AMD vs NVIDIA: China's Ecosystem Could Change Everything #shorts

4 May 2026 · 1 min

Watch on YouTube

1:06

In this short commentary, the speaker argues that AMD’s current lag behind NVIDIA is not due to a unique technical advantage held by NVIDIA, but rather because NVIDIA already benefits from a mature ecosystem of libraries, tools, and developer adoption. This ecosystem creates a lock‑in effect: developers continue using NVIDIA’s platform to avoid the substantial effort required to rewrite code for alternative hardware. The speaker speculates that Chinese AI labs, which are developing large language models (LLMs), may migrate to Huawei’s chipset ecosystem, potentially eroding NVIDIA’s dominance. If Chinese labs adopt Huawei, the ecosystem advantage would shift. The speaker also notes that DeepSync has recently released a 1.5‑trillion‑parameter model that rivals the latest offerings from Anthropic and OpenAI, and suggests that DeepSync would prefer a Huawei chipset over having no suitable hardware, despite the extra work needed to develop compatible libraries.

Chapters

The reason AMD hasn't been as big or as successful at the moment as with NVIDIA is because they don't have the ecosystem yet. But if China, all the LLM model and all the labs move onto the Huawei ecosystem, which they could very well do, and maybe they're already doing it, then the ecosystem mode is not really there. So it's not like NVIDIA has this top secret invention that nobody else can copy, right? It's more about they already have an ecosystem, right? Just like, because other people are using it, so I'm going to use it. It's just too much trouble to, to rewrite all the library and, and, and rewrite all the other codes. But if, if, if Deep6 are against the wall and, and have to choose between having no chipset and a Huawei chipset that will require additional work in terms of writing a library DeepSync will definitely choose the Huawei. And in fact, the recent release of the DeepSync model is a 1.5 trillion parameters model that is very close to the latest model from, from Anthropic and OpenAI, right?