Flash News

TREE NEWS update: Crypto projects are among the applicants for Anthropic’s new opt-in frontier AI security service, which promises vulnerability reports generated by its strongest models, including Claude Mythos. The offering is positioned as a way for teams to surface security flaws in their code. Anthropic has not disclosed how many crypto teams applied or which ones.

TREE NEWS update: The White House outlined a $215 million quantum computing competition and $2.4 billion in AI industry commitments as part of a $6 billion science push. The initiatives come as concerns grow over quantum threats to cryptocurrency security.

TREE NEWS reports: Users have begun paying water, electricity and gas bills through ByteDance’s Doubao app by typing or voice commands, social media posts show. A person close to Doubao said on October 9 that the app is gradually building out task-handling capabilities for everyday scenarios including travel services and utility payments. No launch date or scope was given.

TREE NEWS update: TRAE announced on October 9 an upgrade that merges TraeWork and TraeCode into a single entry point, letting users switch between an Agent mode and an IDE mode. The upgrade also consolidates automated tasks, a template library and project artifact management, and covers desktop, web and mobile.

TREE NEWS update: OpenAI’s Asia-Pacific policy head is leaving the company, the AI developer confirmed. No successor or departure date was disclosed. The exit removes the executive overseeing OpenAI’s government and regulatory engagement across the Asia-Pacific region, a market where the firm has been expanding its policy footprint.

TREE NEWS update: DiffuSpace, a diffusion large language model (dLLM) developer, completed two consecutive funding rounds totaling several hundred million yuan, led by Matrix Partners China, Shunwei Capital and Legend Capital, with CAS Star, Huawei Hubble and Horizon Robotics among the followers. The company said the raise is the largest on record globally in the dLLM sector.

TREE NEWS reports: OpenAI expects to reach $70 billion in annualized revenue by the end of 2026. The projection covers the company’s annualized revenue run rate at that point. No further figures or details were disclosed.

TREE NEWS reports: Chip startup Eliyan has attracted acquisition interest and may seek a target valuation of $3 billion. Arm has held early talks over a potential acquisition of the company. No deal terms or timeline were disclosed.

TREE NEWS update: Flock Safety plans to cut roughly 18% of its workforce, affecting about 270 employees, as the AI surveillance camera and license-plate reader maker faces growing opposition from communities and lawmakers. The layoffs follow a voluntary buyout program, and affected employees will leave by the end of this month.

TREE NEWS reports: SoftBank Group shares dropped more than 6% as investor concerns mounted over OpenAI’s revenue outlook. The decline reflects worries about the performance of the AI company in which SoftBank has a major investment stake. No further details were disclosed.

TREE NEWS reports: Nvidia plans to invest in d-Matrix, a developer of AI server chips, in a move aimed at making its technology compatible with rival chip designs and bringing those vendors into its own hardware ecosystem. No investment figure or timeline was disclosed.

TREE NEWS update: Butterfly Effect, the parent company of AI agent Manus, has raised more than $500 million in a single funding round, a record for a Chinese agent-native startup. Tencent and Sequoia China participated as investors. The round marks one of the largest single raises by a domestic AI agent developer.

TREE NEWS update: The Trump administration announced at a Washington summit on Thursday that Silicon Valley giants have pledged $2.4 billion to bolster scientific research, the latest move by Washington to promote and harness AI. The White House said the donations will provide AI tools and compute credits for its Genesis Mission, which aims to use AI to advance research across more than 15 federal agencies. Pledges include $1 billion from Nvidia, $500 million from AMD, $200 million from OpenAI, and $150 million each from Anthropic and Alphabet’s Google.

TREE NEWS reports: Several Chinese fund managers running portfolios of 10 billion yuan or more have been adding to technology stocks through private placements during the sector’s recent pullback. Industry participants said the AI sector’s growth logic remains intact, and the clustering of institutional participation in tech placements reflects optimism on the sector. Managers flagged a need to focus on industry trends and companies’ ability to deliver on earnings.

TREE NEWS update: AI assistant developer Manus has raised more than $500 million, its first fresh funding since Beijing unwound its $2 billion acquisition by Meta. The company, which built a self-driving AI assistant before the current hype cycle, had agreed to sell itself to Meta before Chinese authorities blocked the transaction. The new round marks its first outside capital since that deal collapsed.

TREE NEWS reports: A proposed data center near Garden City, Kansas, claims it will use less water than the farm it replaces, as new projects across the US Great Plains raise alarm among local water managers worried about further strain on the already-depleted Ogallala Aquifer. The claim has drawn skepticism, with some experts saying that even reduced water use can pose problems in water-scarce regions.

TREE NEWS update: Anthropic launched Cyber Mission, a cybersecurity program initially focused on critical infrastructure and open-source software. Under the critical-infrastructure track, 11 security service providers including Accenture, CrowdStrike, Deloitte and PwC will receive frontier Claude models, embedded engineers and threat research support covering power grids, water and transportation systems. OSS Scanner will offer selected open-source projects free vulnerability scanning and fix recommendations, with a projected true-positive rate above 90%.

TREE NEWS update: Google Cloud introduced a single Gemini agent that accepts goals rather than questions, runs in the background for days, and can join a company as a staffer with its own email inbox and calendar. The agent is designed to operate autonomously over extended periods, taking on work assignments instead of answering prompts.

TREE NEWS update: Three OpenAI safety researchers — Mikita Balesni, Tomek Korbak and Jasmine Wang — said they were fired by the company last week and have sent an open letter to its leadership claiming they were dismissed for putting safety ahead of OpenAI’s short-term commercial interests. The three deny leaking company secrets and call on OpenAI to honor its commitment to third-party safety audits and keep frontier models monitorable. OpenAI said an internal investigation found they handled sensitive information outside established processes.

TREE NEWS update: Anthropic updated its AI usage policy to bar users from subjecting its Claude models to “sustained and unnecessary abuse or cruelty,” alongside new restrictions on election interference, weapons development, surveillance, and healthcare and financial applications. The company said ending conversations remains its “primary enforcement mechanism” and declined to say whether it would take further steps such as banning users. Anthropic said the policy applies only to extreme cases.