Okay, so this Chinese RTX 4080M thing, it’s a mess, right? A desktop card, but it uses a mobile chip, the AD103 GPU. And it’s capped at 100 watts, like, severely capped. That’s the whole point, the mobile chip, it’s not designed for desktop power envelopes. The card, it’s apparently selling for around $300 in China.

That’s a low price point for something with “4080” in the name, even with the “M” for mobile. But you get what you pay for, right? It’s not a real desktop RTX 4080. It’s not even close to a desktop RTX 4070 Ti Super, or anything like that. Performance numbers are coming out, and they’re not great.

The card, this custom Chinese RTX 4080M, it reportedly loses to the AMD RX 9700 GRE by a significant margin. Like, 35% slower. That’s a big gap. The RX 9700 GRE, that’s an interesting card too, not widely available everywhere, but it’s a solid performer for its segment.

So a 4080M, even a desktop version, getting beaten that badly, it just shows the limitations of that 100W power limit. The AD103 chip, it’s capable of more, but not when it’s choked like that. The mobile RTX 4080 in a laptop can draw up to 175W, so this desktop variant is running at significantly reduced power. This whole situation, it highlights the gray market, the custom solutions happening in China. They’re taking mobile silicon and repurposing it for desktop use, probably to meet specific market demands or to use up excess mobile chip inventory.

It’s not an official NVIDIA product for the desktop market, obviously. It’s a third-party creation. And NVIDIA, they’re pushing their own stuff, their official roadmap. We just had that whole thing with the Kyber AI rack. There was a report, you know, floating around, saying it was delayed until 2028.

But NVIDIA, they came out and refuted that today, July 6, 2026. They said, “Our roadmap is intact.” So no delays there, at least according to them. That’s big for their AI strategy, obviously, and for the stock. Speaking of stock, I picked up some NVDA back on January 2, 2024, at like, $481.68 a share. It was a dip, you know, just a small position.

I’m holding that until it hits, I don’t know, $1500, or if there’s a major market correction, then I’ll re-evaluate. But the AI story, it’s still strong. This Chinese 4080M, it’s not going to impact the high-end market, or even the mid-range in most regions. It’s a niche product, for a specific market, probably for budget-conscious gamers who see “4080” and think they’re getting a deal. But the performance just isn’t there.

It’s a mobile chip, in a desktop form factor, with a tight power budget. That’s the takeaway. The core count is there, the memory interface is there, but the clock speeds and sustained boost clocks are just not happening at 100W.It’s a testament to how much power matters for modern GPUs. You can have a powerful chip, but if you can’t feed it enough power, it just can’t stretch its legs.

The reviewer tests showed it performs slightly slower than its competitor from team Red, the RX 9700 GRE, which is a card that itself is designed for a specific performance tier. The 100W cap is the bottleneck. It’s the primary reason for the performance deficit. So, you have a card, it’s $300, it’s called an RTX 4080M, but it’s not a desktop 4080. It’s a mobile chip, power-limited, and it gets beaten by a mid-range AMD card.

Is it a good deal for $300? I mean, maybe for some specific use cases where power consumption is paramount, or if you just need a cheap card with a lot of VRAM, but for gaming performance, it’s a compromise. A big compromise. The desktop RTX 4080, the real one, has a TGP of 320W, for comparison.

This 100W limit on the mobile chip in a desktop card, it’s just too restrictive. It’s like putting a sports car engine in a golf cart and then limiting the fuel flow. What’s the point?