RTX 4080 SUPER is powered by the Nvidia AD103

Previous reports suggested that NVIDIA was gearing up to unveil three GeForce RTX 40 SUPER models: the RTX 4080 SUPER, RTX 4070 Ti SUPER, and RTX 4070 SUPER. Among them, murmurs hinted at the RTX 4080 SUPER transitioning to the AD102, equipped with 20GB memory, and an augmented memory bandwidth stretching to 320 bits. These speculations were subsequently met with skepticism and contradiction.

The discourse surrounding the chipset underpinning the RTX 4080 SUPER now seems to be drawing to a conclusive end. Astute observers have unearthed the GPU model’s PCI ID. The newly revealed GPU, bearing the PCI ID 2703, corresponds unequivocally to the RTX 4080 SUPER—presumably a desktop variant—powered by the AD103.


  • 2684 – AD102 [GeForce RTX 4090]
  • 26b1 – AD102GL [RTX 6000 Ada Generation]
  • 26b2 – AD102GL [RTX 5000 Ada Generation]
  • 26b5 – AD102GL [L40]
  • 26b5 – AD102GL [L40]
  • 26b9 – AD102GL [L40S]
  • 26f5 – AD102GL [L40 CNX]
  • 2703 – AD103 [GeForce RTX 4080 SUPER]
  • 2704 – AD103 [GeForce RTX 4080]
  • 2717 – GN21-X11 [GeForce RTX 4090 Laptop GPU]
  • 2730 – AD103GLM [RTX 5000 Ada Generation Laptop GPU]
  • 2757 – GN21-X11
  • 2770 – AD103GLM [RTX 5000 Ada Generation Embedded GPU]

Earlier whispers intimated that NVIDIA might opt for a fusion of chipsets within the RTX 40 SUPER series. For instance, the RTX 4080 SUPER could potentially harness either the AD102 or AD103, while the RTX 4070 SUPER could oscillate between AD103 and AD104. As of the present juncture, such possibilities remain to be discerned.

The GeForce RTX 4080, unveiled last September, boasts the AD103-300 GPU founded upon the Ada Lovelace architecture. It encompasses 76 SMs, translating to 9,728 CUDA cores. Its base clock is pegged at 2.21GHz, with a boost clock of 2.51GHz. Augmented with 16GB of GDDR6X memory, it features a 256-bit memory bandwidth and a memory speed of 23Gbps. The official Manufacturer’s Suggested Retail Price (MSRP) is slated at $1199.