Daily AI news, sourcedUpdates / Chips & infra
Updates Chips & infra · 115 items · newest issue first · page 4 of 6
AreaAllModelsAgents & codingApps & productsChips & infraResearchPricing & limitsBusiness & fundingPolicy & safetyOther
- SonicBoom, a Python-free C++/CUDA runtime, claims 52 tok/s decode on a laptop RTX 3050 versus 45 for llama.cppUnconfirmedr/LocalLLaMAOct 13Chips & infra
- Developer adds n+2 expert lookahead to SSD-streamed Qwen Flash Next on a 64GB Mac Mini M5Unconfirmedr/LocalLLaMAOct 13Chips & infra
- Developer releases vramble, an open-source arbiter that gives local AI services one GPU lease at a timeUnconfirmedr/LocalLLaMAOct 13Chips & infra
- Four NVLinked Tesla V100s outpace two DGX Sparks and an M5 Ultra running Qwen3.8-Flash-Next, per WccftechReportedWccftechOct 13Chips & infra
- Intel names Crescent Island the GPU D-Series, with D270 offering up to 480 GB LPDDR5X for inferenceOfficialr/LocalLLaMAOct 13Chips & infra
- Alibaba Cloud's Jason Wu says AI training, inference and agents are reshaping infrastructure needs
- NVIDIA and Cisco introduce USFP pluggable optics with 400Gb/s per lane for AI factories
- Developer publishes a lecture and article on building a GPU compiler that lowers PyTorch graphs to CUDAUnconfirmedr/LocalLLaMAOct 13Chips & infra
- Ollama v0.40.4-rc0 updates llama.cpp for decision models and EmbeddingGemma 2 and lists installed models in Codex's model picker
- Microsoft will scale Azure Maia to multi-gigawatt deployments from 2027, making it the primary inference system for its products
- NVIDIA contributes Vera Rubin rack and compute tray designs to OCP and open-sources new DSX AI factory software
- shad0w learns a local LLM's classification answers and serves the repeats in microseconds
- Decapybara open-sources a C++20 and CUDA rebuild of the MegaCapybara inference serverOfficialGitHub giveen/DecapybaraOct 12Chips & infra
- LocalLLaMA user ports Gemma 4 E4B EXL3 via exllamav3 as a Jetson Orin alternative to llama.cppUnconfirmedr/LocalLLaMAOct 12Chips & infra
- QbitAI says Silicon Valley AI competition is now gated by who controls GPU supply, per QbitAIReportedQbitAIOct 11Chips & infra
- LocalLLaMA user posts a final zzpanic vllm-radiance build tuned for one R9700 running Qwen3.8-27BUnconfirmedr/LocalLLaMAOct 11Chips & infra
- Cheaper AI tokens are driving more demand, and that's Jensen Huang's best-case scenario, per The DecoderReportedThe DecoderOct 11Chips & infra
- LocalLLaMA user reports Next Flash IQ3_XXS ~70 tok/s or dual Qwen 3.6 35B-A3B ~145 tok/s on $500 BC-250 boardsUnconfirmedr/LocalLLaMAOct 11Chips & infra
- NVIDIA reportedly ends new RTX 5090 GB202 supply for GeForce, reserves chips for RTX PRO, per VideoCardz
- strata-mlx runs Qwen3.8-Flash-Next GGUF files on Apple Silicon via MLXOfficialGitHub @yibie/strata-mlxOct 10Chips & infra