Google DeepMind和CFS使用AI控制聚变等离子体,目标是到2026年实现净能量目标。 Google DeepMind and CFS use AI to control fusion plasma, targeting net energy by 2026.

Google DeepMind正在与联合启动的英联邦融合系统合作,利用AI,包括其托拉克斯软件,控制Sparc反应堆中的等离子体,目的是到2026年实现净能源收益。 Google DeepMind is teaming with fusion startup Commonwealth Fusion Systems to use AI, including its Torax software, to control plasma in the Sparc reactor, aiming to achieve net energy gain by 2026. AI系统利用强化学习实时管理极端等离子体条件,解决长期阻碍聚变进展的不稳定问题。 The AI system uses reinforcement learning to manage extreme plasma conditions in real time, addressing instability that has long blocked fusion progress. 这一努力支持粮安委在20世纪30年代初提供商业聚变电力的目标,并符合谷歌对可持续、大规模能源为其数据中心提供动力的需要。 This effort supports CFS’s goal of delivering commercial fusion power in the early 2030s and aligns with Google’s need for sustainable, large-scale energy to power its data centers. 这一合作以DeepMind先前的聚合研究为基础,并反映出技术公司利用AI加速清洁能源突破的日益增长的趋势。 The collaboration builds on DeepMind’s prior fusion research and reflects a growing trend of tech companies using AI to accelerate clean energy breakthroughs.