As we know, Nvidia still uses the 12 nm FinFET manufacturing process, while AMD has already made the leap to 7 nm with its Vega and Navi graphics. The green company was at the Credit Suisse Annual Technology Conference 2019 and gave some hints about the leap to 7 nm: When will it do that?
The company answered the big question about the 7 nm and when this node will be used for their graphics cards. Knowing that Turing’s graphical architecture was widely used long before its official announcement, it’s surprising that we still don’t know anything about Amps and its jump to 7 nm.
That’s what the green company has responded to:
We are always busy building our general architectures for many of our different markets, be it games, professionals, or data centers in general. So don’t worry, we’re working on our technology. Our innovations and process technology have been exceptional in recent years. So stay tuned. We will always have something for you in the future, but we want to surprise everyone with our general roadmaps and when things are going well.
Besides our processors being first-class in whatever industry they have worked in, there are continuous improvements that we can make both in the current node and in the near future.
For each of our architectures, we continue the development cycle, select the existing node, or look at the future node to be developed that would be best for us to bring to market. It’s just the way we work, the way we bring architecture to the market. So this is a very important part, a node change or not, to improve overall computing performance.
At the pace that we see the AI progressing and the acceleration progressing, they need this ability to change this overall evolution along the way. They need the ability to program, otherwise, their investment is very, very difficult for them to achieve a return on investment. The return on investment you get when you buy the set of GPUs without full end in the platform is a great help for them. -Colette Kress, CFO, Nvidia.
If we read between the lines, we see that Nvidia doesn’t consider the 12 FinFET nodes “finished”, but there could be another generation of graphics cards with this node. The answer is somewhat ambiguous and can be interpreted in many ways.