NVIDIA is a semiconductor company best known for its GPUs, which are the primary compute hardware for training and running AI models. Its H100 and A100 chips power the majority of AI data centers worldwide.
Madison Huang, a senior director at Nvidia and daughter of CEO Jensen Huang, will visit LG Electronics' robotics hub in Korea on Tuesday for closed-door talks on AI infrastructure and robotics partner…
China's top-ranked hedge fund managers, including Perseverance Asset Management International, cut holdings of Nvidia and US hyperscalers in the second quarter, with Perseverance slashing its Nvidia p…
Satellite operators Lynk and Omnispace completed their merger on August 14 to create Elveo Mobile, a global direct-to-device (D2D) player led by CEO Ramu Potarazu, who said the company has 56 contract…
A Guardian investigation found that Microsoft's AI infrastructure claims exceed its actual deployed chip count, with thousands of advanced AI chips sitting in inventory because data center constructio…
A Reddit user proposes specializing small language models (sub-16B) for dedicated tasks like math, medicine, and writing, orchestrated by a general Mixture-of-Experts model, to reduce cost and complex…
A user who purchased a CMP 170HX mining GPU for AI workloads reports that it performs comparably to a pair of Radeon R9700 GPUs in decoding but has significantly slower prefill performance, and it thr…
Meta Platforms Inc. released Muse Glimmer, an open-source version of its most advanced AI model Muse Spark, designed to run locally on consumer hardware with a single GPU, marking a return to its open…
NxtGen, an Indian sovereign cloud and AI infrastructure provider, launched SpeedCloud Global on August 17, 2026, to offer managed cloud and AI services overseas, claiming 39-48% lower costs than leadi…
Nvidia Corp. is reducing its financial backing for OpenAI, declining to guarantee as much of the AI company's debt amid fears of a compute oversupply and a shift toward more efficient AI workflows. Th…
A practical framework for choosing a computer for local AI in 2026 emphasizes that hardware selection depends on model size, quantization, and VRAM or unified memory, not just brand. The guide, based …
Ukraine's military intelligence agency HUR reported that Nvidia chips are being found in Russian missiles, indicating the integration of AI at the edge for real-time object detection and autonomous na…
Ukraine's military intelligence (HUR) said Wednesday, Aug. 12, that it found an Nvidia Jetson Orin computer module inside Russia's new S-71 Monochrome air-launched cruise missile, potentially indicati…
Mark Cuban said on Aug. 16 that 'chips as an asset class will be the new crypto,' a broad investment thesis as AI drives demand for advanced computing hardware. CoreWeave closed a $2.6 billion delayed…
Nvidia has spent over $40 billion on equity stakes and supply-chain investments in companies that buy its chips, including $30 billion into OpenAI, $21 billion into SpaceX, and $30 billion into Intel,…
Nvidia has dramatically reduced the amount of OpenAI infrastructure financing it may guarantee, scaling back from a previously reported $250 billion commitment, according to a Wall Street Journal repo…
Nvidia announced agreements with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to mobilize over $500 billion for AI infrastructure financing, creating collateralized compute as a n…
Nvidia chips are being found in Russian missile guidance systems, enabling AI-driven targeting with on-board computer vision and sensor fusion. The chips, likely Nvidia Jetson or similar edge processo…
Renting Nvidia's H100 GPU climbed roughly 40% between October 2025 and March 2026, with average rates jumping from $1.70 to $2.35 per GPU-hour, according to data from SemiAnalysis. On-demand H100 capa…
China's Cyberspace Administration (CAC) issued guidance in mid-September 2025 barring major tech firms like ByteDance and Alibaba from purchasing Nvidia AI chips, including the RTX Pro 6000D, pushing …
OpenAI on August 13 unveiled a limited preview of Ultrafast mode for its flagship GPT-5.6 Sol model, running on Cerebras wafer-scale inference chips instead of Nvidia GPUs, delivering up to 750 output…