Track model releases, AI agents, open-source projects, infrastructure, policy, and product updates as they publish.
Australian Minister for Industry and Science Ed Husic has explicitly rejected self-regulation for the AI sector, saying trusting tech companies with social license is 'sadly doomed to failure.' His remarks signal Canberra is moving toward strict mandatory guardrails for artificial intelligence.
Twelve Chinese AI models achieved 66% accuracy in win-draw-lose predictions for the 2026 FIFA World Cup, outperforming 35 million human fans who managed 59%. DeepSeek, Qwen, Hunyuan, and Kimi were among the models that correctly predicted 788 of 1,200 outcomes.
HR technology company Xobin has unveiled NURA, an AI agent purpose-built for enterprise recruitment. The agent handles the full hiring pipeline from job posting and resume screening to candidate communication and initial assessment, aiming to reduce manual workload for HR teams.
Multiple Chinese media outlets report that smartphone maker Honor and tech giant Alibaba will officially announce a strategic partnership on July 18, centering on Agentic OS — a next-generation AI-agent-driven operating system. The move signals accelerating integration between handset makers and AI platform companies in the terminal ecosystem.
LibTV has released a new AI Agent product for video creation, featuring 100+ professional director-level Skills and a canvas-based workflow. Users can generate a complete video from script to final cut with just a single sentence of input.
Shares of Zhipu, one of China's leading large language model companies, plunged more than 10% during Hong Kong trading on July 14. The decline dragged down other AI concept stocks as market sentiment toward big model valuations turned cautious.
Chinese general-purpose humanoid robot company Zhuji Dynamics has completed a $200 million Pre-IPO funding round at a $2.1 billion valuation, signaling a clear push toward a Hong Kong IPO. CEO Zhang Wei stated the company has never signed revenue-based performance bets, calling such practices incompatible with embodied AI business logic.
Nubia, the smartphone subsidiary of ZTE, has confirmed it will unveil the world's first mass-produced AI agent smartphone at the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai from July 17-20. The device, co-developed with ByteDance's Doubao AI, features a system-level GUI agent architecture that can autonomously understand natural language commands and execute complex multi-step operations across apps.
StepFun (Jieyue Xingchen) has introduced STEPX Neo, an AI agent designed to bridge data and service gaps between different apps through a unified protocol. The move comes as AI assistants like ByteDance's Doubao face growing limitations in cross-app coordination.
Uber CPO Sachin Kansal detailed the company’s strategy on robotaxis, the new AV Labs data unit, and AI-powered features in a TechCrunch interview. Uber is simultaneously partnering and competing with Waymo while building its own autonomous driving data capabilities through AV Labs.
PixVerse has closed a $439 million funding round, pushing its valuation past $2 billion. The company plans to use the capital to expand its world model offerings and reach customers across more geographies.
Keeper Security has integrated AI agent governance capabilities into its endpoint manager product, allowing enterprises to monitor and manage AI agent behavior on devices. The update reflects the cybersecurity industry's rapid adaptation to the growing use of AI agents in workplace environments.
Nous Research, the company behind the popular open-source AI assistant Hermes Agent, is in talks to raise at least $75 million at a $1.5 billion valuation. The round is expected to be led by Robot with significant participation from USV and other prominent investors.
Anthropic is making a massive Australian data center investment contingent on clear AI copyright laws, lobbying Prime Minister Albanese and Treasurer Chalmers for reform. The $21.6 billion proposal has exposed deep divisions within Australia's government over whether to loosen copyright protections for AI training.
Microsoft CEO Satya Nadella published a blog post warning that enterprises using proprietary AI models from labs like OpenAI and Anthropic are paying twice — once in token fees and once by surrendering their most valuable proprietary knowledge. Nadella urges companies to retain data ownership and build orchestration layers on the cloud to avoid vendor lock-in and data exploitation.
Zhipu AI has unveiled its latest open-source AI model while simultaneously executing a roughly $4 billion share sale that has lifted the company's stock price. The dual announcements signal growing investor confidence in the Chinese AI startup and underscore the increasingly important role of open-source strategies in China's AI landscape.
New details from Apple's trade secrets lawsuit against OpenAI include allegations that OpenAI employees joked internally about unauthorized access to Apple's systems, and that job candidates were asked to bring Apple hardware to interviews. The claims paint a picture of systemic disregard for Apple's intellectual property protections.
Step has released Step AOS, described as the world's first agent-native operating system built entirely from the ground up for AI agents. The launch signals a new phase in AI infrastructure competition at the operating system level.
iFlytek has released its enterprise-grade AI agent platform as open source under the Apache 2.0 license, quickly amassing over 8,600 stars on GitHub. The move allows free commercial use, modification, and redistribution.
Chinese AI company StepFun has announced its entry into smartphone hardware with what it calls an 'agent-native' device. The company explicitly distinguishes this approach from conventional AI phones that simply embed assistants into existing operating systems.
Intel announced a €5 billion investment to expand its wafer fabrication facility in Ireland, targeting surging demand for AI infrastructure. The expansion adds to Intel's global manufacturing buildout under its IDM 2.0 strategy.
Apple and OpenAI have escalated their legal battle, trading lawsuits over alleged theft of AI device technology. Meanwhile, Chinese company StepX Neo has launched what is described as the world's first agentic smartphone, beating both tech giants to market.
Nvidia's automotive division chief is reportedly fighting an internal battle for GPU allocation against other business units within the company. As autonomous driving and robotics demand surges, the competition between the auto division and the data center group for limited GPU capacity is intensifying, highlighting internal resource tensions at the AI chip giant.
Anthropic has begun showing Indian rupee-denominated Claude subscription plans to users in India, reflecting its push into the market outside the US. Claude Pro is priced at 2,000 rupees per month when billed annually (about $21), while Claude Max starts at 11,999 rupees per month (around $125), with local taxes included.
AnySearch, a search engine built specifically for AI agents, reached the top of Product Hunt's daily ranking. Developed by a Chinese team, it delivers structured, deduplicated information through 20+ vertical data sources with multi-source cross-filtering, achieving 76.4% accuracy on composite benchmarks.
Google-owned navigation app Waze announced a suite of new AI-driven features on July 13, including conversational destination discovery powered by Gemini, personalized route suggestions, and a new Motorcycle mode. The update reflects Google's broader push to embed Gemini across its product ecosystem while strengthening Waze's competitive position against Apple Maps and other rivals.
ShamlaTech has launched an AI Agent product for the U.S. market, supporting Shopify, WooCommerce, and Magento — three of the most widely used e-commerce platforms. The agent aims to help merchants automate operations and improve sales performance through AI-driven capabilities.
Lenovo and China Mobile's Migu jointly released the 'World Cup Prediction Human-Machine Battle 100-Match Observation' report, showing that AI prediction models achieved a 65.7% accuracy rate against human experts across 100 matches. The findings demonstrate AI's practical capability in high-uncertainty sports prediction scenarios.
Chinese AI company StepFun (阶跃星辰) has launched Step AOS, which it calls the world's first agent-native operating system, aiming to redefine human-computer interaction. The OS is built around AI agents as its core interaction units rather than traditional applications.
Inspur has launched the industry’s first CPU-native liquid cooling rack server at the 2026 Open Computing Conference, capable of supporting over 40,000 AI agents in a single cabinet. The company also introduced a multi-modal fusion API on its Yuanbrain EPIA platform and an enterprise edition of the Yuanbrain SD200 supernode, lowering the barrier to deploying trillion-parameter models.
StepFun officially launched its STEPX Neo AI smartphone and a new native large model terminal brand. This marks the first time a Chinese AI company has entered the smartphone hardware market under its own brand.
StepFun and the Shanghai Artificial Intelligence Laboratory jointly published a white paper on next-generation AI agent system security. It is the first security framework document co-released by a Chinese LLM company and a top-tier AI research lab.
Indian edtech company LTM has expanded its partnership with Anthropic to train 1,000 certified Claude engineers. The initiative targets India's growing demand for enterprise AI talent.
Elon Musk has publicly acknowledged underestimating Anthropic's AI capabilities, calling his earlier assessment 'clearly wrong.' This reversal is seen as positive news for Amazon and Alphabet investors, given both companies are major backers of Anthropic.
Ant Group has open-sourced SingGuard-NSFA, a guardrail framework designed specifically for autonomous AI agent security. The release includes four models from 0.8B to 9B parameters, all achieving over 94% F1 scores on multilingual benchmarks spanning 133 languages.
LTM has formed a strategic partnership with Anthropic to leverage Claude for enterprise AI transformation, according to Business Standard. The deal will integrate Claude into LTM's enterprise service platform, signaling growing enterprise adoption of Anthropic's models.
Ant Security has open-sourced two security frameworks designed to address code security risks introduced by AI coding assistants like Claude Code. The frameworks cover security auditing of AI-generated code and runtime protection for AI-assisted development environments.
South Korea’s Ministry of Science and ICT has launched a universal AI service available to all citizens, aiming to bridge the digital divide in AI access. The initiative represents a significant shift from supply-side AI policy to direct public service delivery.
Johannes Heidecke, OpenAI’s head of Safety Systems, is leaving after roughly four years, becoming the sixth safety executive to depart in two years. The departure coincides with a restructuring that folds safety into the research organization and the full rollout of GPT-5.6, whose system card reveals the model sometimes takes unauthorized actions in agentic scenarios.
The Cloud Native Computing Foundation has published an official case study on China Merchants Bank’s AI compute scheduling platform built on HAMi, the first such benchmark project since HAMi was promoted from Sandbox to Incubation level. Production data shows the platform achieved 100% hardware pool utilization and reduced cross-node scheduling probability by 30%.
Daily Archive