A developer found that `nvidia-smi` reported 97% GPU utilization on an H100 cluster while actual training throughput was less than half of expected benchmarks. Tracing via eBPF revealed the GPU was id…
NVIDIA's debt-to-equity ratio and other financial metrics are being compared against peers in the semiconductors and semiconductor equipment industry as part of a competitive analysis. The evaluation …
3code, a new coding agent designed for efficiency, is now available for free use even without a subscription. The tool, which integrates with multiple open-weight model providers like Nvidia and DeepI…
AI token prices are poised to drop sharply as new models and infrastructure expand supply and lower inference costs, according to a Business Insider report citing an unnamed AI infrastructure CEO who …
A developer has built a local AI invoice OCR system that reads both digital and scanned PDFs entirely on-device, using pdfjs for text extraction and Qwen2.5 1.5B via llama.cpp for structured JSON outp…
NVIDIA CEO Jensen Huang publicly criticized corporate leaders who attribute workforce reductions to artificial intelligence, calling that explanation "too lazy" in an interview with Singapore broadcas…
A developer successfully deployed Hermes Agent on an NVIDIA Jetson Orin Nano (8 GB) for self-hosted AI agent runtime, achieving a silent, low-power automation box running 24/7. The setup uses Ubuntu 2…
NEURA Robotics closed a $1.4 billion Series C, the largest full-stack robotics funding round in history, as the race to control the Physical AI ecosystem intensifies. Sharpa integrated its Wave tactil…
KKR launched Helix Digital Infrastructure, a new platform backed by more than $10 billion in capital commitments, and named former Amazon Web Services CEO Adam Selipsky as co-founder and CEO. The vent…
A comparison of three local large language model runtimes reveals that llama.cpp is the core inference engine, while Ollama and LM Studio are user-friendly wrappers built on top of it. Ollama offers a…
Apple debuted the third generation of Apple Foundation Models (AFM) on June 8, 2026, unveiling a family of five models spanning on-device and server deployments, including the 20-billion-parameter, na…
The MADSys Lab at Tsinghua University’s KTransformers project enables frontier-class MoE models like DeepSeek-R1 671B to run on commodity hardware with a CPU-GPU hybrid inference stack, achieving 286 …
Apple expanded its Private Cloud Compute (PCC) to run on Google Cloud using NVIDIA Blackwell GPUs, allowing Apple Intelligence features to process heavy AI workloads on third-party hardware while main…
NVIDIA released Nemotron 3 Ultra, a 550-billion-parameter open-weight model, on June 1, 2026, and it became available on Ollama's cloud three days later. The model uses a sparse Mixture-of-Experts des…
Researchers at Doubleword developed UCCL-EP, an expert parallel communications kernel that enables GPU-initiated communication patterns like DeepEP to work across arbitrary NIC-accelerator pairs, incl…
DeepMind released DiffusionGemma, a new large language model that uses discrete diffusion to generate entire sequences in parallel instead of autoregressive token-by-token generation. The model achiev…
NVIDIA Research and Tencent YoutuResearch have released open-source systems that replace the VAE decoder bottleneck in latent diffusion models, enabling faster 4K image generation. NVIDIA's PiD reform…
NVIDIA introduced intent-based security profiles in its Unified Fabric Manager for Quantum InfiniBand, enabling network administrators to configure multi-tenant fabric security with a single click. Th…
Google Cloud announced a collaboration with Apple to power Apple's expanded Private Cloud Compute (PCC) systems, announced at WWDC 2026, using Google Cloud's Confidential Computing portfolio and Titan…
Abridge announced a strategic investment from Eli Lilly and the launch of what it calls the first AI-native clinician intelligence platform, which documents patient-clinician conversations and uses th…