Learn about NVIDIA's upcoming GeForce RTX 60 series graphics cards, featuring Rubin architecture, TSMC 3nm process, and enhanced ray tracing and DLSS 5 performance.
© A. Krivonosov
NVIDIA is developing the next generation of GeForce RTX 60 series graphics cards, built on the Rubin architecture that has already debuted in AI-focused GPUs. These new cards will utilize TSMC's 3-nanometer FinFET process, enabling clock speeds of 2–3 GHz. The series will feature chips with "GR20x" numbering, with the top model designated as GR202.
The new GPUs come equipped with sixth-generation tensor cores and fifth-generation RT cores. This hardware will accelerate neural rendering, enhance DLSS 5 performance, and significantly boost real-time ray tracing—expect up to a 100% improvement over the current RTX 50 series. Meanwhile, traditional raster graphics will see a 30–35% increase, thanks to IPC gains, higher clock speeds, and improved energy efficiency.
In terms of memory, NVIDIA will stick with the GDDR7 standard but widen the bus and increase bandwidth for most models. The flagship RTX 6090 will feature 32 GB of memory on a 512-bit bus, the RTX 6080 will have 20 GB on a 320-bit bus, and the RTX 6070 will offer 16 GB on a 256-bit bus. These changes promise a substantial boost in memory bandwidth, particularly in the mid-range segment.
The launch of the GeForce RTX 60 series will help NVIDIA maintain its competitive edge in the market, expanding capabilities for ray tracing and AI-enhanced graphics in gaming. The focus is squarely on boosting performance in RT and DLSS 5 modes, which should deliver higher-quality effects and an improved gaming experience across all GPU tiers.