1
腾讯云：将大规模部署国产化算力，布局 NPO 超级节点
IT之家 7 月 21 日消息，据人民财讯报道，腾讯云高管日前出席世界人工智能大会分论坛时表示，为了将推理成本降低到极致，公司会大规模部署国产化算力；公司预计 2026 年第 4 季度部署 NPO（近封装光学）超级节点，并呼吁国内外统一 NPO 行业标准。 另据财联社报道，此次世界人工智能大会期间，…
2
NVCF: Deploy and Route GPU-Accelerated AI Workloads at Scale
英伟达开源NVCF，大模型推理规模化部署与动态路由全搞定，AI基建团队必看。
Article URL: https://github.com/NVIDIA/nvcf Comments URL: https://news.ycombinator.com/item?id=48772345 Points: 1 # Comments: 0
3
How OpenAI Delivers Low-Latency Voice AI for 900M Users
揭秘OpenAI如何用WebRTC为9亿用户打造毫秒级语音AI体验，技术选型干货满满。
In this article, we will look at the entire journey in detail and challenges the OpenAI engineering team faced.
4
Do You Really Need a GPU to Guard Your LLM? CPU-Class Classifiers and Multi-Stage Pipelines for Safety Enforcement at Scale
揭示CPU级分类器在多阶段安全管道中高效防护LLM，打破GPU依赖，量产安全新思路。
arXiv:2512.19011v3 Announce Type: replace-cross Abstract: Safety classifiers that screen LLM inputs for jailbreak attempts have become standard deploy…
5
MinT: Managed Infrastructure for Training and Serving Millions of LLMs
百万级LLM训练与服务的托管基础设施技术报告，揭秘超大规模模型管理的核心架构。
arXiv:2605.13779v2 Announce Type: replace-cross Abstract: We present MindLab Toolkit (MinT), a managed infrastructure system for Low-Rank Adaptation (…