# rackscale Published articles for rackscale. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## AMD Helios Racks Are Running at Meta, With a Custom Instinct GPU and a $1 Billion Stake to Follow DevFeed: [AMD Helios Racks Are Running at Meta, With a Custom Instinct GPU and a $1 Billion Stake to Follow]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-10-29T16:41:34Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [rocm]() Tags: [agent](), [ai](), [ai-agent](), [amd](), [cpus](), [custom](), [deployment](), [enterprise](), [epyc](), [gpu](), [gpus](), [helios](), [instinct](), [meta](), [muse](), [networking](), [pensando](), [personal-ai](), [personal-ai-agent](), [rackscale](), [roadmaps]() ### AI overview AMD says its first Helios rackscale systems are installed and under test at Meta, with a broader deployment planned for the near future; it gives no rack count, site, or deployment date. The partnership spans EPYC CPUs, custom Instinct GPUs, Pensando networking, and ROCm. Meta, which has deployed millions of EPYC processors and is a lead 6th Gen EPYC customer, is also collaborating with AMD on future CPUs including Verano. Meta will deploy a custom Instinct GPU optimized for its workloads; joint software work has optimized recommendation training on MI350X GPUs, with improvements contributed to PyTorch, Triton, and ROCm. AMD also says it provides Day-0 support for Muse Spark Open on Instinct GPUs and Muse Glimmer on Ryzen AI PCs and Radeon GPUs. AMD committed $1 billion to Meta's equity offering. ### Source excerpt AMD says the first AMD Helios rackscale systems are installed and under test at Meta, with a broader deployment to follow, and that the two companies have aligned product roadmaps around Meta's workloads across EPYC CPUs, Instinct GPUs, Pensando networking, and ROCm. The same announcement puts EPYC behind Meta's Muse personal AI agent, adds Day-0 The post AMD Helios Racks Are Running at Meta, With a Custom Instinct GPU and a $1 Billion Stake to Follow appeared first on StorageReview.com. ## Accelerated Server Sales Are Themselves Accelerating DevFeed: [Accelerated Server Sales Are Themselves Accelerating]() Original publisher: [Read original article]() Author: Timothy Prickett Morgan Published: 2026-09-21T19:21:53Z Content type: news Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [GPU](), [Server](), [Computing](), [idc](), [systems](), [cloud-infrastructure](), [data-processing](), [data]() Tags: [2025](), [accelerated](), [compute](), [computing](), [cpu](), [dell](), [gpu](), [gpu-accelerated](), [hpe](), [idc](), [infrastructure](), [inspur](), [lenovo](), [nvidia](), [rackscale](), [server](), [supermicro](), [xpu]() ### AI overview The article examines a sharp rise in global server sales during the second quarter of 2026, including strong growth in GPU-accelerated systems. It reports that GPU system unit sales declined as the market shifted from standalone server nodes toward rackscale systems with shared memory domains. ### Source excerpt Do you know what the difference is between an F-18A taking off from the deck of an aircraft carrier ... ## Startup d-Matrix Will Pair Its Raptor Memory-Based XPU To Nvidia Rackscale Iron DevFeed: [Startup d-Matrix Will Pair Its Raptor Memory-Based XPU To Nvidia Rackscale Iron]() Original publisher: [Read original article]() Author: Jeff Burt Published: 2026-09-10T14:52:10Z Content type: article Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [Artificial Intelligence](), [Inference](), [GPU](), [Hardware](), [data centers](), [xpu]() Tags: [accelerators](), [ai](), [blackwell-gpu](), [compute](), [corsair](), [d-matrix](), [inference](), [latency](), [low-latency-inference](), [lpx](), [mgx](), [nvidia](), [rackscale](), [raptor](), [xpu]() ### AI overview This article reports on d-Matrix pairing its memory-based Raptor XPU with Nvidia Blackwell GPU accelerators in rack-scale systems. The approach assigns compute-intensive inference prefill work to GPUs and token-generation decode work to the XPU, aiming to improve low-latency AI inference and reduce costs. ### Source excerpt Engineers with startup d-Matrix have spent the past seven years putting together the pieces of what ... ## The GenAI Boom Will Lift Supermicro, But It Will Lift Others, Too DevFeed: [The GenAI Boom Will Lift Supermicro, But It Will Lift Others, Too]() Original publisher: [Read original article]() Author: Timothy Prickett Morgan Published: 2026-08-13T02:51:12Z Content type: opinion Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [Generative AI](), [data centers](), [GPU](), [Nvidia](), [cpu](), [Server](), [Software](), [xpu]() Tags: [accelerated](), [accelerators](), [ai](), [business](), [compute](), [cpu](), [datacenter](), [genai](), [gpu](), [nvidia](), [rackscale](), [scale](), [scale-out](), [scale-up](), [server](), [storage](), [supermicro](), [system](), [xpu]() ### AI overview An analysis of how growing demand for generative AI systems is benefiting Supermicro and other original equipment and design manufacturers. It examines Supermicro's systems and subsystems businesses, customer and channel relationships, and its server, rack-scale, storage, switching, CPU, GPU, and accelerator offerings. ### Source excerpt It seems like every business, with the exceptions of a few monster companies that gain a monopoly (e ... ## Just How Rosy Are Those AI Infrastructure Spending Forecasts? DevFeed: [Just How Rosy Are Those AI Infrastructure Spending Forecasts?]() Original publisher: [Read original article]() Author: Timothy Prickett Morgan Published: 2026-07-30T22:02:46Z Content type: opinion Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [AI Infrastructure](), [Artificial Intelligence](), [idc](), [cpu](), [Arm](), [Nvidia](), [NVIDIA Blackwell](), [Hopper](), [x86](), [Network](), [Server]() Tags: [2026](), [ai](), [ai-accelerator](), [ai-host](), [ai-infrastructure](), [amd](), [analysts](), [arm](), [blackwell](), [compute](), [cpu](), [driving](), [google](), [gpu](), [hopper](), [idc](), [infrastructure](), [intel](), [networking](), [nvidia](), [nvl72](), [processors](), [rackscale](), [tpu](), [trainium](), [x86](), [xpu]() ### AI overview The article examines IDC's AI infrastructure spending forecast through 2030 and compares it with projections presented by AMD CEO Lisa Su. It reports higher spending forecasts and describes the growing share of Arm-based AI host systems, including systems using Nvidia GPUs and rackscale platforms. ### Source excerpt The second quarter ended four weeks ago, but the market researchers and analysts of the world - pres ... ## Dell Makes The Profits Up In Volume For Booming AI Servers DevFeed: [Dell Makes The Profits Up In Volume For Booming AI Servers]() Original publisher: [Read original article]() Author: Timothy Prickett Morgan Published: 2026-06-01T15:07:17Z Content type: news Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [Artificial Intelligence](), [dell](), [GPU](), [Server](), [Inference](), [Cloud]() Tags: [ai](), [compute](), [dell](), [genai](), [gpu](), [inference](), [nvidia](), [poweredge](), [rackscale](), [servers](), [training]() ### AI overview The article examines Dell's growing AI server business, arguing that high-volume sales can offset the relatively modest operating margins of these systems. It attributes the growth to strong customer demand and increased access to Nvidia GPUs, while highlighting Nvidia's influence over the supply of AI compute for both training and inference. ### Source excerpt The last time that Dell had a blowout quarter like the one that it has turned in was probably back w ... ## Oracle First In Line For AMD "Altair" MI450 GPUs, "Helios" Racks DevFeed: [Oracle First In Line For AMD "Altair" MI450 GPUs, "Helios" Racks]() Original publisher: [Read original article]() Author: Timothy Prickett Morgan Published: 2025-10-14T17:01:40Z Content type: news Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [Multi-GPU](), [data centers](), [groq]() Tags: [ai-world](), [altair](), [amd](), [compute](), [epyc](), [gpu](), [gpus](), [helios](), [mi450](), [mi450x](), [oci](), [oracle](), [pensando](), [rackscale](), [venice](), [volcano]() ### AI overview Oracle plans to begin deploying 50,000 AMD Altair GPU sockets in OCI in Q3 2026, expanding in 2027 and beyond. The excerpt estimates this is about 700 Helios racks and reports a projected 200 MW draw. Although Oracle says the cluster will use Helios double-wide racks, the article infers it may use the MI450 GPU with up to 432 GB of HBM4 per GPU rather than the MI450X normally associated with Helios; a custom MI450X configuration is another possibility. AMD's stated Helios figures include up to 2.9 exaflops at FP4, 31 TB of aggregate HBM4, and UALink over Ethernet for sharing GPU memory. The cluster is described as general OCI capacity customers can rent, not as dedicated to Oracle's OpenAI contracts. GPU choice, rack configuration, and several network details remain uncertain in the excerpt. ### Source excerpt It is Oracle OpenWorld CloudWorld AI World this week, so we expect a lot of AI infrastructure announcements from Big Red, with AI being the biggest new