AMD denies MI455X delay claims
AMD has pushed back on a report claiming its next-generation AMD Instinct MI455X accelerators are facing manufacturing delays that would limit production in 2026. The claim came from SemiAnalysis, which tied the alleged slip to AMD’s first rack-scale buildout and said broader availability would move out to 2027.
SemiAnalysis said engineering samples and low-volume production of AMD’s first rack-scale MI455X UALoE72 system would land in H2 2026, but that mass production and “first production tokens” would not arrive until Q2 2027. The same piece argued this would leave NVIDIA’s next platform, referenced as Vera Rubin and its rack-scale systems, without an AMD alternative in the same window.
Engineering samples and low volume production of AMD’s first rack scale MI455X UALoE72 system will be in H2 2026 while due to manufacturing delays, the mass production ramp and first production tokens will only be generated on an MI455X UALoE72 by Q2 2027.
— Semi Analysis
AMD’s response came via a post from Anush Elangovan (AMD VP, SW Development), who disputed the delay timeline. He said MI455X is “right on target for shipments in 2H2026,” and he described the bring-up pace as faster than he has seen on comparable complex GPU programs.
BS. Right on target for H2 2026. https://t.co/0OjifOiuYW
— Anush Elangovan (@AnushElangovan) February 17, 2026
Speed is the moat.
MI455X is both fast, and is the fastest I have seen in execution for bringup of a complex GPU platform. MI455X is right on target for shipments in 2H2026 – Irrespective of what @SemiAnalysis_ says – rev your engines because speed is coming.— Anush Elangovan (@AnushElangovan) February 17, 2026
The post does not include a detailed volume forecast, a customer ramp schedule, or a breakdown of what “shipments” means in this context, for example engineering samples versus production allocations. Companies like Intel and AMD sometimes use “shipments” to describe deliveries limited to specific customers or designs, which can blur the line between “launching” and “shipping.”
Source: Anush Elangovan (AMD) , SemiAnalysis
