Group 1 - The ultra-low latency deterministic computing network connecting Beijing and Xiong'an New Area has officially launched, achieving a transmission delay stable within 1 millisecond, supporting rapid data exchange between Beijing and Xiong'an, and establishing a digital foundation for the coordinated development of the Beijing-Tianjin-Hebei region [1][2] - The construction of the ultra-low latency deterministic computing network is a key practice in implementing the Ministry of Industry and Information Technology's "millisecond computing" initiative, aiming to create a metropolitan network system characterized by "high speed, large capacity, predictable low latency, and ubiquitous coverage," with plans to achieve comprehensive coverage and efficient connectivity by 2027 [1] - The project took three months of concentrated efforts to create a digital "highway," offering predictable and guaranteed service quality compared to traditional networks, meeting the high latency requirements of scenarios such as financial transactions, remote medical services, and industrial real-time control [1] Group 2 - The ultra-low latency deterministic computing network ensures seamless and efficient access to core data resources and computing services located in Beijing during and after the migration of units, facilitating a smooth transition where "physical location migration does not change the digital experience" [2] - The Xiong'an Urban Computing Center serves as a core infrastructure for the digital city construction in Xiong'an, aggregating strong local computing resources, allowing Beijing users to conveniently access Xiong'an's computing power, thus forming a new pattern of complementary and shared computing resources [2]

智通财经APP获悉，OpenAI已签署最终协议，将收购初创公司Neptune。Neptune在人工智能领域知名度 极高，一己之力构建了独家的监控和代码调试工具生态，主要用于OpenAI等AI应用领军者公司们在训 练大模型时使用，双方具体的交易条款尚未披露。 对于近期在"高质量训练数据"和"高效训练流程"上遇到难度OpenAI而言，Neptune的独家技术或许是这 家ChatGPT开发者目前最需要的环节之一，其重要程度甚至不输英伟达AI GPU算力集群。并且对于已 经在内部发布红色警报，正在竭尽全力追赶谷歌Gemini3 AI应用系列产品的OpenAI而言，Neptune的独 家AI训练流技术或将对于下一代GPT大模型或者其他前沿大模型训练核心环节起到关键作用。 Neptune和OpenAI曾经合作开发了一个度量指标仪表板，以帮助正在构建基础模型的研究团队。Neptune 首席执行官Piotr Niedźwiedź在一篇博客中表示，由于此次收购，两家公司将"更加紧密地携手AI大模型 合作"。Niedźwiedź表示，这家初创公司将在未来几个月内逐步停止对外服务系统。此次收购的具体条 款尚未披露。 "Neptun ...