasn't really "anti-trust", merely anti-monopoly. That is the playbook for Intel and Nvidia that allows AMD to continue existing. They exist so that the other two aren't monopolies. They are kept alive as long, some market segments will have some competition and other will simply not be touched by Intel and AMD, dynamically decided, kind of like splitting territory, so that AMD can always survive.
So that Intel and Nvidia don't became actually illegal monopolies. This is a very conservative playbook as since the 90s they could have very well become monopolies and the neolibs wouldn't have squeaked. They would have cheered !
That's why you're not getting that actually threatening to Nvidia amounts of VRAM and vGPU from AMD and why they're dropping support for their 2018 datacenter GPU if they get a little too much of a good deal on the used market.
The solution is simple, destroy nvidia, not a calculated and gentle trust bust, no, break the company so it stops existing as a coherent entity.
Nvidia is a bunch of software and PNGs on top of TSMC. Break Nvidia, break Intel, break microsoft, break cisco, broadcom, break everything in silicon valley and put it all in a blender. And if anything grows too big again, break it the duck up again.
the playbook for Intel and Nvidia that allows AMD to continue existing.
This isn't a good explanation for why AMD would act stupid. But there still needs to be one.
I feel dumb for it but I just can't keep track of all the controversies and scandals right now:
Which scandal were they referring to the week of August 11 for why nobody in the US state department would respond?
We really need a spreadsheet to track all this shit.
Whaaaat. US protectionism is purely performative and ineffective? I’m shocked! Surely this administration couldn’t be incompetent?
In the video alot of people said that it is slowing them down
My god, this is the LOTR Extended Edition of tech journalism videos
That was a lot of CAPITAL LETTERS
Mr Burke is going to get black bagged.
Any TL;DR on this? Nvidia secretly ships sanctioned GPUs to China?
GPUs that are banned from export to China can be easily acquired in low numbers in the country due to individuals bringing in banned GPUs which get sold just like unbanned GPUs after passing through several people. Also possible that some banned cards are repaired QC failures from the Chinese factories that produce these cards (the cards are produced in China, but are supposed to be exported and banned from export back into China). The export restrictions present more of a barrier for building out large AI clusters with many, many GPUs.
Nvidia almost certainly knows about this but turns a blind eye because they want to sell more cards and China is a large market.
There are some repair shops in China that can not only repair the GPUs, but can also transfer the GPU die onto an aftermarket PCB and give it double the amount of VRAM to create a card that Nvidia doesn't even sell.
Individual banned GPUs are easier to get than the full systems that Nvidia also sells. AMD GPUs are much less popular, Intel GPUs are basically non-existent. The overall most popular card for smuggling into China seems to be the RTX 4090.
Thanks!
Sad to see all this effort has to be put into making sanctions and then into evading them.
I can't wait for this fucking bubble to burst.
Much easier black market than drugs, because it is legal to import it, and no one checks when you export it.
Nvidia has 18%-28% of its sales to Singapore. They defend this by saying only 2% of their product is shipped there, but they don't say which countries receive the bulk of shipments that are billed through Singapore.
some 5090 cards in Hong Kong (from OP) are the same price as the US, which makes it hard to understand why US people have been charged with smuggling, and shows complicity of entire supply chain to funnel to China. Premiums over MSRPs in US would be zero or lower if this was all above board. You also can't find a 5090 on Amazon, but you can in Hong Kong boutique?
Prohibition for the loss, as always.
Aaand they got their video taken down, link to post on YT
Here's a mirror on internet archive.
Thanks! +1
I've added the link to the post🙌
Why did you link an image instead of the video?
Good question and thanks for asking!
As far as I'm currently aware, due to YouTube constantly changing things, YT thumbnails often will not load unless directly linked to the video's thumbnail (at least here on the Fediverse).
Because of this, it allows me to link to alternative sources within the post such as Invidious (YT alt ad-free frontend), Peertube, Bandcamp, etc.