Reports
AI-generated structured vendor updates
NVIDIA's French AI Push: Open Models as a Trojan Horse for Hardware Lock-in
NVIDIA partners with French entities to deploy GB200, Blackwell B300, and Vera Rubin NVL72 systems, while promoting the Nemotron open model coalition. This builds an NVIDIA-centric AI infrastructure ecosystem in Europe, masking hardware lock-in with open model rhetoric.
NVIDIA and Coherent Scale 6-Inch InP Fab, Optical Interconnect Becomes AI Infrastructure's New Bottleneck Breaker
NVIDIA invests $2B and commits multi-billion purchases to Coherent's expanded 6-inch indium phosphide fab in Texas, scaling production of lasers and optical modules for AI interconnects. This addresses copper's distance and power limitations in large GPU clusters (e.g., Vera Rubin Ultra NVL576), pushing co-packaged optics into volume manufacturing.
AMD MLPerf 6.0: MI350 GPUs Achieve 3.5x Leap with MXFP4, Debut Multi-Node Training
AMD submitted its most comprehensive MLPerf Training 6.0 results, including first multi-node training (FLUX.1 on 512 GPUs) and MXFP4 training recipe. MI355X delivers 3.5x generational leap over MI300X on Llama 2-70B, within 5% of NVIDIA B200. 10 ecosystem partners validated reproducibility.
NVIDIA and HPE Expand AI Factory with Vera CPU for Agentic AI, Full-Stack Integration
NVIDIA and HPE expand the HPE AI Factory with the Vera CPU, the first CPU built for agentic AI, plus the NVIDIA Agent Toolkit, Confidential Computing, and full-stack NVIDIA integration (Spectrum-X, BlueField, ConnectX). This turnkey solution targets enterprise agentic AI production, locking customers into NVIDIA's hardware-software stack.
NVIDIA Blackwell Sweeps MLPerf: NVLink and NVFP4 Redefine AI Training Economics
NVIDIA Blackwell dominates MLPerf Training 6.0, submitting across all seven benchmarks including MoE workloads. GB300 NVL72 delivers up to 1.6x faster training than GB200, with fifth-gen NVLink unifying 72 GPUs as one giant GPU. NVFP4 low-precision training and massive scale (8,192 GPUs) set new industry standards.
NVIDIA Optimizes Google's DiffusionGemma for 1,000 tok/s Parallel Text Generation
NVIDIA optimizes Google DeepMind's DiffusionGemma, a diffusion-based text model generating 256 tokens per step in parallel. On a single H100, it achieves 1,000 tok/s, with deployment via NIM and NeMo. This breaks the sequential token bottleneck, slashing serving costs and latency for real-time AI.
Intel Reclaims AI Control Plane: Xeon 6+ and E835 Target Agentic Orchestration
Intel launches Xeon 6+ (288 E-cores on 18A), E835 200GbE controllers, and Crescent Island GPU. The strategy repositions the CPU as the control plane for agentic AI orchestration and data movement, while using E835 Ethernet to standardize AI data center networking.
Cisco-AMD Benchmark Shifts AI Fabric Control from GPU to SmartNIC and Switch
Cisco and AMD jointly release benchmarks for AI scale-out fabrics using N9000 800G switches, Pensando Pollara 400 smartNICs, and MI300X GPUs. IBPerf and MLPerf tests show P01/P99 bandwidth near 400Gbps line rate under incast congestion, proving deterministic performance that eliminates GPU stalls.
Intel at Computex 2026 Emphasizes CPU's Critical Role in AI Compute
Intel will outline its vision for the AI-driven computing era at Computex 2026, centering on the resurgence of the CPU as a critical AI engine. It emphasizes CPU-GPU/accelerator synergy to build efficient, scalable AI systems atop the broad x86 ecosystem.
Intel Collaborates with ChatPPT to Launch Hybrid AI PC Edition, Driving AI Workload Localization
Intel partnered with AI app ChatPPT to launch a hybrid AI PC edition using Intel's AI Super Builder technology. This version offloads certain AI workloads (e.g., formatting) from the cloud to the local PC, reducing cloud token costs by over 50%, boosting usage duration by 32%, and enhancing data privacy.
Intel and Google Deepen Collaboration to Define Core of Heterogeneous AI Infrastructure
Intel and Google announced a multiyear collaboration to advance next-generation AI and cloud infrastructure. The core is reinforcing the central role of CPUs and custom IPUs in heterogeneous AI systems, optimizing performance and efficiency through multi-generational Xeon processors, and expanding co-development of ASIC-based IPUs to improve efficiency and predictable performance at hyperscale.
Intel and Google Deepen Collaboration on CPU and IPU for Heterogeneous AI Infrastructure
Intel and Google announced a multi-year collaboration to advance next-generation AI and cloud infrastructure through aligned Xeon processor roadmaps and expanded co-development of custom ASIC-based IPUs. This reinforces the central role of CPUs in AI system orchestration and the critical value of IPUs in offloading infrastructure tasks to improve efficiency at hyperscale.
Intel and SambaNova Announce Heterogeneous Inference Architecture for Agentic AI
Intel and SambaNova have announced a collaborative blueprint for Agentic AI production workloads. The heterogeneous design combines GPUs, SambaNova RDUs, and Intel Xeon 6 processors to address performance, efficiency, and software compatibility issues, with availability expected in H2 2026.
AMD Announces Breakthrough MLPerf Inference 6.0 Results, Showcasing Multinode Scaling and Multimodal Capabilities
AMD's MLPerf Inference 6.0 submission, powered by Instinct MI355X GPUs, surpassed 1 million tokens per second for the first time on models like Llama 2 70B and GPT-OSS-120B. The results highlight efficient multinode scaling, rapid enablement of new workloads (e.g., text-to-video model Wan-2.2-t2v), and reproducible performance across a broad partner ecosystem.
AMD Achieves Breakthrough MLPerf Inference Results
AMD reports its Instinct MI300X accelerators achieved outstanding performance in MLPerf Inference 6.0 benchmarks, setting new records in natural language processing tasks. This demonstrates AMD's growing technical competitiveness in AI inference infrastructure.
Intel Demonstrates AI Performance with Xeon 6 and Arc Pro GPUs in MLPerf Inference
Intel showcased the performance of its Xeon 6 CPUs and Arc Pro B-Series GPUs in the MLPerf Inference v6.0 benchmarks, particularly in handling large language models (LLMs). The results indicate that a system with four Arc Pro B70 GPUs can process 120B parameter models, delivering up to 1.8x higher inference performance in multi-GPU setups.