Huawei Begins Delivering CloudMatrix 384 AI Clusters to Chinese Clients; Costs Triple That of NVIDIA’s GB200 NVL72 System

Okay, so here’s the deal with Huawei’s new gizmo – CloudMatrix 384 AI cluster, or whatever they’re calling it. Seems like the big tech kids in China are flocking to it like bees to honey. I guess they’ve had enough of NVIDIA’s stuff, or maybe they just want something homegrown. Who knows? Could be a bit of both.

So, picture this: Huawei, not exactly shy are they? They’re pushing their way into AI like they own the place. Their CloudMatrix 384 is supposedly built from scratch – by them, with their own goodies. If you believe the rumors, it’s giving NVIDIA a run for their money, at least in China. It’s like a local superhero beating up on the foreign villain. Or something like that, anyway.

Anyway, back to those ten mysterious companies who snapped this up. Nobody’s saying who they are yet – top secret, of course. But word on the street is they’re big names in the tech scene. I don’t have a crystal ball, so your guess is as good as mine on who they are.

Here’s a quick recap: CloudMatrix 384 is strutting around like it’s the boss, taking on NVIDIA’s powerhouse GB200 NVL72. It’s like they’re saying, “Hey world, we don’t need you to hold our computers.” Which, if you think about it, is kind of a big deal.

And there’s this nerdy spec stuff. The CloudMatrix 384 has these 384 Ascend 910C chips – all tangled up in an “all-to-all topology.” Whatever that means. I’d explain it, but I’m pretty sure only rocket scientists get it. Twice as powerful in some areas, but it’s guzzling energy like nobody’s business. Not great if you’re an environmentalist, but who’s perfect, right?

Now brace yourself, the price tag! We’re talking $8 million for the CloudMatrix 384. It’s like buying a dozen sports cars or something. Bigger price but more local pride, maybe? Seems Huawei isn’t about the cheap thrill – they want to show they’re not playing around, even if it means burning a hole in your pocket.

So yeah, that’s the scoop. Less about being cost-savvy, more about waving their own flag and telling NVIDIA, “We got this.”