The morning briefing.
While you slept: AWS addresses AI agent limitations, Sam Altman defends scaling, and China pushes domestic GPU and AI development amidst a shifting global market.
AI Agent Evolution. AWS has launched new services, Continuum and Context, to address critical gaps in AI agents, specifically their lack of business context and security vulnerabilities. These services aim to improve agent reliability by building knowledge graphs from corporate data and automatically fixing code vulnerabilities. This move acknowledges the current limitations of AI agents that generate code quickly but often inaccurately, highlighting a key area for enterprise AI development.
Scaling & Market Dynamics. OpenAI's Sam Altman argues that underestimating scaling potential held back AI research for a generation, citing recent mathematical breakthroughs as evidence. This perspective comes as the AI token subsidy war appears to be ending, with Google threatening significant price cuts that expose structural asymmetries between AI startups and tech giants. The semiconductor equipment market is also seeing a shift in pricing power towards sellers due to surging AI-driven demand.
China's AI & Hardware Push. China is rapidly advancing its domestic AI and hardware capabilities, with Moore Threads reporting a 243% revenue surge in 2025 for GPUs amid an AI boom and export restrictions. Alibaba has strengthened its AI leadership by adding scientist Zhou Jingren as a partner and placing CEO Wu Yongming in direct control of its new Token Foundry AI unit. Despite significant investment, with RMB 46 billion poured into embodied AI in H1 2026, widespread factory deployment remains elusive due to data scarcity and technical immaturity.
AI Policy & Societal Impact. The Pentagon is actively using AI, with officials boasting about employing GenAI.mil to generate Congress reports, indicating growing government adoption of AI tools. Meanwhile, security experts warn of increasingly sophisticated AI-boosted scam campaigns that can trick even the smartest victims, highlighting the urgent need for enhanced awareness and protection. The Guardian also reports that brands are quietly deploying AI-generated influencers on social media, raising concerns about transparency and authentic consumer experiences.
AWS Launches New Services to Address AI Agent Business Context and Security Gaps
AWS unveiled Continuum and Context at its New York summit, designed to detect and fix code vulnerabilities and build knowledge graphs for AI agents. These services aim to resolve issues where agents write code quickly but often produce incorrect results.
Sam Altman Argues Researchers Underestimated AI Scaling Potential
At a Stanford talk, Sam Altman asserted that a generation of researchers hindered AI progress by failing to grasp the power of scaling. He cited OpenAI's recent disproof of a mathematical conjecture as supporting evidence for his view.
Zscaler Executive Warns Companies Must Align with European AI Requirements
Casper Klynge of Zscaler stated that companies failing to meet European AI requirements will face difficulties in the European market. His comments addressed questions of sovereignty and the future of AI regulation in the region.
Security Experts Warn of AI-Boosted Scam Campaigns Tricking Smart Victims
AI-powered scams are rapidly evolving, tricking victims within minutes through convincing fake identities. Security experts caution that simply being aware is no longer enough to protect against these advanced campaigns.
Cainiao Deploys ZeeBot Climbing Robots to Boost Warehouse Efficiency
Cainiao has deployed over 100 ZeeBot climbing robots in a European fulfillment center, claiming a threefold increase in floor efficiency and double the labor productivity. These robots are transforming global warehousing operations.
AI Token Subsidy War Nears End as Google Threatens Price Cuts
Token subsidies are collapsing in the AI market as Google threatens an 80% price cut, revealing structural asymmetries between AI startups and tech giants. This shift is reshaping pricing power and competitive dynamics.
Unitree Robotics Faces AI Challenge Despite Cost Engineering Profitability
Unitree has achieved profitability through aggressive cost engineering in the humanoid robot sector. However, the company faces critical gaps in AI capabilities as competition intensifies.
AI Demand Shifts Pricing Power in Semiconductor Equipment Market
AI-driven demand is disrupting the semiconductor equipment market, leading to a shift in pricing power from buyers to sellers. This marks a significant change for the industry after decades.
Brands Deploy AI-Generated Influencers on Social Media Without Disclosure
An investigation found that companies are increasingly using AI-generated influencers to promote products on social media. These campaigns often purport to show genuine customer experiences without clear indication that the individuals are not real.
Chinese Startup Uses AI to Address Fusion Energy Software Bottleneck
A Chinese startup is leveraging AI to overcome the software bottleneck in fusion energy development, aiming to shorten the costly trial-and-error cycle. Fusion theorist Xie Huasheng believes better software is key to advancing the industry.
Alibaba Strengthens AI Leadership with New Partner and CEO Oversight
Alibaba has appointed AI scientist Zhou Jingren and CFO Xu Hong as new partners, with CEO Wu Yongming now directly managing the new Token Foundry AI business unit. This move signals a strategic focus on AI development.
Embodied AI Sees Billions in Investment, But Factory Deployment Lags
China's embodied AI sector attracted RMB 46 billion in investment during the first half of 2026. However, real-world factory deployment is still lagging due to persistent challenges with data scarcity and technical immaturity.
Moore Threads Reports Significant Revenue Growth Amid China's GPU Boom
Moore Threads reported 1.5 billion RMB in revenue for 2025, a 243% increase, driven by exploding domestic GPU demand in China. This growth is fueled by the AI boom and ongoing export restrictions.
Pentagon Officials Use AI to Generate Congress Reports via GenAI.mil
Pentagon Chief Technology Officer Emil Michael encourages the use of GenAI.mil, with 1.5 million Pentagon workers now using the platform. Officials are openly discussing its use for generating reports for Congress.