Lenovo isn’t a brand that we would typically associate with the manufacturing of its own graphics card but lo and behold, a leaked image of the China-based brand’s own GeForce RTX 4090 Legion has made its way online. Yes, this isn’t an off-season April’s Fool joke, that is a full-sized graphics card, made especially for the company’s own Legion-themed gaming PC.
The image, unsurprisingly, comes by way of Twitter from user Алексей (@wxnod). While the name of the card has been pixelated out, the partially exposed end of the Legion-themed component leaves little doubt in our mind that the card that we’re looking at is a custom-made NVIDIA GeForce RTX 4090. That, and the fact that the fonts actually match up with the RTX 40 Series’ new logo design.
Lenovo Saver Suspected of 4090 New Machine https://t.co/Voqp52nIPN pic.twitter.com/kvSL2RtNwj
— Алексей (@wxnod) September 14, 2022
There are a couple of points that we can unpack from this stand-alone image. Firstly, it confirms that the RTX 4090 is going to be huge; the sheer length of the card barely leaves any clearance at the front of the case and on top of that, the massive thickness of the cards can be seen taking up nearly four slots on the motherboard. A point that ties in with earlier speculation about the GPU’s size.
Another feature that is clearly being shown here is the 16-pin microFit MOLEX PCIe power port, or 12HVPWR connector, that is visible on the spine of the card. It’s an important sighting to us, at least, due in part to the fact that earlier leaked images of the RTX 4090 don’t show off the port.
As we’ve mentioned in past reports, NVIDIA is expected to announce the GeForce RTX 4090 on 20 September during its GeForce broadcast. Specs-wise, the card will have 24GB GDDR6X graphics memory, 16384 CUDA cores, as well as an AD102 GPU. It is also expected to have a TDP of 450W, making one of the GPU brand’s most power-hungry graphics cards to date, and while not explicitly mentioned, it is likely to be complemented with some next-generation ray-tracing (RT) and Tensor cores.
(Source: Videocardz)
