OpenGEMM, an open-source project from developer aramesh10, released CUDA GEMM kernels for NVIDIA's B200 GPU (sm_100a), installable via pip and requiring PyTorch 2.8+ and CUDA 12.9+. The library emits …
NVIDIA's NIM 2.0.12 optimized serving stack delivered up to 2.5x higher output-token throughput than the open-source baseline when serving the Nemotron 3 Ultra model on four B200 GPUs, reaching 1,997 …
SwarmOS, backed by GPT-6 Astra, generated a Rust/CUDA fully sharded data-parallel trainer for Qwen3.8-27B that achieved over 11.5× higher throughput than the eager PyTorch FSDP2 baseline and 21.5% hig…
Anduril Industries, a defense technology company, is hiring a Senior AI Infrastructure Engineer in Costa Mesa, California, offering $166k–220k/yr, to lead GPU cluster robustness and automated resilien…
A new arXiv paper (submitted 27 Aug 2026) benchmarks confidential computing on NVIDIA B200 GPUs using Intel TDX and NVIDIA Confidential Computing, finding that correctly configured confidential infere…
Etched AI, a chip startup founded in 2022, exited stealth on June 30, 2026, unveiling its transformer-only ASIC, Sohu, which the company claims delivers 500,000 tokens per second on an 8-chip server f…
KernelArc, a multi-agent framework for autonomous GPU kernel optimization, achieved first-place rankings on representative L1, L2, Quantization, and FlashInfer tasks at the public SOL-ExecBench leader…
South Korea's Ministry of Science and ICT (MSIT) eliminated Motif Technologies from its sovereign AI competition, 'Dokpamo', after an evaluation round from August 8 to 11, leaving LG AI Research, SK T…
A looming AI debt bubble, driven by massive infrastructure spending on GPU clusters and data centers, threatens systemic financial stability, according to analysis of the tech sector's capital expendi…
A solo developer with no professional GPU background placed 12th of 183 in GPU MODE's batched QR decomposition contest, beating the cuSolver-backed baseline by 232x (419,000 µs down to 1,805 µs on NVI…
Sankalp placed 12th among 183 entrants with a time of 1,804.779 microseconds on an NVIDIA B200 in GPU Mode's qr_v2 competition, using OpenAI Codex to make over 1,500 submissions over 14 days. His resu…
Baseten has built the fastest API for GLM-5.2, achieving peak speeds of 280 tokens per second and average speeds around 100 tokens per second, more than double the performance of the launch-day API as…
Helion, PyTorch's high-level DSL for writing performance-portable ML kernels, partnered with Google to build a TPU backend that compiles Helion kernels to Pallas, enabling PyTorch-friendly TPU kernel …
Unsloth achieves up to 7.3x training speedup over standard Transformers for MoE models like gpt-oss-20B on an NVIDIA B200, according to published benchmarks, while Axolotl delivers up to 1.45x speedup…
AMD Instinct MI355X GPUs with ATOM and ATOMesh achieve competitive inference performance for MiniMax-M3, a 428-billion-parameter multimodal MoE model, outperforming NVIDIA B200 and B300 in per-GPU thr…
Self-hosting open-weight large language models can save millions of dollars compared to using inference providers, according to a practical guide by Cline that analyzes the economics using Kimi K2.6 a…
The Soofi Consortium, coordinated by KI Bundesverband and funded by the German Federal Ministry for Economic Affairs and Energy, released Soofi S 30B-A3B, an open hybrid Mamba-Transformer Mixture-of-E…
A developer released the first open-source training kernels for MiniMax Sparse Attention (MSA) on Hopper and Blackwell GPUs, enabling efficient million-token training with sparse attention. The kernel…
Researchers introduced cuTile Rust, a tile-based system for safe, idiomatic GPU kernel authoring in Rust that extends Rust's ownership discipline to GPU kernels. On the NVIDIA B200 GPU, cuTile Rust ac…
AWS invented Parallel-EAGLE (P-EAGLE), a speculative decoding method that parallelizes draft token generation, achieving up to 1.69x throughput speedup over vanilla EAGLE frameworks. Amazon SageMaker …