When AMD began using TSMC’s 7nm process for the Zen 2 family of processors launched in November 2019, one of the most important messages of that launch was that it was important to be at the forefront of process node technology to be competitive. That move to the TSMC N7 was aided by the small chips used in desktop processors at the time, ensuring higher yields and better binning curves for desktop and business processors. However, between now and then, we’ve seen other companies take advantage of TSMC’s 5nm, 4nm process and talk about TSMC’s 3nm process coming to market in the next 12-24 months. During our roundtable discussion with CEO Dr. Lisa Su, I asked if the need to stay in the lead was still true.
To put this into perspective, AMD announced in late 2021 that it would use TSMC’s 5nm process for its Zen 4 chips in business CPUs in the second half of 2022. Then in early 2022, the company re-used Zen 4 chips, but this way back into desktop processors by the end of 2022. This is a significant delay between the first use of TSMC 5nm by smartphone manufacturers, which reached mass production in the third quarter of 2020, and Apple and Huawei were the first to take advantage. Even today, if we go beyond 5nm, Mediatek has already announced that its upcoming Dimensity 9000 smartphone chip on TSMC is 4nm and will hit the market earlier this year. TSMC’s 3nm process is expected to increase production in late 2022, for launch to consumers in early 2023. According to those indicators, AMD is behind one or two process nodes until Zen 4 chips come to market later this year.
I asked Dr. Sua at our roundtable on whether the need to be at the top of the process is critical to be competitive for them. After innovating around the chipset, I asked if being a lead partner with foundry partners and packaging partners (known as OSAT) was of great importance, especially when the lead competition seems willing to throw money at TSMC to take over. How could AMD aggressively establish a market-leading position in light of the complexity of production and the financial strength of the competition?
Dr. Su stated that AMD continues to innovate in all areas. For AMD looks, a leader in chiplet technology helped to connect the package. She further said that AMD had a strong delivery of 7nm 6nm presentation, monitored Zen 4 and 5nm, talking about 2D chips and 3D chips – AMD has all these things in the tool and uses the right technology for the right application. Dr. Su stepped it up technological roadmaps are all about making the right choices and the right intersections, and explicitly stated that our 5nm technology is highly optimized for high performance computing – not necessarily the same as some other 5nm technologies out there.
Although it is not explicitly stated that the need for leadership is no longer critical, this message follows an improved story from AMD that in the era of chiplets it becomes important how they are combined and packaged, which is undoubtedly more important than what exactly the process node is used. We’ve already seen this message from AMD’s main competitor Intel, where back in 2017 the company said it would rely heavily on optimized chips for every use case – this further crystallized in 2020, suggesting 24-36 chips on a single consumer desktop processor for dedicated client designs. With that in mind, it is constantly rumored that Intel will be a big buyer of TSMC 3nm in the coming years, so it will be interesting to see where AMD can take advantage of years of expertise in chips and packaging tools compared.
Friendly communicator. Music maven. Explorer. Pop culture trailblazer. Social media practitioner.