Flash News

TREE NEWS update: AI startup Temporal has closed a $550 million late-stage funding round at a $12.55 billion valuation, more than doubling its valuation in seven months. Lightspeed Venture Partners led the round, with Wellington Management, Goldman Sachs Alternatives’ Growth Equity and Tiger Global as co-leads. SV Angel and T. Rowe Price Associates advised accounts also participated, while Andreessen Horowitz, Sequoia Capital and GIC joined as existing investors.

TREE NEWS reports: US data centers could add 15 billion cubic feet per day of natural gas demand by 2035, more than double a prior forecast. About 69% of the power for newly grid-connected data centers is expected to come from natural gas, while LNG exports are set to further tighten supply pressure.

TREE NEWS update: OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei met with U.S. Senator Elizabeth Warren last weekend. The meeting brought the two leading AI executives together with the senator, who has been vocal on technology and financial oversight issues. No details of the discussion were disclosed.

TREE NEWS update: Anthropic has signed a six-year compute procurement agreement with Rum Group, formerly Rumble, worth $13.7 billion. The deal brings Anthropic’s total compute procurement agreements over the past year to at least $517 billion.

TREE NEWS update: European Central Bank President Christine Lagarde said Europe cannot simply buy AI technology from the United States and must build some capacity of its own, warning that otherwise it risks becoming vulnerable. Speaking Monday, Lagarde said importing new technology rather than producing it is not inherently wrong, but AI is “special.” She flagged access rights as a key concern, since a US withdrawal of access or change in terms once AI is widely used would hit every industry at once.

TREE NEWS update: US President Donald Trump said the United States holds a large lead in artificial intelligence. He made the remark in a statement warning against killing “the goose that lays the golden eggs.” No further details, figures or policy measures were given.

TREE NEWS update: Xiaomi founder, chairman and CEO Lei Jun visited Unitree Robotics on Thursday, touring the company and watching a humanoid robot demonstration alongside Unitree founder Wang Xingxing. Earlier in the day Lei Jun appeared at Bilibili’s headquarters, where the company greeted him with a banner reading “Welcome Bilibili UP host Lei Jun” and chairman and CEO Chen Rui received him. In the afternoon he traveled to Hangzhou to hand over cars to more than 30 Xiaomi Pengcheng owners.

TREE NEWS reports: Several of AI’s largest labs are publicly calling for a slower pace of development. The shift raises the question of whether AI risks are catching up with the technology, or whether the economics of the race are becoming harder to justify. No specific companies, figures or dates were named.

TREE NEWS reports: The European Union is considering restricting access to social media platforms and chatbot tools for users under the age of 15. The measure under discussion would place age-based limits on minors’ use of both services, though no legislative text or effective date has been set. The proposal is part of broader EU scrutiny of how online platforms and automated chat tools affect younger users.

TREE NEWS update: Consensus will return to Hong Kong in early February 2027, marking its third consecutive year in the city. The event, which drew more than 10,000 attendees in both 2025 and 2026, will focus on institutional adoption of digital assets and artificial intelligence, with AI expected to carry greater weight than in previous years. Hong Kong’s regulatory framework and status as a global financial hub have made it a leading hub for digital asset adoption in Asia.

TREE NEWS reports: Anthropic may still go public in 2026, with its IPO timetable unchanged even as AI safety issues have become a prominent topic of the era, multiple sources said. The company’s listing schedule has not been altered.

TREE NEWS update: Canadian Prime Minister Mark Carney said he has held discussions on AI safety with OpenAI, Anthropic and Google. The talks cover the safety of artificial intelligence systems, Carney said, without giving further details on timing or the content of the conversations.

TREE NEWS update: US President Donald Trump said the government has already stopped AI from doing bad things. The remark, made without naming a specific company, model or incident, is the only detail given. No further information was provided on what action was taken or when.

TREE NEWS update: US President Donald Trump said the only control AI needs is a strong president, adding that his administration already holds regulatory authority over these companies. In the same remarks he accused Anthropic CEO Dario Amodei of pretending to be a “perfect little angel.” No policy measures or regulatory actions were announced.

TREE NEWS update: Microsoft has established a set of guidelines to place limits on its future artificial intelligence models. The company spent roughly five months drafting the rules, which govern how its AI models may be used and developed.

TREE NEWS reports: Nvidia, Palantir and Booz Allen Hamilton are set to restrict their use of OpenAI models. The three companies — a chipmaker, a data analytics firm and a defense contractor — are reported to be limiting their reliance on the AI developer’s models. No further details on the scope or timing of the restrictions were given.

TREE NEWS update: Aurora Optoelectronics (奥瑞德) said it plans to issue up to 825 million A-shares to no more than 35 selected investors, raising no more than RMB 868 million. Proceeds would fund RMB 282 million for an Inner Mongolia sapphire production base, RMB 346 million for a western China domestic server cluster project, and RMB 240 million for working capital. The board approved the plan; it still needs shareholder and Shanghai Stock Exchange approval plus CSRC registration.

TREE NEWS update: WUS Printed Circuit Co. said its products for 1.6T switches have entered mass shipment, with pre-research underway on components for 3.2T switches. In computing products, the company said core products supporting 224Gbps transmission are in small-batch supply and it has begun pre-research on products supporting 448Gbps. The disclosures cover its work on super-node architectures spanning compute, switching and interconnect.

TREE NEWS reports: China Telecom Research Institute released its 2026 report on AI infrastructure in the agent era, projecting China’s annual token consumption will reach 1 quadrillion in 2026 and exceed 350 quadrillion by 2030 — a 350-fold increase over four years, a compound annual growth rate of about 333%. The report describes a strategic shift in China’s AI industry from large-model and compute competition toward scaled agent deployment and application monetization.

TREE NEWS update: China’s National Technical Committee 260 on Cybersecurity released the AI Safety Governance Framework 3.0 on September 14 at the opening ceremony of National Cybersecurity Week 2026. The framework was drafted under the guidance of the Cyberspace Administration of China, with contributions from the China Academy of Cyberspace Studies, the CAC’s Data and Technical Support Center, research institutes and industry firms, to address new AI trends and governance challenges.