Reports
AI-generated structured vendor updates
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.
Z.ai GLM-5.2 Ships Usable 1M-Token Context, No Benchmarks, Two Thinking Levels
Z.ai releases GLM-5.2 with a claim of usable 1M-token context and two thinking-effort levels. No standard benchmarks are provided, raising concerns about real-world performance. The model targets replacing chunking-based RAG with native long-context reasoning.
NVIDIA AgentPerf Benchmark: Blackwell Ultra Delivers 20x More Agents per Megawatt vs Hopper
NVIDIA and Artificial Analysis unveil AgentPerf, the first benchmark for agentic AI workloads. Results show the GB300 NVL72 platform delivers up to 20x more concurrent agents per megawatt than the HGX H200 when running DeepSeek V4 Pro, using real coding agent trajectories to measure throughput and responsiveness.
NVIDIA and SK Hynix Lock Down HBM4/5 Roadmap, Cementing Vera Rubin Supply Chain
NVIDIA and SK Hynix sign a multi-year agreement to co-define HBM4 production and HBM5 pre-research for Vera Rubin GPUs. Samsung also enters HBM4 supply as a second source. The deal elevates SK Hynix from vendor to co-developer, potentially creating a de facto memory standard barrier that marginalizes Micron and others.
NVIDIA's UK Sovereign AI Play: From Chip Vendor to National Infrastructure Controller
NVIDIA partners with the UK government to deploy sovereign AI infrastructure via Isambard-AI (5,400 GH200 superchips) and the Sovereign AI Fund, backing local startups. This move establishes a national AI control plane, locking compute into NVIDIA's ecosystem and bypassing traditional hyperscalers like AWS and Azure.
NVIDIA Nemotron 3 Ultra: A MoE-Based Control Plane for Cost-Efficient AI Agent Orchestration
NVIDIA launches Nemotron 3 Ultra, a 550B-parameter MoE model (55B active) purpose-built for AI agent orchestration. Featuring Multi-Teacher On-Policy Distillation (MOPD) and a Hybrid Mamba-Transformer architecture, it achieves 5x throughput and 30% cost savings on tasks like SWE-bench, signaling a shift of reasoning control to a layered agent system.
NVIDIA DGX Spark Update: One-Click Local AI Agents, Multi-Node Cluster for 400B Models
At Computex 2026, NVIDIA updates DGX Spark with NemoClaw for one-click local AI agent setup, 2.6x throughput boost for Qwen3.6-35B via vLLM optimizations, and Sync cluster assistant to connect 2-4 nodes over ConnectX-7 200Gbps RoCE, enabling local deployment of large models and multi-agent pipelines.
AMD Backs SPEC CPU 2026 Benchmark, Emphasizing Open, Trusted Performance Measurement
AMD published a blog endorsing the upcoming SPEC CPU 2026 industry benchmark, emphasizing the critical role of open, reproducible CPU performance standards for customer infrastructure decisions in the AI era. The new benchmark updates its application suite and strengthens support for bare-metal cloud environments and parallel computing.
Google Launches Gemma 4 Open Models, Accelerating Local AI Agent Deployment
Google released the Gemma 4 open model family under Apache 2.0 license, introducing MoE architecture for the first time. It aims to deliver high-performance AI agent capabilities directly to mobile and edge hardware, reducing reliance on cloud clusters and enabling new local, private AI applications.
AMD and OpenAI Contribute MRC Protocol to OCP for Scalable AI Networking
AMD, in collaboration with OpenAI, Microsoft, and others, contributed the MRC (Multipath Reliable Connection) protocol, designed for large-scale AI training, to the Open Compute Project (OCP). AMD co-authored the specification and has already deployed MRC on its programmable Pensando DPU/NIC products, positioning its networking technology as a key enabler for resilient and adaptive AI infrastructure.
AMD and OpenAI Introduce MRC, a Next-Gen Transport Protocol for AI Training
AMD, in collaboration with OpenAI, Microsoft, and other industry leaders, has released the specification for the Multipath Reliable Connection (MRC) protocol. MRC addresses performance bottlenecks of RoCEv2 in hyperscale AI training clusters through intelligent packet spraying, selective retransmission, and network-signaled congestion control, aiming to improve bandwidth utilization and job resilience.
NVIDIA Extreme Co-Design: Vera Rubin Platform Targets Agentic Inference TCO Inflection
NVIDIA unveils an extreme co-design stack for agentic systems, featuring Vera Rubin NVL72, NVLink 6, ConnectX-9, BlueField-4, and Spectrum-X. By disaggregating inference, optimizing KV cache management, and deploying low-latency fabrics, it aims to break the throughput-interactivity tradeoff, making high-context token processing economically viable.
AMD Showcases Heterogeneous Computing Strategy for Enterprise AI with Dell
At Dell Technologies World, AMD highlighted its heterogeneous computing portfolio, aiming to match the right compute engine to specific enterprise AI workloads, while emphasizing hardware-based security and manageability. This signals a shift in AI infrastructure from generic solutions to fine-tuned, scenario-specific deployments.
AMD Proposes New AI Infrastructure Networking Paradigm: From Lossless Fabrics to Intelligent Endpoints
AMD published a blog outlining seven key questions for building large-scale AI infrastructure, arguing that traditional lossless Ethernet or InfiniBand architectures face cost and complexity bottlenecks. It advocates shifting network intelligence and reliability functions from expensive, specialized switches to intelligent NICs, enabling reliable transport over standard (potentially lossy) Ethernet to reduce TCO and simplify operations.
NVIDIA Launches Nemotron 3 Nano Omni, Targeting AI Agent Perception Layer
NVIDIA released the open-source multimodal model Nemotron 3 Nano Omni, featuring a 30B-A3B hybrid MoE architecture. It unifies vision, audio, and language processing into a single model, designed to act as the 'eyes and ears' for AI agents. It claims to eliminate latency and context fragmentation from multi-model collaboration, achieving up to 9x higher throughput while maintaining interactivity, thereby reducing AI agent deployment and inference costs.
Apple-Google Multi-Year Partnership Confirmed: Gemini to Power New Siri
Apple and Google confirm multi-year partnership with Google Cloud as preferred provider. Google is building a custom 1.2 trillion parameter Gemini model for Apple, 8x Apple's current cloud model. Siri will gain Gemini capabilities in 2026 with iOS 27. Privacy architecture unchanged—Gemini runs on Apple-controlled servers with data protection guarantees. Device compatibility limits exclude hundreds of millions of older iPhone users.
NVIDIA Shifts AI Infrastructure Metric from FLOPS to Cost Per Token
NVIDIA advocates for "cost per token" as the primary economic metric for AI infrastructure, replacing "FLOPS per dollar." This shift moves the focus from computational inputs to business outputs, requiring full-stack optimization across hardware, software, and networking to lower enterprise AI inference TCO.
Google Launches Gemma 4 Open Models, Targeting Edge Inference and AI Agent Architecture
Google introduces the Gemma 4 open model family, with four sizes from 2B to 31B parameters, emphasizing breakthrough intelligence-per-parameter and native support for agentic workflows, multimodality, and long context. The small models are engineered for edge devices, aiming to bring frontier reasoning to mobile and IoT scenarios.
Google Launches Gemma 4 Open Model Family
Google introduces Gemma 4 open model family with four size variants, optimized for edge and mobile devices. The series supports multimodal processing, long context windows and 140+ languages under Apache 2.0 license.
Nvidia Launches Nemotron 3 Super for Agentic AI Inference Optimization
Nvidia releases Nemotron 3 Super, a 120B parameter model with hybrid MoE architecture combining Mamba and Transformer layers, delivering 5x throughput improvement. Designed for multi-agent workflows with 1M token context window to prevent task drift. Open weights and cloud deployment lower enterprise adoption barriers.