Nvidia silently launched a new 12GB version of its RTX 5070 mobile GPU yesterday, featuring 24Gb (3GB) GDDR7 chips instead of the existing 16Gb (2GB) ones. The company also swapped from a 128-bit wide bus to a 192-bit bus, bringing the 5070 mobile up to parity with its desktop counterpart. That took the memory throughput from 384 GB/s to around 576 GB/s on an otherwise unchanged GPU. Apparently, that upgrade is worth an additional $500, at least, according to Framework's pricing.

The 12 GB 5070 mobile is otherwise identical to the 8 GB model — they share the same 4,608 CUDA core count and maximum TGP of 100W, with a boost clock of 2.4 GHz. The value Framework adds comes from the modularity, serving as a drop-in upgrade for existing hardware. Since literally no other brand offers similar graphics modules or add-in cards, it's a difficult product to compare directly. More rationally speaking, however, there are obviously better deals to be had, but Framework has clarified that the high prices are beyond its control.

Article continues below

The reason behind the price increase is pretty easy to understand. GDDR7 is already expensive, so using higher-density 3 GB chips (that were likely meant for a 50-series Super SKU at some point) is sure to add to the cost. And we're amidst an AI boom right now that has caused a memory shortage worldwide, which further exacerbates the situation.

While replying to a comment on the original announcement post, the company even warned that once its existing allocation of GDDR7 runs out, even the 8 GB RTX 5070 mobile will be restocked at a higher cost. Such is the reality of the times we're living in — the major memory manufacturers have sold their production lines to serve data center needs, so regular consumers are left with slim (and expensive) pickings.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Hassam Nasir is a die-hard hardware enthusiast with years of experience as a tech editor and writer, focusing on detailed CPU comparisons and general hardware news. When he’s not working, you’ll find him bending tubes for his ever-evolving custom water-loop gaming rig or benchmarking the latest CPUs and GPUs just for fun.