USA companies optiong for chinese models

Jul 14, 2026

View on YouTube

Chapters

00:00 The Unfolding AI Revolution 00:45 Western Firms Switch to Chinese AI 01:30 AI's Pivotal Historical Moment 02:15 The Internet vs. AI Impact 03:00 Geopolitical Stakes of AI Access 03:45 The Story: Western Companies' AI Shift 04:30 How China Catches Up: Distillation Technique 05:30 Notable Companies Making the Switch 07:00 Why This Shift is Happening: Price & Performance 08:00 Strategic Workload Offloading 08:45 Fragmentation of the AI Ecosystem 09:30 The Bottom Line: Team West's Challenge 10:15 Geopolitical Implications and Future Advocacy

Transcript

Western companies are quietly switching to Chinese AI models. Now there's a reason why I keep coming back to AI and all the evolution of AI, because one is happening so fast, it's bloody interesting. And I'm involved in it, so I'm like, oh, shit, this is great. But I think it is quite, it is a pivot in history. We are going through, we don't really know it, but we're going through a bit of a pivot in history when it comes to this technology. It's kind of bigger in my humble opinion than the internet. And you know what happened with, when the internet came through, it was a pretty big, it was like the industrial revolution, right? It was like when we were, we pivoted to engines. You know, when engines hit the population of the world and our electricity, you know, these major moments in our history, they were pretty big. Now I think this one is right out there. Like, you just haven't seen what is coming. And I mean, I get to see a little bit of it and I'm like, wow, this is so crazy. Absolutely crazy. So I think our lives are all about to be tipped upside down. And yeah, it's going to be pretty interesting. So that's why I keep coming back to this subject. It's also going to impact geopolitics in the biggest way, very strongly, team, very strongly. Because if you don't have access to these tools in this capacity, then you're in trouble. You're really in trouble because you don't have the ability to compete. So that's why I keep coming back to this. But what is the story? This is a quick one. This is all about a notable list of Western companies that are now moving significant AI workloads to Chinese developed models. This shift is driven primarily by superior price performance ratios, rapidly improving capabilities from the Chinese frontier labs. They essentially, they use this technique called distilling. They come in, they crank up 20 or 30,000 subscriptions to anthropic and then they use a technique to ask a whole bunch of questions and study the way that it answers the question, the response. And then it uses that to derive from that interchange how the model is actually architected underneath. And then they basically take that and they apply those learnings to their own models and that's how they catch up. So it's not difficult to do. It's actually really easy to do. It's paid 30,000 times 20 bucks a month. You run that for three months, bang, you've got a model. So it's a very, very clever way of doing it. Of course, all the American companies now are like, they're entrepreneurs and they're like putting in anti-distillation techniques to try to detect this scarolous, these bloody Chinese they keep on doing clever stuff like this. Now, let's go back to the notable companies making this switch. Lindy, I don't know who all these are, but Lindy has adopted deep seek V4. Kerser is now using Kimi K2, 2.5, Coinbase. Wow, Coinbase has integrated GLM5-2 and Kimi-2-7. Shopify is now using Queen. What? How is that even legal? Somebody's got to do something. Come on, small hands, you've got to do something. Uber eats as you're running on Queen-2. Siemens, Siemens isn't that a German or I don't know, some Nordic companies, but it's like a German company, isn't Germany? Yes, Siemens is using a combination of deep seek and Queen. We need to stop this Germany. Chaps version is using Queen. Microsoft is testing. They're not using, they're currently testing deep seek V4. Why this is happening? So Chinese models like all these are at 15 times cheaper. Basically that's why. Core-centred applications like coding systems, customer support, internal tools, interference everywhere, all that sort of stuff. It just makes sense to offload that. If you need really difficult stuff done, give it to the premium models, the frontier models. But if you don't, then just dump that workload out onto the cheapest commodity token that you can find that does the job reasonably well. Many of these companies are not fully replacing their Western models but are strategically shifting parts of their workload to Chinese alternatives where the performance gap is small or non-existent. This reflects a broader market reality where the quality difference is like 5%, but you're paying 15x on, it's impossible to justify that. So, strategically and geopolitically, well, this drives deep fragmentation of fragmentation. It's just like fragmentation. We need deep fragmentation. But fragmentation, chopping up of the global AIA ecosystem, Western companies are effectively voting with their wallets for cheaper Chinese models. That's the issue with this one. So, bottom line, major companies doing the switch, I don't like it for team west, come on team west, don't like it because basically this means that we're missing out on this opportunity to do this sort of stuff exclusively. And we're letting the Chinese come in and basically do it. And the other issue is it drops the price per token because if you can get a token for like, you know, we'll call it a dollar, a million tokens for a dollar from China, but it costs you $15 from the Western models, you're always going to buy from China. That's the problem. So, I think we need to, you know, form a cartel and legislate to make sure that in the west, you can only use the cartel. I'm joking, I'm joking team, I don't want that. I want about to use whatever I like, but you know what, that's what these companies will be advocating for. I'll be lobbying for exactly what I just said. So, I'll bring it to you, if they succeed at that, I'll bring it to you and you'll know. So, that's why this is such an important geopolitical subject. I hope you've enjoyed this little show today. Wherever you are, this amazing world of ours. I hope you have a great day, cheers.