AI Industry Development & Policy Dynamics

According to Reuters, Japan announced on October 23 it will establish a team to simulate cyberattacks using AI to identify weaknesses following critical PHP vulnerabilities (CVE-2026-109618 and CVE-2026-109619) that triggered widespread outages. SoftBank suspended its emergency mobile alert

OpenAI reduced its proposed valuation by $150 billion to $1.25 trillion from $1.4 trillion, according to an October 23 Bloomberg report, as the company approaches final stages of a major fundraising round. The approximately 10.7% reduction reflects investor caution as OpenAI seeks to raise $30 billi

According to Gang.fun's official announcement, the BNB Chain Agent launchpad Gang.fun launched Gang Pair on Oct 23. The platform now enables trading pairs with tokenized stocks such as NVDAB, TSLAB, and AAPLB, as well as any tokens with liquidity pools on the BSC network. A 1% trading fee applies to

According to Guru Club, the State Council Information Office issued on October 23 an opinion aimed at mitigating employment displacement from artificial intelligence applications. The policy prioritizes AI deployment in labor-scarce and high-hazard roles, while exercising caution in

Japanese stocks declined on the 23rd, with the Nikkei 225 index falling 214.68 points, or 0.34%, to close at 62,693.16, marking the second consecutive session finishing in the 62,000 range. The decline was driven by selling pressure in artificial intelligence and semiconductor-related stocks

According to Guru Club, Kimi (developed by Dark Matter) released K3.1 today (October 23) with a new medium reasoning mode, bringing the total reasoning intensity options to four: low, high, max, and medium. The announcement was confirmed via an official video posted by Kimi's account, which

According to The Wall Street Journal, OpenAI dismissed 3 security researchers following an internal investigation into suspected leaks of sensitive model information. The company's security team discovered the researchers had been communicating with external AI safety researchers via Signal's

According to Axios, major U.S. banks including Morgan Stanley, JPMorgan Chase, Bank of America, and Wells Fargo recently discovered unauthorized third-party AI agents accessing customer accounts without bank approval. In some cases, the AI agents' activities appeared to exceed user

Baek Yeon-joo, a research fellow at the Korea Institute of Finance, stated on Oct. 23 that artificial intelligence agents could accelerate deposit outflows from banks and erode profitability. Baek made the remarks at '2026 Payment Insight,' a seminar marking the 40th anniversary of the Korea

According to ChainCatcher, Bytedance's Volcano Engine launched Seedance 2.5 4K, upgrading its video generation model from 1080P to 4K resolution for film and television production. The new version enhances facial microexpressions, subtle movements, and image details through specialized training. Th

According to Anthropic's September report, artificial intelligence could push the U.S. unemployment rate to approximately 12%. In its IPO prospectus, the company noted that capital raised could displace jobs. Anthropic is targeting a valuation as high as $2 trillion,

According to CCTV-affiliated Yuyuan Tantian account on October 22-23, Chinese AI developer DeepSeek and Hong Kong-listed MiniMax received state approval for global use of their AI models. The approvals followed China's October 16 proposal of export controls on 23 AI technologies, though the

According to CNBC, Nvidia CEO Jensen Huang visited Kazakhstan on Friday for the first time, announcing a $10 billion AI infrastructure investment plan in partnership with U.S.-based AI cloud company Firebird. The initiative aims to build a computing cluster powered by 100,000 advanced GPUs across Ka

According to ChainCatcher, robotics company Galactic General and Unitree Technology announced recently at the 2026 World Physical AI Conference that they jointly developed Galbot ET1, the world's first native AI agent robot. The product debuted at the World Robot Conference in August. The

According to Evan Conrad's article published on October 22, most AI startups will be unable to secure computing capacity within the next six months as global GPU supply has been depleted. Startups are being asked to make increasingly large upfront payments—Conrad noted seeing contracts

According to Anthropic's Frontier Red Team research released on October 22, Claude has successfully completed formal verification of 123 open-source libraries at the RISC-V assembly level, marking the first known AI-driven large-scale formal verification of production software at the assembly

According to People Finance, Zhou Hongyi, founder of 360 Group, commented today on DeepSeek's decision to open-source its internal AI inference framework DeepInfer. Zhou said the move reflects "technological confidence and an open-source philosophy of shared benefits," signaling that the era of tech

According to Bitcoin Magazine, on October 22, Square Tech, GOAT Network, and DAC Labs signed a Memorandum of Understanding (MOU) to explore AI agent-powered stablecoin payment services in South Korea as part of Square Tech's Digital Solutions business. As an initial priority, the partners will