Alibaba has brought day-zero support for its Qwen-3.8 27B AI model to its RISC-V-based XuanTie C950 chip, achieving 30 tokens per second decode speed and a 1.9-second time to first token. The chip, re…
Philip Kiely's new book, 'Inference Engineering,' is now available as a digital download, offering a comprehensive guide to the technologies and techniques powering AI inference across runtime, infras…
Google's Gemma 4 26B A4B, a 25.2B-parameter multimodal mixture-of-experts model with 3.8B active parameters and 128 experts, is now available with DekaLLM offering the cheapest API pricing at $0.06 pe…
Soros Fund Management's Q2 2026 13F filing reveals a $497 million cluster of AI infrastructure investments across five under-followed names, led by a $96.9 million stake in Digital Realty Trust and a …
Roboflow Inc. explains that on-premise computer vision keeps inference local, with production images processed on hardware inside the facility, which can keep sensitive data local, remove network roun…
NVIDIA released TensorRT Model Connect (TRTMC) in public preview, an Apache-2.0 project that converts a supported Hugging Face or local checkpoint to end-to-end TensorRT inference in two commands, eli…
Qualcomm open-sourced its Modular AI software platform at ModCon 2026, making the Apache 2.0-licensed code available on GitHub. The move aims to create a cross-platform AI stack that runs on various h…
PantheonGPU, a new GPU health-check tool, runs 45+ targeted tests on NVIDIA CUDA and AMD ROCm hardware to detect stability issues and configuration bottlenecks that telemetry tools miss. The tool stre…
A kernel is a single function that runs on a GPU, and the steep memory hierarchy—with a 1,600x gap between L2 cache and HBM—makes kernel design crucial for AI performance. FlashAttention, developed by…
NVIDIA has delayed mass shipments of its Blackwell NVL72 rack system to Q1 2025 due to overheating issues in the 72-GPU configuration, which exceeds conventional air cooling limits. The company confir…
RadixArk released Miles v0.1, an open-source framework for aligning AI training and inference, on GitHub. The company, founded by Ying Sheng and Banghua Zhu, aims to commercialize open-source infrastr…
Enterprise SSD prices rose another 5 percent in July 2026, leaving flash pricing about 6.5 times higher than in Q3 2025, according to VDURA's Flash Volatility Index, which prices a 30TB TLC SSD at $22…
OpenAI has joined NVIDIA's Open Weights and American AI Leadership initiative, an industry-backed policy effort advocating for the availability of downloadable AI model weights. The coalition, which n…
Craton Bolt v0.7.0, a pure Rust JIT-compiled GPU SQL engine, compiles each query into a fresh NVIDIA PTX kernel at runtime via the CUDA driver API, eliminating C++ shims and precompiled kernels. The e…
NVIDIA introduced the ALCHEMI Toolkit, a PyTorch-native suite for GPU-accelerated atomistic simulations with machine learning interatomic potentials, and demonstrated that AI coding agents can build s…
HIVE Digital Technologies announced a five-year GPU cloud services agreement worth approximately $350 million through its subsidiary BUZZ HPC during its Fiscal Q1 2027 earnings call on August 17. The …
Velaura AI, a startup focused on chip design for data centers, has crossed the $1 billion valuation threshold after closing a $110 million Series A round led by Seligman Ventures, with participation f…
FORT Robotics Inc. is going public via a merger with Newbury Street II Acquisition Corp., a special-purpose acquisition company, to scale its safety technology stack for autonomous robots. The Philade…
A developer built a team of three AI agents using the Hermes Bot Mode desktop plugin, creating a Researcher → Risk Analyst → Thesis Editor pipeline that hands off work automatically. The test demonstr…
Google rolled out a phone feature that recognizes sign language in real time, enabling Deaf users to communicate via camera without typing or third-party apps. A startup began renting home-cleaning ro…