Reports
AI-generated structured vendor updates
NVIDIA Launches Nemotron 3.5 Lightning and NeMo Switchyard to Fortify Local AI Control Plane
NVIDIA introduces Nemotron 3.5 Lightning, an open-weight 30B MoE model with 4x faster token generation, and NeMo Switchyard, an open-source routing library that dynamically selects models to optimize cost and performance for local agentic AI on RTX PCs and DGX systems.
NVIDIA Reveals Vera Rubin NVL72 Specs, GA Slips to H2 2027
NVIDIA detailed the Vera Rubin NVL72 rack-scale platform, pushing GA to H2 2027 due to packaging transition and Vera CPU yield. The platform offers 72-GPU coherent domain, setting new benchmarks in performance per watt and token cost, but the slip extends Blackwell's lifecycle and impacts deployment timelines.
NVIDIA Employee Detained in Taiwan B300 Smuggling Probe, Export Control Reaches Chipmaker
Taiwan prosecutors detain a NVIDIA employee for allegedly helping smuggle Super Micro servers with **GB300** chips to China. Seven people are held, involving $21 million. This marks the first time export control enforcement directly hits a chipmaker, shifting compliance risk upstream.
DeepSeek Claims China Will Phase Out NVIDIA in One Year with Huawei Ascend 950 and Self-Developed Tile Language
DeepSeek founder claims China will phase out NVIDIA within a year, with Huawei Ascend 950 SuperPod matching GB200/GB300 performance. DeepSeek is developing its own inference chips and Tile Language to reduce CUDA dependency, marking a critical shift from import substitution to an independent AI compute ecosystem.
NVIDIA Vera Rubin NVL72 Enters Production, 10x Energy Efficiency Reshapes AI Infrastructure
NVIDIA has announced full production of the Vera Rubin NVL72 platform, with first shipments to CoreWeave, Microsoft, Amazon, and Oracle. Each rack integrates 72 Rubin GPUs and 36 Vera CPUs, delivering 10x improvement in energy efficiency and inference cost over Blackwell for agentic AI workloads, marking a new era of rack-scale AI infrastructure.
NVIDIA and Wistron Launch US-Based GB300 Production, Shifting AI Manufacturing Landscape
NVIDIA partnered with Wistron to open a $700M factory in Texas, achieving L6 integration of the GB300 Grace Blackwell Ultra system. The first US-made unit contains 1.5M parts, weighs 2 tons, and costs $4M. This milestone accelerates US AI capex localization from 5% to 30%, reshaping the global AI supply chain.
Tiered AI Chip Market Emerges as US Allows H200 Exports to China with 25% Levy
The US Commerce Department approved NVIDIA H200 exports to China with a 25% sales tax, while maintaining a ban on Blackwell. This formalizes a tiered AI chip market, making H200 the best available imported chip for China, but the performance gap and added tax burden increase deployment costs and complexity for Chinese AI infrastructure.
NVIDIA DGX GB300超级计算机在海军研究生院上线
...
NVIDIA-OpenAI $100B Partnership: 10GW Vera Rubin AI Factories Reshape Ecosystem
NVIDIA and OpenAI announce a strategic partnership to deploy at least 10GW of NVIDIA systems using the Vera Rubin platform (Rubin GPU, Vera CPU, HBM4, NVLink 6). NVIDIA will invest up to $100B. First facilities go online in H2 2026, powering OpenAI's next-gen models, marking the era of multi-GW AI factories.
NVIDIA公开Rubin GPU架构细节：3360亿晶体管，智能体AI性能较Blackwell提升10倍
...
NVIDIA Reveals Vera Rubin GPU and Vera CPU: 3360B Transistors, 88-Core Olympus, 10x Agentic AI Efficiency
NVIDIA fully discloses Vera Rubin GPU and Vera CPU specifications. The GPU features 3360B transistors, HBM4 288GB, and 10x agentic AI efficiency over Blackwell. The CPU has 88 custom Olympus cores, delivering 2.2x faster agentic AI performance than Intel Sapphire Rapids. This solidifies NVIDIA's full-stack strategy against x86 incumbents.
NVIDIA and Wistron Open US Factory for GB300 and Vera Rubin AI Superchips
Wistron opens its first US manufacturing facility in Fort Worth, producing NVIDIA GB300 Grace Blackwell Ultra and Vera Rubin superchips. The $700M plant aims for tens of thousands of boards monthly, marking NVIDIA's strategic shift to domestic AI hardware production.
NVIDIA Invests $2B in CoreWeave, Debuts 'Compute Central Bank' Model
NVIDIA invests $2 billion in CoreWeave and launches the AI Compute Partner Program, featuring credit enhancement, revenue sharing, and GPU buyback. This transforms NVIDIA from a hardware vendor into a 'compute central bank', tightening control over the AI cloud leasing ecosystem and squeezing intermediaries.
Moonshot AI Launches Kimi K3: 2.8T Parameter Open-Source MoE Model at $3/$15
Moonshot AI unveils Kimi K3, a 2.8T parameter open-source MoE model with 896 experts (16 active), native vision, and 1M context. API pricing at $3/$15 input/output per million tokens undercuts rivals. Open weights release July 27. GPU capacity exhausted within 2 days. Arena score 1679 tops Fable 5.
NVIDIA Agent Toolkit Shifts AI Agent Control from Cloud to Local DGX Station
NVIDIA launches Agent Toolkit for DGX Station, comprising NemoClaw, Nemotron 3 Ultra, Omniverse Libraries, and OpenShell. It enables local AI agent deployment in 30 minutes, locking developers into NVIDIA's hardware-software stack and shifting control from cloud services to on-premises hardware.
测试来源：36氪
...
AMD与OpenAI达成6GW算力供应历史性协议 1.6亿认股权证可获10%股权 股价盘前涨35%
...
NVIDIA AI Compute Partnership: Revenue Share and Credit Backstop to Lock Cloud Providers into DSX AI Factories
NVIDIA launches AI Compute Partnership with revenue sharing and credit backstop, shifting from hardware sales to recurring service revenue. Initial projects include 40K GB300 chips for Sharon AI and 170K GPUs for Firmus, totaling 200K+ high-end chips. NVIDIA is becoming the 'central bank' of AI compute, squeezing cloud brokers.
Anthropic Claude Goes Exclusive on Azure, Microsoft Locks AI Model Distribution via GB300
Anthropic's Claude models are now generally available on Azure Foundry, powered by NVIDIA GB300 NVL72 clusters with over 4600 Blackwell Ultra GPUs. Initial models include Opus 4.8 and Haiku 4.5 with prompt caching and extended thinking. Microsoft gains exclusive enterprise distribution, strengthening its competitive position against AWS and Google Cloud.
NVIDIA and AWS Default GPU Vector Search with cuVS, G7 Instances Deliver 4.6x Inference
NVIDIA and AWS collaborate to embed cuVS as default GPU-accelerated vector search in OpenSearch Serverless, delivering 10x faster indexing at 1/4 cost. New EC2 G7 instances with RTX PRO 4500 Blackwell GPUs achieve up to 4.6x inference performance. AWS achieves GB300 Exemplar Cloud status for training.