Hyperscalers and AI-focused infrastructure operators are increasingly developing proprietary AI accelerators and XPUs to keep pace with large language models and complex reasoning workloads. To address the physical, memory, and networking challenges of deploying proprietary silicon at scale, NVIDIA has detailed its NVLink Fusion interconnect and NVHBM custom base-die memory architecture. The unified portfolio is

Amazon Web Services and NVIDIA have announced a major expansion of their joint infrastructure initiatives, outlining plans to deploy an additional 2 million NVIDIA GPUs across AWS data centers between 2027 and 2028. Building on a 16-year engineering relationship, the expanded collaboration targets end-to-end scaling of AI infrastructure, encompassing accelerator capacity, host CPU architectures, high-bandwidth

AMD announced that its upcoming Versal adaptive SoCs will natively support Universal Chiplet Interconnect Express (UCIe) 1.1 interfaces. The open industry standard, which Intel is also building into Diamond Rapids and Wildcat Lake, establishes low-power, high-bandwidth die-to-die communication, allowing co-packaged AMD adaptive SoCs to interface directly with specialized silicon. This transition targets the engineering constraints

Nutanix has announced the general availability of Nutanix Enterprise AI (NAI) 2.8 alongside the upcoming release of Nutanix Kubernetes Platform (NKP) 2.19. The updates focus on bridging the gap between legacy virtualized environments and modern containerized AI workloads. By implementing a dual-native architecture, Nutanix aims to allow organizations to run virtual machines and containerized AI

Cerebras Systems has officially introduced the CS-4, the fourth generation of its wafer-scale AI supercomputing platform, designed for frontier training and high-concurrency inference workloads. The system is engineered around three newly developed Wafer Scale Engine 3 (WSE-3) Turbo processors, housed in a ground-up redesign called the Nexus rack-scale platform. Cerebras claims the system delivers up

Proxmox rolled out version 9.2 of its virtualization platform at the start of this month, which adds official support for 64-bit ARM (arm64/aarch64). Previously, Proxmox VE was available only for x86-64 (amd64). With this release, arm64 builds share the same code base, package repositories, and release lifecycle as x86-64. Proxmox VE 9.2 for arm64 is

Lancium has announced a strategic collaboration and investment from NVIDIA to accelerate the deployment of large-scale accelerated computing infrastructure across Lancium’s data center portfolio. Lancium, a Blackstone portfolio company supported by Blackstone Energy Transition Partners and Blackstone Multi-Asset Investing, will integrate NVIDIA’s full-stack hardware and software architectures directly into its power-ready campus footprint. The collaboration

Alongside its Groq 3 LPX news at Hot Chips 2026, NVIDIA announced that SpaceXAI will deploy NVIDIA Vera CPUs to accelerate its next generation of agentic AI applications. The company also plans to expand the infrastructure behind Grok on the Vera Rubin platform as it scales toward gigawatts of capacity, and to extend that architecture

NVIDIA announced the production release of the NVIDIA Groq 3 LPX, a dedicated interactive inference accelerator built to complement the Vera Rubin NVL72 platform. As enterprise AI workloads shift from raw model pre-training toward real-time reasoning and autonomous agent orchestration, infrastructure requirements are changing. Agentic workflows require iterative, multi-step execution where latency during token generation