Samsung announced several updates to its process technology roadmap today at its Samsung Foundry Forum. Like Intel, Samsung intends to offer multiple generations of a single process node — but unlike Intel, it isn’t debuting a new strategy in lieu of maintaining a quick transition roadmap.
First, there’s interesting news concerning a fourth-generation 14nm product, 14LPU. For those keeping score at home, Samsung released 14nm Low Power Early (14LPE) first, followed by 14nm Low Power Plus (14LPP), which was broader ramp with more customers and up to 10% improved performance. Earlier this year, the company announced it would build a lower-cost variant of 14nm that didn’t sacrifice on power or performance, 14LPC. This fourth-generation 14LPU is meant explicitly for customers who are building “high performance, compute-intensive” applications. 14LPU is said to offer better performance than 14LPC, but Samsung hasn’t published details on how all four of its processes compare with one another; only 14LPP and 14LPE are listed on its website.
It’s not hard to read between these particular lines, however. When foundries add new iterations of existing process nodes, it’s typically because they’ve got customers who are interested in those process nodes and want to take advantage of them. Adding a higher power line could mean that Samsung wants to compete for more business from AMD, Nvidia, or even Qualcomm’s server processor — or it could be that Samsung meant relatively high-performance parts, and that it’s still focusing on lower power hardware at relatively low TDPs. Either way, Samsung clearly believes that 14nm will stick around as a long-lived node, and is worth investing in.
The company is still pushing ahead with its plans to roll out 10nm technology, however, and has also announced a new 10nm node, 10LPU. The company describes this as follows:
Samsung seems to be implying that 14LPU is a higher performance node, while 10LPU is a cost-optimized node. Meanwhile, the company also showed off a 7nm EUV wafer and provided an “update” on its 7nm EUV progress, but neglected to tell us anything about what that update was. This is one area where there’s notable difference between the various foundries — Intel says it intends to push to 7nm without EUV, but will deploy that tech at 5nm. TSMC has said something similar, but Samsung remains resolute that it can introduce EUV at the 7nm node.
The question of who will use that capacity is less clear. The number of companies deploying cutting-edge process nodes has dwindled as design costs have exploded, and there’s money to be made in improving yields and performance at older process nodes, if anyone can figure out how to do it.
It’ll be interesting to see how the TSMC / Samsung / GlobalFoundries match-up shifts over the next 12 to 18 months. Samsung beat TSMC to the 14/16nm node and reaped substantial financial benefits from doing so, while GF had to license Samsung tech rather than continue developing its own 14XM. Now, GlobalFoundries says it’ll bring its own 7nm to market, while TSMC is gunning hard for 10nm. With fewer companies designing on cutting-edge nodes, it means there’s less room for error than ever.