EARLY IN THE NEW YEAR,
we’re expecting Nvidia to release the GeForce RTX 4080 Ti. The regular RTX 4080, while awesome, was deemed too expensive, with a launch price of $1,199. It was expected to follow the established pricing pattern for GeForce cards, and significantly undercut the top-dog RTX 4090 (which launched at $1,599). The previous generation’s RTX 3080 started at $699, so even allowing for inflation, it was an unexpected move. It also generated an uncomfortable gap in performance, with the RTX 4080 having 9,728 shaders, while its big brother 4090 had a 16,384 count for not much more money. This made the the 4080 look particularly poor value, and left a huge performance gap in the 4000-series range.
Adding an RTX 4080 Ti gives room to drop the RTX 4080’s price by $100-$200, and even out the pricing structure at the top end. It would also let the 4080 compete with AMD’s Radeon RX 7900 XTX, which is as good at rasterized 1440p gaming, and $150 or more cheaper (see this month’s build on page 16). The RTX 4080 Ti is expected to host a AD102 GPU, with 14,080 shaders, run to around 450W, and sell for $1,1001,200. This would put it closer to a 4090 than the 4080 in performance, but much depends on memory and bus speed. The RTX 4080 was never popular, as people knew they were being sold a dud. You either spent the extra, and went big with the monster RTX 4090, or jumped ship to AMD. Nvidia needs to get this right.
–CL