Nvidia just announced a new GPU variant in the weirdest way possible: buried it in a game driver update blog post.
Alongside the release of its Game Ready 596.36 WHQL driver, the company also confirmed the launch of a 12GB GDDR7 configuration of the GeForce RTX 5070 laptop GPU.
Why is Nvidia doing this now?
The short answer: memory supply. Nvidia has already acknowledged that demand for its RTX GPUs exceeds the availability of the 16Gb G7 modules, which are currently used in most RTX 50-series products.
The new 12GB variant sidesteps the issue by using 24Gb G7 modules built on a different manufacturing process that produces 3GB chips instead of 2GB. The variant is sourced from a separate supply chain that Samsung and Micron have ramped up.
The result is a third configuration sitting above the baseline 8GB model, giving laptop manufacturers more options to space out their variants and potentially making them a tad cheaper than those equipped with the higher configuration.
The 12GB RTX 5070 laptop GPU won’t replace the 8GB variant; both will coexist.
Is more VRAM better here?
Not necessarily. According to an Engadget report, the 12GB variant might not provide a noticeable performance improvement over the 8GB variant, unless Nvidia widens the memory bus to 192-bit, which isn’t confirmed for the new GPU yet.
Laptops with the 12GB GeForce RTX 5070 laptop GPU are expected to begin shipping around June 2026. Manufacturers like Lenovo, MSI, and XMG will be among the first to use the graphics chip.
Everything else aside, I was most baffled by the fact that Nvidia hid a new GPU variant inside a driver blog post, rather than a typical launch. This speaks to how supply-chain pressures are reshaping product strategy. The new GPU, in my opinion, is a memory sourcing solution designed to benefit consumers.
If the pricing stays reasonable, the 12GB GeForce RTX 5070 could genuinely fill a gap. If it doesn’t, it could cannibalise the RTX 5070 TI’s already slim value proposition.