# AMD expands AI hardware across data centre and edge > Live situation record from CLSTR: https://clstr.news/situations/amd-ai-hardware-rollout > Updated: 2026-09-10T10:59:40.000Z. Sources: 298. Developments: 38. AMD’s AI push has accelerated through new hardware, software, and partnerships. In early July, the company released the Ryzen AI Halo mini-PC, a $3,999 workstation built around the Ryzen AI Max+ 395 APU. AMD also unveiled its 2nm Zen 6-based EPYC “Venice” server line and the Helios rack-scale platform, which integrates 72 Instinct MI455X GPUs. Anthropic has committed up to $5 billion and 2 GW of MI455X GPUs for Claude model training, with the first gigawatt slated for early 2027. At the Advancing AI 2026 event, AMD expanded its Helios solution, claiming it delivers 30% more inference tokens per dollar than competitors. The company also introduced the Kria AI system-on-module and a Robotics Developer Platform using Ryzen AI Embedded X100 processors. To bolster inference, AMD entered a definitive agreement to acquire Toronto-based startup Taalas, which hard-wires AI model weights into silicon using a structured ASIC approach to reduce compute and memory bottlenecks. This technology is intended for integration across Instinct GPUs, Helios systems, and EPYC CPUs. Recent software updates include Day 0 support for the Qwen 3.8 27B model, enabling local execution on Ryzen AI Max+ 395 and Radeon AI PRO R9700 hardware. Developers can also use the Lemonade hardware-aware inference layer to optimize load distribution. Further infrastructure developments include a joint reference design with Schneider Electric for the Helios platform to assist in deploying high-density AI environments. The launch of ROCm 10 and the ROCm.AI toolkit introduced Hyperloom, an autonomous agentic system designed to automate performance tuning and reduce optimization time from weeks to hours. AMD reports that the Instinct MI355X platform achieved a 3.3x increase in inference throughput for models like GLM-5 and DeepSeek-R1, alongside a 2.4x gain in training efficiency for models such as DeepSeek-V3. ## Claims - AMD unveiled the Threadripper Halo Station at IFA 2026. (corroborated by 16 sources) - The Threadripper Halo Station can run AI models with over one trillion parameters locally. (corroborated by 16 sources) - The workstation features a Ryzen Threadripper PRO 9995WX CPU with 96 cores and 192 threads. (corroborated by 15 sources) - The system can be configured with up to four Instinct MI350P accelerators, providing 576GB of HBM3e memory. (corroborated by 14 sources) - The total system memory can reach up to 2.6 TB. (corroborated by 14 sources) - Qualcomm and Amazon have entered a multi-year partnership to develop custom AI inference chips. (corroborated by 13 sources) - AMD signed a definitive agreement to acquire Taalas on August 6, 2026. (corroborated by 11 sources) - Taalas builds chips that hard‑wire AI models directly into silicon for inference. (corroborated by 11 sources) - The acquisition is subject to customary closing conditions and regulatory approvals. (corroborated by 11 sources) - AMD plans to integrate Taalas technology into its Instinct GPUs and Helios rack‑scale systems. (corroborated by 10 sources) - Vamsi Boppana, senior vice president of AMD’s AI Group, said the acquisition strengthens AMD’s AI portfolio. (corroborated by 10 sources) ## Timeline ### 2026-09-10: AMD unveils Threadripper Halo Station AI workstation AMD unveiled the Threadripper Halo Station, a liquid-cooled workstation capable of running trillion-parameter AI models locally to reduce cloud dependency. 4 sources. https://clstr.news/cluster/amd-unveils-threadripper-halo-station-ai-workstation ### 2026-09-07: Qualcomm and Amazon partner on AI chips; AMD unveils Threadripper Halo Station Qualcomm and Amazon have partnered to develop custom AI inference chips and optical connectivity for AWS, while AMD unveiled its Threadripper Halo Station workstation for local large-scale AI models. 37 sources. https://clstr.news/cluster/amd-expands-ai-hardware-with-workstation-prototype-and-european-supercomputer ### 2026-08-31: AMD expands AI capabilities with Ryzen AI X100 and on-device moderation AMD is advancing its physical AI and gaming capabilities with the Ryzen AI X100 SoC, on-device real-time voice moderation via CtrlVox, and the AI-optimized ROCm 10 software platform. 10 sources. https://clstr.news/cluster/amd-targets-robotics-and-physical-ai-with-new-soc-and-rocm-10-software ### 2026-08-26: AMD launches ROCm 10 and ROCm. AI toolkit to boost performance AMD has released ROCm 10 and the ROCm. AI toolkit, promising up to 3.3x faster AI inference and 2.4x faster training on Instinct GPUs, supported by recent developer innovations from the DevMaster Hackathon. 4 sources. https://clstr.news/cluster/amd-launches-rocm-10-and-rocm-ai-toolkit-to-boost-performance ### 2026-08-18: Schneider Electric and AMD launch Helios platform reference design for AI Schneider Electric and AMD have released a new reference design for the AMD Helios rackscale solution to accelerate the deployment of high-density AI environments and data center infrastructure. 2 sources. https://clstr.news/cluster/schneider-electric-and-amd-launch-helios-platform-reference-design-for-ai ### 2026-08-15: AMD enables local execution of Qwen 3.8 27B AI model AMD has launched Day 0 support for the Qwen 3.8 27B AI model, enabling high-performance local execution on Ryzen AI Max+ processors and Radeon AI PRO GPUs. 2 sources. https://clstr.news/cluster/amd-enables-local-execution-of-qwen-38-27b-ai-model ### 2026-08-11: AMD and Intel navigate shifting AI market demands The AI sector is shifting focus toward inference and specialized hardware. AMD is expanding its robotics and inference capabilities, while Intel seeks $20 billion to bolster its foundry business. 10 sources. https://clstr.news/cluster/amd-advances-ai-ecosystem-through-developer-sessions-and-ryzen-ai-halo-testing ### 2026-08-11: Wiwynn unveils AMD-powered servers for Agentic AI at OCP APAC Summit At the OCP APAC Summit, Wiwynn introduced AMD-powered servers for Agentic AI, while OCP focused on scaling hardware management for dense, power-constrained AI data centers. 3 sources. https://clstr.news/cluster/wiwynn-unveils-amd-powered-servers-for-agentic-ai-at-ocp-apac-summit ### 2026-08-11: AMD to acquire Taalas to enhance AI inference capabilities AMD is acquiring Toronto-based Taalas to integrate specialized AI inference chip technology into its Instinct GPU portfolio, aiming to reduce memory bottlenecks and boost performance for specific models. 7 sources. https://clstr.news/cluster/amd-to-acquire-taalas-to-expand-ai-inference-hardware ### 2026-08-06: AMD to acquire AI inference chip startup Taalas AMD agreed to buy Toronto AI chip startup Taalas, which hard‑wires model weights into silicon, to boost its AI inference hardware and compete with Nvidia. 27 sources. https://clstr.news/cluster/amd-acquires-ai-inference-chip-startup-taalas-to-boost-accelerator-roadmap ### 2026-08-03: Weka partners with Andromeda to standardize AI data storage Weka and Andromeda have teamed up to make Weka’s NeuralMesh the primary AI storage layer for Andromeda’s GPU clusters, delivering uniform performance, rapid deployment and high throughput across global compute‑ 2 sources. https://clstr.news/cluster/weka-partners-with-andromeda-to-standardize-ai-data-storage ### 2026-08-01: AMD launches Helios AI rack and integrated robotics platform AMD unveiled its Helios rack‑scale AI solution and a new integrated robotics platform based on Ryzen AI X100 processors, promising up to 30% more inference efficiency and a 3.4× speed edge over Nvidia. 2 sources. https://clstr.news/cluster/amd-launches-helios-ai-rack-and-integrated-robotics-platform ### 2026-07-29: WEKA‑Andromeda and Cerebras‑AMD Announce AI Infrastructure Partnerships WEKA partners with Andromeda to provide consistent AI storage across global GPU clusters, while Cerebras teams with AMD to deliver a disaggregated inference system that boosts performance fivefold. 9 sources. https://clstr.news/cluster/wekaandromeda-and-cerebrasamd-announce-ai-infrastructure-partnerships ### 2026-07-28: AMD chips drive robotics and server AI platforms AMD’s AI‑focused processors are powering Foundation Future Industries’ Phantom robots and Supermicro’s new H15 servers, offering on‑board intelligence for industrial, defense and high‑performance AI workloads. 2 sources. https://clstr.news/cluster/amd-chips-drive-robotics-and-server-ai-platforms ### 2026-07-27: Schneider Electric and AMD unveil AI‑ready Helios reference design Schneider Electric and AMD released a reference design for the AMD Helios platform, enabling high‑density AI deployments up to 10.4 MW with advanced cooling and digital‑twin tools. 3 sources. https://clstr.news/cluster/schneider-electric-and-amd-unveil-aiready-helios-reference-design ### 2026-07-26: AMD Unveils AI‑Focused EPYC Venice CPUs and Helios Rack System AMD unveiled EPYC Venice CPUs (up to 256 cores) and the Helios rack with 72 MI455X GPUs, partnering with Anthropic for up to 2 GW GPU capacity, while announcing collaborations with VAST Data, ASUS, Advantech, C 39 sources. https://clstr.news/cluster/amd-unveils-future-epyc-zen-78-roadmap-and-helios-ai-rack-system ### 2026-07-23: AMD launches ROCm.ai AI development platform for its hardware AMD unveiled ROCm.ai, an AI‑native platform that boosts inference 3.3× and training 2.4× on its hardware, integrating CLI tools, AI code‑assistant knowledge and automated optimization, available globally from A 4 sources. https://clstr.news/cluster/amd-launches-rocmai-ai-development-platform-for-its-hardware ### 2026-07-23: AMD launches Helios rack‑scale AI platform and new EPYC processors AMD unveiled its Helios rack‑scale AI system, Zen 6 EPYC CPUs, and new partnerships—including Schneider, Samsung and Anthropic—aiming to challenge Nvidia in a $2 trillion AI market. 92 sources. https://clstr.news/cluster/amd-launches-helios-ai-rack-to-challenge-nvidias-datacenter-dominance ### 2026-07-23: AI infrastructure race reshapes AMD‑Nvidia competition AI infrastructure is moving from chips to rack‑scale systems, with AMD investing $60 bn to challenge Nvidia, while firms grapple with fragmented data, security, and talent needs to power next‑gen AI workloads. 2 sources. https://clstr.news/cluster/ai-infrastructure-race-reshapes-amdnvidia-competition ### 2026-07-22: AMD unveils 256‑core Zen 6 Epyc 9006 “Venice” server processor AMD introduced the Zen 6‑based Epyc 9006 “Venice”, a 256‑core, 512‑thread server CPU with 2 nm cores, 6 nm I/O, PCIe 6.0, DDR5‑12800 and AI‑focused performance gains. 2 sources. https://clstr.news/cluster/amd-unveils-256core-zen-6-epyc-9006-venice-server-processor ### 2026-07-19: AMD Helios AI System Powers Microsoft Azure and Anthropic Partnership AMD’s Helios AI rack system will run on Azure and power Anthropic’s Claude, with up to $5 bn investment and 2 GW of GPUs slated for deployment by 2028. 97 sources. https://clstr.news/cluster/mexicos-aidriven-tech-exports-to-united-states-surge ### 2026-07-15: AMD unveils 2nm Zen 6 EPYC ‘Venice’ servers with up to 256 cores for AI workloads AMD’s upcoming EPYC Venice server CPUs, built on 2 nm, will top out at 256 cores and 1.6 TB/s memory bandwidth, targeting AI workloads that need varied CPU profiles. 3 sources. https://clstr.news/cluster/amd-unveils-new-epyc-server-cpus-designed-for-aiagent-workloads ### 2026-07-06: AMD launches Ryzen AI Halo mini‑PC for local AI computing AMD unveiled the $3,999 Ryzen AI Halo mini‑PC with 128 GB unified memory and a Ryzen AI Max+ 395 SoC for local AI workloads, sold only in the US via MicroCenter. 2 sources. https://clstr.news/cluster/amd-launches-ryzen-ai-halo-minipc-for-local-ai-computing ### 2026-07-02: Enterprise AI Gains Momentum with AMD PCs and AWS Investment AMD’s AI‑ready PCs and a $1 billion AWS investment push enterprise AI deployment, enabling local inference, faster rollout and enhanced security for businesses and government. 3 sources. https://clstr.news/cluster/enterprise-ai-gains-momentum-with-amd-pcs-and-aws-investment ### 2026-06-25: Future, AI CROSS and AMD Unveil Major AI Initiatives in 2026 Future launches an AI research lab, AI CROSS releases an MCP server for its RCS platform, and AMD debuts the Helios rack‑scale AI system at Interop Tokyo 2026, marking major AI initiatives across Japan and the 3 sources. https://clstr.news/cluster/future-ai-cross-and-amd-unveil-major-ai-initiatives-in-2026 ### 2026-06-23: AMD Ryzen AI 9 HX 470 Drives Strong Performance in BOSGAME VTA-439 Mini PC The BOSGAME VTA-439 mini PC, powered by AMD Ryzen AI 9 HX 470 and Radeon 890M, delivered 35‑78 fps in Cyberpunk 2077 benchmarks, showing strong performance for compact systems. 2 sources. https://clstr.news/cluster/amd-ryzen-ai-9-hx-470-drives-strong-performance-in-bosgame-vta-439-mini-pc ### 2026-06-23: AMD Accelerates AI Server Growth with New CPUs, GPUs and Supercomputer Wins AMD targets strong AI server growth with Instinct accelerators and EPYC CPUs, reporting 57% data‑center revenue rise and AMD tech now in 4 of the top‑10 TOP500 supercomputers. 4 sources. https://clstr.news/cluster/amd-powers-new-supercomputer-milestones-with-epyc-cpus-and-instinct-gpus ### 2026-06-19: GMKtec launches Evo‑X3 AI mini PC with up to 128 GB RAM GMKtec’s Evo‑X3 AI mini PC, featuring an AMD Ryzen AI Max+ 395 APU, up to 128 GB RAM and 16 TB SSD, opens early‑access registration on 22 June 2026 with a $20 discount; pricing is yet to be announced. 2 sources. https://clstr.news/cluster/gmktec-launches-evox3-ai-mini-pc-with-up-to-128-gb-ram-1 ### 2026-06-18: GMKtec launches EVO‑X3 AI mini PC with up to 128 GB RAM GMKtec’s EVO‑X3 AI mini PC, featuring an AMD Ryzen AI Max+ 395 CPU, up to 128 GB RAM and 50 TOPS NPU, is slated for global launch on 29 June after early‑access registration begins 22 June. 3 sources. https://clstr.news/cluster/gmktec-launches-evox3-ai-mini-pc-with-up-to-128-gb-ram ### 2026-06-17: AMD expands AI portfolio with MEXT acquisition and Rackspace partnership AMD bought AI memory startup MEXT and teamed with Rackspace to launch 30 MW of AMD‑based AI compute for regulated‑industry cloud services. 2 sources. https://clstr.news/cluster/amd-expands-ai-portfolio-with-mext-acquisition-and-rackspace-partnership ### 2026-06-16: AMD boosts AI hardware with MEXT acquisition and Ryzen AI Halo platform AMD acquired MEXT for AI memory tech and unveiled the €3,700 Ryzen AI Halo platform, lifting its market value above $900 bn and sending shares to record highs. 3 sources. https://clstr.news/cluster/amd-boosts-ai-hardware-with-mext-acquisition-and-ryzen-ai-halo-platform ### 2026-06-14: AMD launches Ryzen AI Halo mini PC for AI workstations AMD’s Ryzen AI Halo mini PC, featuring a Ryzen AI Max+ 395, 128 GiB unified memory and 2 TB SSD, is on pre‑order in the US for $3,999.99, with delivery from July 2026. 2 sources. https://clstr.news/cluster/amd-launches-ryzen-ai-halo-mini-pc-for-ai-workstations ### 2026-06-13: OneXPlayer launches X2 Mini Pro 3‑in‑1 gaming handheld with AMD Ryzen AI Max+ OneXplayer's X2 Mini Pro, a 3‑in‑1 handheld/ tablet/ mini‑PC with AMD Ryzen AI Max+ 388, launches mid‑June via Indiegogo at $2,399‑$2,799, featuring an 8.8‑inch OLED, up to 64 GB RAM and optional liquid cooling 4 sources. https://clstr.news/cluster/onexplayer-launches-x2-mini-pro-3in1-gaming-handheld-with-amd-ryzen-ai-max ### 2026-06-13: AMD launches Ryzen AI Halo mini PC with 128 GB unified memory for $3,999 AMD opened pre‑orders for the $3,999 Ryzen AI Halo mini PC, featuring 128 GB unified memory, Windows or Linux support, and targeting AI developers as a cheaper alternative to Nvidia’s DGX Spark. 4 sources. https://clstr.news/cluster/amd-launches-ryzen-ai-halo-mini-pc-with-128-gb-unified-memory-for-3999 ### 2026-06-11: PELADN launches HO5 mini PC with AMD Ryzen AI 9 processor in China PELADN unveiled the HO5 mini PC in China, powered by an AMD Ryzen AI 9 HX470 CPU, 32 GB RAM, 1 TB SSD, priced at 7,599 yuan, with Wi‑Fi 7, Bluetooth 5.4 and extensive I/O. 2 sources. https://clstr.news/cluster/peladn-launches-ho5-mini-pc-with-amd-ryzen-ai-9-processor-in-china ### 2026-06-02: AMD Launches Ryzen AI Halo Platform and Max PRO Processors for Local Agentic AI AMD unveiled the Ryzen AI Halo platform and Max PRO 400 series processors, delivering up to 128 GB unified memory and high‑performance AI compute for desktop development of large models. 2 sources. https://clstr.news/cluster/amd-launches-ryzen-ai-halo-platform-and-max-pro-processors-for-local-agentic-ai ### 2026-06-02: Supermicro and AMD unveil 72‑GPU Helios rack for AI at Computex Supermicro and AMD showcased the 72‑GPU Helios rack‑scale AI platform at Computex, promising high‑density compute and improved efficiency for large‑scale models. 3 sources. https://clstr.news/cluster/supermicro-and-amd-unveil-72gpu-helios-rack-for-ai-at-computex ### 2026-05-20: AMD launches $3,999 Ryzen AI Halo mini‑PC for local AI development AMD's $3,999 Ryzen AI Halo mini‑PC, with a Zen 5 CPU, 128 GB memory and 50‑TOPS NPU, launches in June to run large AI models locally, challenging Nvidia's DGX Spark. 8 sources. https://clstr.news/cluster/amd-launches-3999-halo-ai-workstation-to-rival-nvidia-dgx-spark --- Cite as: AMD expands AI hardware across data centre and edge. CLSTR, https://clstr.news/situations/amd-ai-hardware-rollout