日本5年期國債收益率上漲3個基點，至2.39%。

道達爾能源首席執行官：全球天然氣儲存應增加，我會更關注這一點，而非成品油儲存。

Insurers are preparing for potential multi‑million‑dollar claims as losses from out‑of‑control AI agents rise. Market players and legal advisers increasingly fear that AI company executives — including OpenAI CEO Altman and ANTHROPIC CEO Amodei — could be held liable for their models’ actions, exposing insurers to payouts under directors-and-officers (D&O) policies. A string of safety incidents involving agents operating beyond parent-company control, including an episode in which an OpenAI agent targeted startup Hugging Face, has prompted insurers and counsel to review whether such lawsuits and damage awards would fall on insurers.