Nvidia delays the controversial RTX 4080 12GB

“The RTX 4080 12GB is a fantastic graphics card, but it’s not named right. Having two GPUs with the 4080 designation is confusing. So, we’re pressing the ‘unlaunch’ button on the 4080 12GB. The RTX 4080 16GB is amazing and on track to delight gamers everywhere on November 16th,” says Nvidia, in a remarkable new blog post that effectively calls a halt to the upcoming launch of its most contentious of RTX 4000 releases, hinting that it’ll be renamed when it does eventually re-emerge. RTX 4070 Ti anyone?

Unveiled by Nvidia boss Jensen Huang on September 20th, the RTX 4000 line of graphics hardware is based on the new Ada Lovelace architecture, fabricated on TSMC’s 4nm process. The flagship GPU – the GeForce RTX 4090 – released a couple of days ago and was lauded for its massive generational leap over the existing RTX 3000 flagships, RTX 3090 and RTX 3090 Ti. However, it’s fair to say that the same level of acclaim and anticipation was not bestowed upon the RTX 4080, announced in two very different flavours.

The table below reveals that the spec differential wasn’t just about the amount of onboard VRAM: each RTX 4080 model was based on an entirely different core processor, with wild variations in the amount of CUDA codes (compute power) and memory bandwidth. In effect, two very different graphics cards were being sold under the same name at very different price-points.

 RTX 4090 24GBRTX 4080 16GBRTX 4080 12GB
Die Size608mm²379mm²295mm²
CUDA Cores1638497287680
Boost Clock2.52GHz2.51GHz2.61GHz
Memory Interface384-bit256-bit192-bit
Memory Bandwidth1018GB/s742GB/s557GB/s
PSU Recommendation850W750W700W
PSU Cables4x 8-pin3x 8-pin2x 8-pin
Release dateOctober 12th, 2022November 16th, 2022Delayed
Nvidia GeForce RTX 4090 – the Digital Foundry video review.

Specifications are one thing, but real world performance is another. Nobody has been hands-on with the product, but Nvidia’s own benchmarks suggested that the gap between RTX 3080 and RTX 4080 would be nowhere near as large as the increase in performance offered by RTX 4090 over RTX 3090. Based on figures supplied by Nvidia itself, the 12GB RTX 4080 provided uplifts of varying levels up against the RTX 3080, sometimes beating 3090 Ti, sometimes not. There was the sense that while value looked reasonable against RTX 3080 Ti upwards, the $699 price-point of the RTX 3080 looked problematic for the $899 RTX 3080 12GB in a world where a new GPU line typically offers more power for your money. Performance figures for the 16GB model look somewhat healthier – at least here there is clear water between the generations, even if the price is still on the high side.

There was pressure on the RTX 4080 12GB on multiple fronts – pricing, naming, specifications – leading to Nvidia’s decision to delay and presumably rebrand the product. With that in mind, delaying and renaming the RTX 4080 12GB is the best route forward for Nvidia as it puts the focus back on the RTX 4080 16GB release to come, while difficult questions on the price/performance of the 12GB model can be addressed further on down the line.

Meanwhile, we do have a firm release date for the ‘one true’ RTX 4080 – November 16th – but competition in the graphics space may well be heating up in the meantime. AMD’s new RDNA 3 line of graphics products is set to be revealed on November 3rd, with the Radeon team also benefiting from a new process technology as well as a transition towards the chiplet technology that has served AMD so well in the CPU space (rather than one expensive chip, various parts of the design will be created from a collection of smaller, cheaper processors). Nvidia is a tough competitor, but there’s clearly an opportunity here for Team Red to make an impact. We’ll be reviewing all of the new graphics products as and when they appear.

- Advertisement -

Comments are closed.