I’ve always believed that we desperately needed competition in the GPU space, which is why I was so happy when Intel entered the GPU market – even though it only makes mid-range and budget cards.
New competitors from China are forming, but should NVIDIA be afraid?
The history of Chinese GPUs
While this may be the first we’ve heard of Lisuan and its intentions to release a consumer-grade GPU, it’s come a long way. For decades, China’s graphics processing field relied entirely on Western imports, with domestic efforts largely limited to academic research or strictly military applications. But it turns out that the initial shift to self-reliance actually started in 2006 with Jingjiawei, which successfully produced the JM series of GPUs. These are specialized chips intended for aerospace and military use and lack the DirectX support or driver maturity required by the consumer market, so they are not really suitable for gaming or general consumer computers. But it’s something.
The real turning point came around 2018, accelerated by the escalation of U.S. trade disputes and the global artificial intelligence boom. This period saw the rise of the “Four Tigers,” startups such as Biren Technology and Moore Threads founded by returning veterans of NVIDIA and AMD. Moore Threads first launched the MTT S80 to consumers in 2022.
Moore thread
While it marked a historic milestone as the first domestic gaming card to feature PCIe 5.0, it struggled with immature drivers and unstable performance, and was unable to run modern games smoothly.
That’s where Lisuan, a startup founded in late 2021 by a team of former Silicon Valley engineers, comes into play. Learning from the driver issues of its predecessors, Lisuan focused on software optimization based on the “TrueGPU” architecture, and finally launched the Lisuan G100, which may be China’s most solid attempt at actually usable GPUs. They just started shipping today, so it’s not vaporware, but an actual product that exists.
But is there any benefit?
Lucas Gouveia / Geek Guide
This particular GPU is…a bit of a mixed bag. It should be built using the 6nm process. According to Lisuan, the G100 is the first domestic chip that is truly comparable to the NVIDIA RTX 4060 in raw performance, providing 24 TFLOPS of FP32 computing. It even introduces support for Windows on ARM, a feature that even its major Western competitors haven’t fully prioritized.
However, it appears to have fallen short of its marketing promises. A purported Geekbench OpenCL listing shows the G100 scoring just 15,524, a performance rating that effectively ties it to the GeForce GTX 660 Ti released in 2012. This puts the “next generation” Chinese GPU on par with 13-year-old hardware, giving it one of the lowest-scoring entries in the modern database. Leaked specs further muddy the waters, showing the device uses just 32 compute units, a confusingly low 300 MHz clock speed, and an almost unusable 256 MB of video memory. We’ll likely see more benchmarks as the GPUs get into customers’ hands.
These “meager” numbers may represent engineering samples that were unable to report correctly due to driver immaturity – a Ryzen 7 8700G testbed configuration on Windows 10 supports this theory. But if true, the underlying chip may still simply not hit the promised RTX 4060 performance targets, regardless of actual reported specs.
Does it have a future?
However, here’s the thing. Even though this card performs similarly to the GeForce GTX 660 Ti, it’s still pretty good. Creating a functional GPU architecture from scratch (especially one that relies on a domestic 6nm supply chain rather than TSMC) is a huge engineering triumph. In fact, the G100’s ability to boot and run Windows and execute modern APIs such as DirectX 12 is a “zero-to-one” breakthrough. The Lisuan G100 has a host of issues that can be fixed through software, and even if the hardware and chips are the culprits, Lisuan can definitely put something together in a few years that can rival NVIDIA and AMD in at least some areas.
On the bright side, thanks to the Chinese government’s innovative plans, the country is actively replacing tens of millions of foreign computers with domestic alternatives. Lishu already has a large number of real-world users and a guaranteed captive market, which will be the key to increasing its research and development efforts. This would give Lisuan access to the capital and user data it needs to iterate quickly, insulating it from the cutthroat open market competition that has killed Western startups like 3dfx.
When will we see a decent mid-range or even flagship chip from Lishu? There is no way to explain clearly. It could be two years, it could be five years, it could be ten years. But if the company plays its cards right, it might soon be catching up?
Can we actually check them out in the US? Never say never, but considering how little Chinese electronics we actually get, it’s a gamble you’ll probably lose.
