The GeForce RTX 5070, a new GPU for gamers, creators, and developers based on the “Blackwell” architecture that NVIDIA has advertised as being twice as fast as the Ada-based GeForce RTX 4070, has been delayed, according a new rumor from an insider, one that claims the GPU will now be available for purchase in “early March” rather than late February.
“The RTX5070 will be delayed,” a tweet from the @Zed__Wang account dated February 11 reads. “Instead of February, it will be on the shelf in early March.”
“The GeForce RTX 5070 Ti GPU with 1,406 AI TOPS and GeForce RTX 5070 GPU with 988 AI TOPS will be available starting in February at $749 and $549, respectively,” NVIDIA wrote in its press release for the GeForce RTX 50 Series in January, although if MEGAsizeGPU is to be believed, those plans have changed due to the Radeon RX 9070 Series—incoming RDNA 4 competition from AMD that red team has confirmed will be launching soon, in early March.
Here’s what NVIDIA’s Andrew Burnes had to say about the GeForce RTX 5070 when it was revealed alongside the rest of the GeForce RTX 50 Series on January 6:
The GeForce RTX 5070 sports the powerful new cores of the NVIDIA Blackwell architecture 12GB GDDR7 memory, and has 672 GB/sec of total memory bandwidth, compared to the GeForce RTX 4070’s 504 GB/sec.
At 2560×1440, with full ray tracing and other settings maxed, and DLSS Multi Frame Generation enabled, GeForce RTX 5070 owners can play Black Myth: Wukong, Alan Wake 2, and Cyberpunk 2077 at high frame rates, with performance that is twice as fast on average compared to the GeForce RTX 4070.
The GeForce RTX 5070 will be available starting at $549.