News/NVIDIA
Nvidia unveils Rubin CPX GPU with 128GB memory for AI inference
Nvidia has unveiled the Rubin CPX GPU, a new compute-focused graphics processor featuring 128GB of GDDR7 memory specifically designed for enterprise AI inference workloads. The announcement positions Nvidia to address the growing demand for long-context AI applications in software development, research, and high-definition video generation, with shipments planned for late 2026. What you should know: The Rubin CPX represents Nvidia's first GPU to reach 128GB memory capacity, delivering up to 30 petaFLOPs of NVFP4 compute performance.• The GPU integrates hardware attention acceleration that Nvidia claims is three times faster than the GB300 NVL72.• Four NVENC and four NVDEC units are...
read Sep 5, 2025Nvidia fights GAIN AI Act that would prioritize US chip orders
Nvidia has criticized the proposed GAIN AI Act, arguing it would restrict global competition for advanced chips and harm U.S. economic leadership. The legislation, part of the National Defense Authorization Act, would require AI chipmakers to prioritize domestic orders before supplying foreign customers, mirroring restrictions from the AI Diffusion Rule that allocated computing power limits to different countries. What you should know: The GAIN AI Act would mandate that U.S. companies receive priority access to advanced AI chips before they can be exported internationally. Short for Guaranteeing Access and Innovation for National Artificial Intelligence Act, the legislation stipulates exporters must...
read Sep 5, 2025Broadcom scores $10B AI chip deal, shares surge 11% premarket
Broadcom shares surged 11% in premarket trading after the chipmaker announced a massive $10 billion AI chip order from a new customer, marking one of the largest deals in the company's history. The blockbuster agreement reinforces Broadcom's position as a leading custom chip provider as Big Tech companies seek alternatives to Nvidia's expensive and supply-constrained processors, while potentially adding over $160 billion to the company's $1.44 trillion market valuation. What you should know: The deal represents a significant milestone in Broadcom's AI strategy and could reshape competitive dynamics in the custom chip market. If premarket gains hold, Broadcom would cross...
read Sep 3, 2025Congress proposes US priority access to Nvidia’s AI chips
Congress is considering legislation that would force Nvidia and AMD to prioritize sales of their most powerful GPUs to US customers before exporting them to foreign markets, including China. The GAIN AI Act of 2025, now included in the Senate's draft of the National Defense Authorization Act of 2026, aims to address supply shortages that leave American businesses, startups, and universities waiting months for critical AI chips while these same processors are sold abroad. What you should know: The legislation would establish a "right-of-first-refusal" system for US customers seeking the most advanced GPUs needed for AI development. Nvidia and AMD...
read Sep 2, 2025Eggs, basket: Just 2 customers drive 39% of Nvidia’s record $46.7B quarterly sales
Nvidia disclosed that just two customers accounted for 39% of its record-breaking $46.7 billion second-quarter revenue, highlighting the chip giant's heavy dependence on a narrow customer base. The revelation underscores the concentration risk facing the AI chip leader, even as demand for its data center hardware continues to surge. What you should know: Nvidia's customer concentration extends beyond the top two buyers, creating significant revenue dependency on a small group. One customer alone represented 23% of Q2 revenue, while another accounted for 16%. Four additional customers each contributed between 10-14% of quarterly revenue. For the first half of fiscal 2026,...
read Aug 28, 2025Nvidia CEO says Trump talks on China chip sales will take a while
Nvidia CEO Jensen Huang revealed that discussions with the Trump White House about allowing sales of a less advanced version of the company's next-generation Blackwell AI chips to China "will take a while." The talks represent a potential shift in U.S.-China tech policy, as Nvidia seeks to maintain access to the Chinese market while navigating ongoing export restrictions on advanced semiconductor technology. What you should know: Huang confirmed that preliminary conversations with the White House have begun regarding Blackwell chip exports to China. The discussions focus on allowing sales of a "less advanced version" of Nvidia's next-generation Blackwell chips, suggesting...
read Aug 27, 2025Leveraged AI ETFs face $260B test as Nvidia earnings loom
The AI-driven boom in leveraged ETFs tied to individual stocks is set to face a critical test with Nvidia's Wednesday earnings report, as investors have poured billions into speculative products designed to amplify their AI-themed bets. This surge reflects unprecedented retail investor appetite for AI exposure, with 112 new leveraged ETFs launching in 2025 compared to just 38 in all of 2024, creating what analysts warn is an increasingly crowded and volatile market. The big picture: AI companies now dominate the leveraged ETF landscape, with more than half of all 190 single-stock leveraged and inverse ETFs in the U.S. connected...
read Aug 27, 2025China halts Nvidia H20 orders after U.S. “addiction” comments discovered
China has suspended new orders for Nvidia's H20 chips following inflammatory comments by U.S. Commerce Secretary Howard Lutnick, who publicly stated that America's strategy was to get Chinese developers "addicted" to weaker U.S. technology. The move threatens Nvidia's revenue stream from a market that represents at least 15% of the company's total sales and accelerates the tech decoupling between the world's two largest economies. What triggered the restrictions: Lutnick's mid-July 2025 televised remarks proved to be the catalyst for China's swift regulatory response. "We don't sell them our best stuff, not our second-best stuff, not even our third-best," Lutnick told...
read Aug 25, 2025AI and VR converge to reshape graphics at Vancouver’s Siggraph 2025
Siggraph 2025 in Vancouver showcased how artificial intelligence and extended reality (XR) are converging to reshape the graphics industry, with major announcements from Nvidia, Meta, and Arm demonstrating the future of neural rendering and immersive computing. The conference highlighted a fundamental shift toward AI-accelerated graphics processing, with new hardware and software solutions designed to make high-fidelity rendering more accessible and efficient across platforms. What you should know: Nvidia unveiled new Blackwell-powered RTX PRO servers and workstation GPUs specifically designed for professional AI and rendering workloads. The RTX PRO 6000 servers deliver 4x improvements in real-time rendering FPS and 6x better...
read Aug 25, 2025Nvidia’s Jetson Thor delivers 7.5x AI boost for robotics at $3.5K
Nvidia has launched its Jetson AGX Thor developer kit and production modules, a next-generation robotics platform powered by Blackwell GPUs designed to enable millions of robots, including humanoids, with advanced AI capabilities. The platform delivers record-breaking performance and efficiency, reinforcing Nvidia's position in AI infrastructure as analysts project stronger earnings driven by its GB200 and Blackwell product ramps ahead of the company's August 27 earnings report. Key performance specs: The Blackwell GPU-powered platform delivers up to 2,070 FP4 teraflops of AI compute with 128GB of memory in a 130-watt power envelope. This represents 7.5 times the AI compute and 3.5...
read Aug 20, 2025Nvidia stock drops 5% as AI sector selloff rattles investors
Nvidia's stock plunged more than 3% in early trading Wednesday, extending its weekly decline to over 5%, while other AI industry leaders including Google, Meta, Tesla, and Palantir also saw significant drops. The selloff has rattled investors across the AI sector, with analysts struggling to pinpoint specific causes behind the sudden retreat from what had been a meteoric rise in AI-related stocks. What you should know: The decline comes despite Nvidia's position as the dominant supplier of AI hardware, making its struggles particularly concerning for the broader industry. Nvidia has been the primary beneficiary of the AI boom, selling the...
read Aug 20, 2025Canada builds sovereign AI infrastructure to keep data cozy, warm and at home
Canada is making a bold play for AI independence. Buzz High Performance Computing (Buzz HPC), a subsidiary of publicly-traded Hive Digital Technologies, has partnered with Bell Canada to build nationwide AI infrastructure that keeps sensitive data entirely within Canadian borders—a concept known as "sovereign AI." This partnership addresses a growing concern among governments and enterprises: maintaining control over their most sensitive data while accessing cutting-edge AI capabilities. Rather than relying on foreign cloud providers, Canadian organizations will soon have access to powerful AI computing resources hosted exclusively in Canadian-owned facilities. The infrastructure backbone The collaboration centers on Bell AI Fabric,...
read Aug 19, 2025Nvidia develops new B30A chip for China with half the flagship’s power
Nvidia is developing a new AI chip for China based on its latest Blackwell architecture that will outperform the H20 model currently sold there, according to sources familiar with the matter. The move comes as President Trump recently opened the door to allowing more advanced Nvidia chips in China, though regulatory approval remains uncertain amid ongoing tensions over AI technology access. What you should know: The new chip, tentatively called the B30A, will use a single-die design delivering roughly half the computing power of Nvidia's flagship B300 accelerator. A single-die design places all main integrated circuit components on one continuous...
read Aug 19, 2025Nvidia’s 9B parameter AI model offers toggleable reasoning on single GPU
Nvidia has released Nemotron-Nano-9B-v2, a compact 9-billion parameter language model that features toggleable AI reasoning capabilities and achieves top performance in its class on key benchmarks. The model represents Nvidia's entry into the competitive small language model market, offering enterprises a balance between computational efficiency and advanced reasoning capabilities that can run on a single GPU. What you should know: Nemotron-Nano-9B-v2 combines hybrid architecture with user-controllable reasoning to deliver enterprise-ready AI at reduced computational costs. The model was pruned from 12 billion to 9 billion parameters specifically to fit on a single Nvidia A10 GPU, making deployment more accessible for...
read Aug 15, 2025NVIDIA releases 1M-hour speech dataset for 25 European languages
NVIDIA has released Granary, an open-source multilingual speech dataset containing approximately one million hours of audio, alongside two new AI models designed for transcription and translation across 25 European languages. The release addresses a critical gap in speech AI development, as only a tiny fraction of the world's 7,000 languages are currently supported by AI language models, with particular focus on underrepresented European languages like Croatian, Estonian, and Maltese. What you should know: The Granary dataset represents a massive leap forward in multilingual speech AI training data, providing developers with ready-to-use resources for production-scale applications. The dataset includes nearly 650,000...
read Aug 14, 2025Ai2 secures $152M from NSF and NVIDIA for open scientific AI research
Ai2, a Seattle-based nonprofit AI research institute, has secured $152 million in combined funding from the National Science Foundation ($75 million) and NVIDIA ($77 million) to build a national-level open AI ecosystem for scientific research. The partnership will establish the Open Multimodal AI Infrastructure to Accelerate Science (OMAI) project, positioning Ai2 to advance both AI-driven scientific discovery and the fundamental science of AI itself through fully transparent, reproducible models. What you should know: The OMAI project represents a major federal investment in open-source AI infrastructure specifically designed for scientific applications. Led by Dr. Noah A. Smith, Senior Director of NLP...
read Aug 13, 2025AI2’s MolmoAct 7B enables robots to think in 3D space, challenging rivals like Nvidia
The Allen Institute for AI (AI2) has released MolmoAct 7B, an open-source robotics AI model that enables robots to "reason in space" and "think" in three dimensions. This Action Reasoning Model challenges existing offerings from tech giants like Nvidia and Google by providing robots with enhanced spatial understanding capabilities, achieving a 72.1% task success rate in benchmarking tests that outperformed models from Google, Microsoft, and Nvidia. What makes it different: MolmoAct represents a significant departure from traditional vision-language-action (VLA) models by incorporating genuine 3D spatial reasoning capabilities. "MolmoAct has reasoning in 3D space capabilities versus traditional vision-language-action (VLA) models," AI2...
read Aug 13, 2025Nvidia unveils Cosmos Reason model to help robots think before acting
Nvidia has unveiled new agentic AI and physical robotics models at the Siggraph 2025 conference, including the Cosmos Reason model, updated Omniverse libraries, and the RTX Pro 6000 Blackwell Server Edition GPU. These advances target digital twins, physical AI systems, and enterprise computing, positioning Nvidia to capture growing demand for AI that can reason and act in the physical world. What you should know: The Cosmos Reason model represents a significant leap in physical AI capabilities, designed to help robots and vision agents think before they act. The customizable 7-billion-parameter vision-language model enables machines to "reason like humans" using prior...
read Aug 12, 2025AI cloud provider CoreWeave reports $1.21B revenue, $30.1B backlog
CoreWeave, an Nvidia-backed AI cloud computing company, exceeded Wall Street revenue expectations for the second quarter, reporting $1.21 billion against analyst estimates of $1.08 billion. The strong performance reflects surging demand for GPU-based infrastructure needed to train and run large AI models. What you should know: CoreWeave operates exclusively in GPU-based cloud computing, providing access to Nvidia chips that are essential for AI development. The company currently operates 33 AI data centers across the United States and Europe, focusing solely on GPU infrastructure rather than traditional CPU-based cloud services. Revenue backlog reached $30.1 billion as of June 30, indicating strong...
read Aug 12, 2025Trump weighs allowing downgraded Nvidia Blackwell chips to China
President Donald Trump has signaled openness to allowing Nvidia to sell a downgraded version of its most advanced Blackwell AI chip to China, potentially reducing performance by 30-50%. This development comes as part of ongoing negotiations over semiconductor exports, with Nvidia and AMD already agreeing to pay the U.S. government a 15% revenue cut from Chinese chip sales in exchange for export licenses. What you should know: Trump indicated he would consider approving a "somewhat enhanced — in a negative way" Blackwell processor for Chinese markets. The president said he plans to meet with Nvidia CEO Jensen Huang specifically about...
read Aug 11, 2025Amazon uses AI and synthetic data to boost “zero-touch manufacturing”
Amazon Devices & Services has deployed a groundbreaking AI-powered manufacturing solution at one of its facilities that enables robotic arms to autonomously audit devices and integrate new products into production lines using only synthetic data. The technology represents a major advancement toward "zero-touch manufacturing," where robots can handle diverse products without requiring physical prototypes or hardware changes, significantly accelerating production timelines and reducing costs. How it works: The solution combines Amazon's custom software with NVIDIA digital twin technologies to create photorealistic, physics-enabled simulations of devices and factory workstations. NVIDIA Isaac Sim generates over 50,000 synthetic images from CAD (computer-aided design)...
read Aug 11, 2025US takes 15% cut from Nvidia and AMD China chip sales under new deal
Nvidia and AMD have agreed to pay the U.S. government 15% of revenues from sales of advanced AI chips to China, including Nvidia's H20 and AMD's MI308 processors, according to a U.S. official. The arrangement serves as a condition for obtaining export licenses after the Trump administration halted such sales in April, raising questions about whether chip exports to China constitute a national security risk or simply a revenue opportunity. What you should know: The revenue-sharing agreement applies specifically to advanced AI chips that were previously restricted from Chinese markets. Nvidia's H20 chips, described by Commerce Secretary Howard Lutnick as...
read Aug 8, 2025Deutsche Telekom plans AI data center with Nvidia for 2026 launch
Deutsche Telekom plans to begin construction of a large-scale AI data center in 2026, developed in partnership with Nvidia and investment firm Brookfield as part of what the companies describe as the world's first industrial AI cloud for European manufacturers. The facility represents a significant push into AI infrastructure by the German telecom giant, targeting the growing demand for specialized computing power among manufacturing companies across Germany and mainland Europe. What you should know: Deutsche Telekom is targeting a 2026 construction start for its AI "Gigafactory" using Nvidia's advanced chip technology. CEO Tim Höttges confirmed the timeline during an investor...
read Aug 8, 2025Tesla shuts down Dojo supercomputer team, focuses on AI5 and AI6 chips
Tesla has shut down its Dojo supercomputer team and reassigned workers to other projects, with team leader Peter Bannon departing the company. The move reflects CEO Elon Musk's strategic pivot toward consolidating AI chip development rather than dividing resources across multiple custom silicon projects, as the company increasingly relies on external partners like Nvidia and Samsung for its AI computing needs. The big picture: Tesla is streamlining its AI strategy by abandoning its custom Dojo supercomputer in favor of focusing development efforts on its next-generation AI5 and AI6 chips.• Musk explained on X that "it didn't make sense for Tesla...
read