#AMD Says โ€œ#Heliosโ€ Racks And #MI400 Series #GPU On Track For 2H 2026
AMD has been working on the Open Rack Wide v3 specification, a double-wide rack that AMD calls โ€œHeliosโ€ and that is the platform for delivering #rackscale compute based on the โ€œAltairโ€ MI400 series GPU accelerators. This series includes the #MI450, #MI430X, and the #MI455X for the Helios rackscale systems, which will have 64, 72, or 128 GPUs per system, and possibly MI440X for eight-way system nodes.
https://www.nextplatform.com/2026/02/23/amd-says-helios-racks-and-mi400-series-gpus-on-track-for-2h-2026/
AMD Says โ€œHeliosโ€ Racks And MI400 Series GPUs On Track For 2H 2026

While releasing an update to its InferenceX AI inference benchmark test, formerly known as InferenceMax and thus far only having Nvidia and AMD testing

The Next Platform

[MI455X์— LPDDR5X ๋ชจ๋“ˆ 24๊ฐœ๊ฐ€ ๋ฐ•ํ˜€์žˆ๋‹ค.

MI455X์— LPDDR5X ๋ชจ๋“ˆ 24๊ฐœ๊ฐ€ ํƒ‘์žฌ๋˜์–ด ์žˆ์œผ๋ฉฐ, KV ์บ์‹œ ์šฉ๋„๋กœ ์‚ฌ์šฉ๋  ๊ฒƒ์œผ๋กœ ์˜ˆ์ƒ๋ฉ๋‹ˆ๋‹ค. ์ตœ๋Œ€ ๋Œ€์—ญํญ์€ 1.63TB/s๋กœ ์ถ”์ •๋˜๋ฉฐ, AMD์˜ ๋ฒ ๋‹ˆ์Šค ๊ฐ€์†๊ธฐ์™€ ๋น„๊ต๋˜์–ด ์–ธ๊ธ‰๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

https://news.hada.io/topic?id=25990

#lpdrd5x #kvcache #amd #mi455x #memory

MI455X์— LPDDR5X ๋ชจ๋“ˆ 24๊ฐœ๊ฐ€ ๋ฐ•ํ˜€์žˆ๋‹ค.

<ol> <li>KV ์บ์‹œ ์šฉ์œผ๋กœ ์“ฐ์ผ ๊ฒƒ์œผ๋กœ ์•Œ๋ ค์ง€๊ณ  ์žˆ์Œ.</li> <li>์ŠคํŽ™์€ ๋‡Œํ”ผ์…œ์ด์ง€๋งŒ ์‚ผ์„ฑ์ด 8533 64GB lpddr5x๋ฅผ ํ™๋ณดํ•œ ์ ์ด ์žˆ์œผ๋ฏ€๋กœ ์ตœ๋Œ€ 1.63...

GeekNews
#AMD threatens to go medieval on Nvidia with #Epyc and #Instinct: What we know so far
AMD teased its next-generation #AI accelerators at #CES2026, with CEO Lisa Su boasting the #MI500-series will deliver a 1,000x uplift in performance over its two-year-old #MI300X #GPU.
if AMD wants to stay competitive with Nvidia, MI500-series will need to deliver performance on par with if not better than Rubin Ultra Kyber racks.
AMD joins the #rackscale race with #MI455X #Helios racks.
https://www.theregister.com/2026/01/07/mi500x_amd_ai/
AMD threatens to go medieval on Nvidia with Epyc and Instinct: What we know so far

: AMD boasts 1000x higher AI perf by 2027 and pulls the lid off Helios compute tray ahead of 2H 2026 launch

The Register
#AMD touts Instinct #MI430X, #MI440X, and #MI455X #AI accelerators and #Helios #rackscale AI architecture at CES โ€” full MI400-series family fulfills a broad range of infrastructure and customer requirements
The newly disclosed MI440X and MI455X are set to be optimized for low-precision workloads, such as FP4, FP8, & BF16. The previously disclosed MI430X targets both sovereign AI and #HPC thus fully supports FP32 and FP64 technical computing and traditional #supercomputing
https://www.tomshardware.com/tech-industry/artificial-intelligence/amd-touts-instinct-mi430x-mi440x-and-mi455x-ai-accelerators-and-helios-rack-scale-ai-architecture-at-ces-full-mi400-series-family-fulfills-a-broad-range-of-infrastructure-and-customer-requirements
AMD touts Instinct MI430X, MI440X, and MI455X AI accelerators and Helios rack-scale AI architecture at CES โ€” full MI400-series family fulfills a broad range of infrastructure and customer requirements

There's an Instinct MI400X-series GPU for everyone in 2026

Tom's Hardware
#HPE to ship #rackscale #AI system using #AMD's #Helios in #2026
Announced ahead of HPE's Discover event in Barcelona this week. Helios is intended to operate a rack full of nodes fitted with accelerators as if they were one single large #GPU, like Nvidia's DGX GB200 NVL72 system, to which Helios is pitched as a rival. Helios will be a vehicle for its next-generation Instinct #MI455X GPUs and its 6th-gen Epyc CPUs, codenamed Venice, both of which are due next year.
https://www.theregister.com/2025/12/02/hpe_amd_helios_racks/
HPE backs AMD's Helios AI rack with Juniper's scale-up switch

: Hardware bundle ties next-gen accelerators to an Ethernet fabric arriving in 2026

The Register