NVIDIA
NVIDIA designs the GPUs that train and run virtually every major AI model, from the H100 to the Blackwell generation. Its CUDA software stack and data-center chips made it one of the most valuable companies on Earth and the bottleneck of the AI boom. All our NVIDIA coverage: chips, earnings and ecosystem news.
habrGPT: training a 0.5B LLM from scratch on Habr articles — Karpathy’s nanochat and an fp8 vs bf16 comparison
Nvidia pairs GPUs with startup d-Matrix's inference chips instead of competing with it
NVIDIA released Audex-30B — a unified audio-text MoE model based on Nemotron-Cascade-2
Hugging Face: transformers backend in vLLM now runs at native speed
Hugging Face announced performance parity: the `--model-impl transformers` flag delivers native vLLM speed across 450+ architectures without rewriting code.
AI chip startup SambaNova Systems raises $1 billion at an $11 billion valuation
AI chipmaker SambaNova Systems has closed a $1 billion funding round at an $11 billion valuation as investors bet on long-term growth in demand for specialized AI…
China's DeepSeek secretly develops its own AI inference accelerator
Reuters reports that DeepSeek has been secretly designing its own AI inference accelerator for about a year — the company is in talks with partners and hiring chip…
Amazon raises at least $25 billion for AI infrastructure through bond issuance
Chinese Companies Ready to Allocate Up to 46% of AI Chip Budget to Local Suppliers Instead of Nvidia
AI Data Center Projects Collapse Worldwide Amid Energy Shortages and Protests
NVIDIA Vera CPU Boosts Throughput of AI Factories for Agentic Workloads
NVIDIA Dynamo vs Triton Inference Server: what changed for agent inference
NVIDIA released open-source framework Dynamo for agentic LLM inference with KV-cache routing — unlike Triton, not designed for parallel agent calls.
Fable creates first GPU megakernel — experts discuss AI self-improvement loop
Fable created the first true GPU megakernel; Anthropic co-founder Jack Clark called it 'the beginning of a recursive self-improvement loop' — a step toward…
Inside Chinese AI: ByteDance, Z.ai, humanoid robots requiring passport ID and the server GPU card market
Chipmaker Stocks Fall Amid Doubts About AI Spending Sustainability
South Korea to Create Future Fund From Semiconductor Boom Tax Revenue
LeRobot v0.6.0 from Hugging Face: robots learn to predict the future and evaluate themselves
Shampoo and Cookies: Procter & Gamble Transitions to AI in Product Development
Procter & Gamble and other major manufacturers are restructuring their R&D laboratories: algorithms now participate in creating shampoo formulas, cookie recipes…
Meta Decided to Sell AI Computing Power to Third Parties — Asian Chipmakers' Stocks Crashed
Meta's announcement of plans to monetize AI infrastructure as a commercial service triggered a sell-off of semiconductor stocks in Asia amid fears of market…