AMD confirms it is using TSMC 16nm FinFET for some future chips

Posted on Friday, Oct 21 2016 @ 13:52 CEST by Thomas De Maesschalck
AMD logo
Over a month ago AMD made some costly changes to the Wafer Supply Agreement (WSA) with GlobalFoundries to give itself more flexibility. Among other things, the new framework makes it possible for AMD to have certain products made with another foundry.

Not a lot of specifics were given but some questions about the WSA amendment came up during yesterday's earnings call with financial analysts. Overall, AMD CEO Lisa Su answered rather evasively but she did mention that over the last six months, AMD ramped five new FinFET products and that it's going really well thanks to having two sources ramping at the same time.

Furthermore, Su explicitly mentioned AMD is sourcing multiple products at 14nm and 16nm nodes across foundries. A month ago, we didn't know for sure whether AMD would pick Samsung or TSMC but this comment from Su settles the discussion. The 16nm FinFET is only offered by TSMC so now we know for sure that some of AMD's future products will be fabbed by TSMC.

Here's the full conversation:
Stacy Rasgon
Hi guys, thanks for taking my questions. I first wanted to ask about the second sourcing ability embedded in the new Wafer Supply Agreement. So you’ve said that you’re going to be doing some second sourcing, starting in 2017. I wonder, did that push to actually seek out that supply diversification come from you, or was it from specific requests from your customers? And given that, how do you guys make the decision on which products to manufacture at GLOBALFOUNDRIES versus manufacture somewhere else?

Lisa Su
Sure, Stacy. Let me start with that. I think, relative to our second sourcing or our supply sourcing flexibility, I think we make it on a product-by-product basis, based on where we are in the business. So we will have multiple products in 14-nanometer and 16-nanometer that will be sourced across foundries, and similarly when we talked about the Wafer Supply Agreement, we mentioned 7-nanometer as being a key target node for that. Relative to how we make the decisions, I think it’s a combination of factors. It includes the complexity of the product. It includes the timing, customers, all kinds of things. So I think that’s part of our product planning process.

Stacy Rasgon
Got it. But did the customers themselves have a hand in driving you to make that push to second source?

Lisa Su
I think that’s one element, but frankly, I think what’s more important to me is, I need to commit a strong, five-year product road map to the customers, and so we want to make sure that we have all the flexibility to ensure nothing happens. I’ll give you just a little bit of context, Stacy, because I think you know our business well. In this past six months, we’ve ramped five new FinFET products. I mean it’s the fastest transition we have ever made in a process node, and it’s gone really, really well. And I think what’s helped us with that is the fact that we’ve had two sources ramping at the same time.

About the Author

Thomas De Maesschalck

Thomas has been messing with computer since early childhood and firmly believes the Internet is the best thing since sliced bread. Enjoys playing with new tech, is fascinated by science, and passionate about financial markets. When not behind a computer, he can be found with running shoes on or lifting heavy weights in the weight room.

Loading Comments