A technology disadvantage?
Between 2012 and 2016, Advanced Micro Devices (AMD) lost significant market share in the CPU (central processing unit) and GPU (graphics processing unit) markets, lagging behind Intel (INTC) and NVIDIA (NVDA) in terms of technology.
AMD had a huge technology disadvantage because it was stuck on the outdated 28 nm (nanometer) node while Intel transitioned to the 14 nm node in 2014.
AMD’s Ryzen, EPYC CPU, and Polaris GPU put it on the 14 nm process technology course alongside its rivals. At similar nodes, the scope of performance differentiation is narrow.
AMD has thus managed to give strong competition to Intel, though the latter’s 14 nm node was more mature. This also helped AMD gain some market share from Intel and NVIDIA.
Can AMD catch up?
AMD is not looking to grow by beating the competition with technology leads. Instead, it’s looking to gain market share by being competitive. For this, it has to move along with competitors to the next technology node.
However, NVIDIA is already using TSMC’s (TSM) 12 nm node to build its Volta GPU. Intel would also start production on its 10 nm node by the end of 2018. It’s thus important for AMD to transition to the 12 nm or even the 7 nm node to stay competitive.
In September 2017, AMD’s foundry partner, GlobalFoundries, announced its 12 nm LP (low-power) node, which will likely begin production in 2018. AMD announced that it will build its Ryzen CPUs and Vega GPUs on the 12 nm node. The company did not mention its EPYC, which leaves uncertainty as to whether the server processor will also get an upgrade in process node.
Although AMD’s original technology roadmap does not include a transition to the 12 nm node, the company appears to be grabbing the opportunity to stay competitive until Global Foundries’ 7 nm node is ready. AMD is already creating competition for Intel at the 14 nm node, and if AMD moves to 12 nm or 7 nm technology, it will be even more competitive.
Next, we’ll assess AMD’s Ryzen family.