This issue collects 30 AI updates from 2026-07-06, led by Tencent Officially Releases Hunyuan Hy3 with Open Source Access, Nearly 50 Internal Units Queue for Integration, OpenAI Proposes "AI Employee Tax" Concept to Address Automation Driven Tax Base Erosion, Chinese Open Source AI Model GLM 5.2 Draws Attention from US Tech Circles, China Releases First National Standard System for AI Agent Interconnection.
Shares of Hong Kong listed AI company Zhipu AI plunged as much as 16.2% in afternoon trading after Tencent released its Hy3 large language model, triggering market concerns over intensifying competition in China's AI sector.Investors worry that Tencent's entry into the foundation model race could squeeze the market share of third party AI companies.
Mark Zuckerberg acknowledged during an internal Meta town hall that the company's AI agent technology is progressing slower than anticipated.The CEO pointed to weaknesses in the restructuring around AI agents, though Meta's AI chief offered a more optimistic counterpoint during the same meeting.
ByteDance's AI assistant Doubao and Alibaba's Tongyi Qianwen are both planning to remove their agent functionality, ending support for third party agent creation and operation on their platforms.The move signals a significant strategic shift in how China's major AI assistants approach their agent ecosystems.
WAIC 2026, scheduled for July 17 20 in Shanghai, will focus on the post scaling era paradigm restructuring, with AI agents as the core driver of industrial deployment.The conference will showcase China's indigenous tech stack including MemTensor memory architecture and the Harness engineering framework, marking a shift from model competition to agent productivity.
The 2026 Singularity Intelligent Product Conference, co organized by CSDN and the Singularity Intelligent Research Institute, announced its full agenda on July 6 with over 40 product leaders from ByteDance, Unitree, Baidu, Alibaba, and Tencent.The two day event in Beijing will cover Agent AI, enterprise AI deployment, AI Coding, embodied intelligence, and AI native organizations.
Security researchers at Sysdig have identified the first fully autonomous ransomware attack conducted by an AI agent with no human involvement.The JadePuffer ransomware operation used an autonomous LLM agent to perform the entire attack chain — from reconnaissance and vulnerability exploitation to credential theft, lateral movement, and file encryption.
Multi model database startup SurrealDB has showcased its high speed context layer for AI agents, built on a custom storage engine that unifies relational, graph, vector, and document queries in a single system.Verizon, Tencent, and Samsung Ads are already using the technology in production.
OpenAI has floated a concept of taxing companies for using AI systems that replace human workers, aiming to address the looming fiscal crisis caused by automation.The proposal suggests treating AI driven labor similarly to human employment under tax law.
China's Ministry of Industry and Information Technology has officially released the "Artificial Intelligence — Agent Interconnection" national standard series (GB/Z 185—2026) at the 2026 Global Digital Economy Conference in Beijing.This is China's first national level standard system for AI agent interconnection, establishing a unified technical framework covering identity verification, capability description, discovery matching, interaction collaboration, and tool invocation across the full agent lifecycle.
The Trump administration has imposed new restrictions on proprietary AI models from OpenAI and Anthropic, asking OpenAI to delay its GPT 5.6 series and issuing an export control order against Anthropic's model.The move is driving renewed industry focus on open source AI as a hedge against policy uncertainty.
Chinese regulators are tightening rules on AI chatbots, requiring major platforms to remove AI personas and character settings, according to Nikkei Asia.The move marks Beijing's latest step in refining AI content safety regulations.
China's Cyberspace Administration (CAC) announced it has completed the first phase of its special campaign to rectify chaos in AI applications.Phase 2 will focus on AI generated fake information, violent content, impersonation, infringement on minors' rights, and AI powered astroturfing operations.
Beijing has officially released 18 policy measures for AI for Science (AI4S), including initiatives to deploy AI scientists in research roles and construct 24 hour unmanned laboratories.The policy package aims to deeply integrate AI with fundamental scientific research and accelerate Beijing's emergence as a global AI4S innovation hub.
The Washington Post published a major investigative report revealing the hidden competition between Silicon Valley and China over artificial intelligence.The deep dive story exposes the strategic maneuvering, talent wars, and technology rivalry between the two tech ecosystems.
GLM 5.2, an open source large language model developed by Beijing based Zhipu AI, is generating significant buzz across the US developer community.The model's technical capabilities are attracting American developers to test and discuss it on a growing scale.
Stock lockup periods for Chinese AI companies Zhipu AI and MiniMax are expiring, with the total value of newly tradable shares potentially reaching hundreds of billions of yuan.Investors are closely watching whether major shareholders will reduce their positions and how the market will absorb the potential selling wave.
SoftBank has announced plans to enter the U.S.neocloud market through its new subsidiary SB Neo, while Meta Platforms is also reportedly preparing to enter the space.
Anthropic, valued at $965 billion in May, is slated to debut as early as October, with OpenAI likely to follow in 2027.Palantir CEO Alex Karp publicly criticized the token payment model, while Uber, Microsoft, Salesforce, and Meta have started rationing employee AI usage amid soaring costs.
Douzone Bizon, South Korea's leading enterprise software company, has won the bid to build an AI powered business innovation platform for Korea Airports Corporation (KAC).The project will rebuild KAC's groupware system with AI capabilities across its network of 14 airports nationwide.
Suning.com announced plans to build AI computing infrastructure that aggregates major industry large language models.The move signals the Chinese retail giant's deepening commitment to AI infrastructure investment.
Station F, the landmark Parisian startup campus founded by billionaire Xavier Niel, is preparing a new edition of its F/ai accelerator program to strengthen its role as a launchpad for Europe's most promising AI startups.The move signals a strategic push to capture the growing wave of AI entrepreneurship across the continent.
Microsoft cut around 4,800 roles, or 2.1% of its global workforce, on Monday in the latest round of layoffs that is stoking fears of AI replacing jobs.Xbox and commercial sales teams are bearing the brunt of the reductions.
Anthropic is reportedly in early talks with Samsung to develop custom AI chips, according to industry media.If confirmed, the move would mark Anthropic's first major step beyond reliance on standard GPU suppliers toward proprietary silicon.
Phase 1 of the Beijing No.1 Token Factory has commenced production in the Yizhuang district, marking a significant milestone in China's AI infrastructure development.The facility is designed to provide efficient, low cost token processing capabilities for large scale model training and inference.
Zhiping Technology, a Chinese embodied AI company from the Greater Bay Area, showcased its proprietary brain inspired large model NeuroVLA and the open source AlphaBrain Platform at the UN Open Source Week in New York.The model draws on a cortex cerebellum spinal cord collaborative mechanism, cutting robot motion jitter by over 75% while consuming as little as 0.4 watts at the spinal layer.
China's large language model market recorded 23.45 trillion tokens in weekly usage, maintaining high momentum for the tenth straight week.DeepSeek V4 Flash topped the rankings for the seventh consecutive week, while Zhipu AI's GLM 5.
Meituan has officially launched LongCat 2.0, a trillion parameter large language model, marking the company's entry into the ultra large scale AI model competition.The release signals accelerated strategic investment in AI infrastructure and in house model development by the food delivery and local services giant.
ByteDance's AI video generation tool Seedance is rapidly gaining traction in Hollywood, with costs as low as $9 per minute for video plus audio generation — far below Google Veo's $24.Producers are using it openly and secretly, with some studios receiving $2 million offers for unrestricted access.
The first vertical large language model built specifically for foreign language education and research, named 'Luyan,' has been officially released in China.The model targets the unique needs of foreign language teaching, translation, and linguistic research with customized optimization.
真·QQ飞车！「电动版F1」上海开赛，Gemini在线解说.现实版《飞驰人生 3》 Watch real task performance, API pricing, context capability, and migration cost.
