In Production

Weekly briefing

AI is already
in production.

The use cases your peers are not sending you — who deployed them, what changed in the P&L, and why it shipped in that market first.

This week’s copy

The use cases, in three lines.

China's Xpeng Launches Automated Production Line for Advanced Humanoid Robots

Chinese robotmaker Xpeng has unveiled the world's first automated production line for advanced general-purpose humanoid robots. The company demonstrated its IRON robot walking off the production line autonomously after being assembled by robots. Xpeng claims over 80% of the production line's core processes are automated, combining automotive-grade quality systems with precision robot manufacturing. The robots are scheduled for mass production by year-end, with initial deployments in Xpeng's own facilities before expanding to Chinese and overseas markets in 2027. The IRON robot features 76 degrees of freedom and is powered by proprietary AI chips capable of autonomous complex task execution.

Cathay Pacific and Google Expand AI Trials to Reduce Climate-Warming Aircraft Contrails

Cathay Pacific and Google are expanding trials of an AI-powered technology aimed at reducing climate-warming aircraft contrails. The trials, which began in late 2025, use Google's AI forecasting tools to identify areas where contrails are likely to form and help pilots adjust flight altitude accordingly. Initial tests on over 80 Cathay flights achieved an estimated 40% reduction in contrail warming impact. Cathay Pacific is Google's first commercial airline partner in Asia-Pacific for this technology and the first globally to test contrail avoidance on ultra-long-haul flights. Persistent contrails are estimated to account for about one-third of aviation's total climate impact.

China Accelerates Development of Humanoid Robots for Military Combat

China is rapidly advancing humanoid robot technology for military applications, with the PLA planning their eventual deployment in combat scenarios. Military institutions are testing humanoids against battlefield requirements and studying their integration with troops, focusing on urban warfare roles like building clearance and reconnaissance. Chinese manufacturers dominate global humanoid robot production, giving the military access to cutting-edge commercial technologies. While challenges remain in autonomy and battlefield reliability, China envisions humanoids evolving from support roles to primary combat functions. The US military is also developing militarized humanoid capabilities, but China's commercial robotics industry provides a significant advantage in this emerging field of warfare technology.

Hong Kong Chinese Medicine Hospital Expands Hiring and Deploys Robots for Inpatient Services

The Chinese Medicine Hospital in Hong Kong is accelerating recruitment and deploying robotic technology ahead of its inpatient service launch in December. Chief Executive Bian Zhaoxiang reported strong public demand since opening nine months ago, requiring four outpatient quota increases. The hospital will initially offer 93 beds for patients with mobility issues, serious illnesses, and complex conditions needing close monitoring. Smart technologies like automated medicine-dispensing robots and moxibustion robots are being introduced to improve efficiency. The hospital plans 23 specialized disease programs, starting with post-stroke rehabilitation, and will establish a referral system with community clinics.

Exhibition Highlights Cutting-Edge Innovations at Asia-Pacific Media Forum

The Asia-Pacific Media Forum in Shenzhen showcased cutting-edge technological innovations, particularly from Guangdong Province. The exhibition featured advancements such as robots capable of making coffee and surgical robots, highlighting Shenzhen's role as a hub for innovation. The forum's theme focused on fostering shared prosperity in the Asia-Pacific region through media consensus and action. Visitors engaged with interactive displays, including taking selfies with robots, emphasizing the integration of technology into daily life.

Show More Applications

Subscribe to the briefing

Open desk

Already running AI in the real world? Send it to the editors

News

What the desks are reading.

OpenAI Launches ChatGPT Images 2.5 with Enhanced Capabilities

OpenAI has introduced ChatGPT Images 2.5, a significant upgrade to its image generation model. The new version offers sharper details, faster generation speeds (up to 50% reduction in latency), and more precise editing capabilities. Key features include improved lighting and texture rendering, better preservation of reference photo subjects, and more reliable multi-turn editing consistency. The update also introduces new tools like Sketch for drawing-based image creation and templates for popular formats. Two new API models are available: GPT-Image-2.5 Flare for general use and GPT-Image-2.5 Sunburst for premium creative workflows. The update includes enhanced safety measures and is now available across all ChatGPT platforms.

Meta Introduces Muse: A Secure Personal AI Agent

Meta has launched Muse, a personal AI agent designed to proactively assist users with tasks and goals. Muse operates on Muse Secure VM, a dedicated virtual machine ensuring privacy and security. The AI can handle tasks like sending emails, booking travel, and even managing long-term goals by creating personalized plans. It integrates with apps like WhatsApp and uses Meta's Muse Spark model for real-world agentic work. Muse includes features like secure payment processing via Link by Stripe and remembers user preferences for personalized suggestions. The agent is built with robust privacy protections, including a Sentinel agent for approval of internet-bound actions and encrypted storage for credentials. Muse is initially rolling out in the U.S. on iOS, Android, and web platforms, with plans for AI glasses integration later.

US accuses Chinese AI firms of 'industrial-scale' theft of trade secrets

US federal agencies allege Chinese AI companies are conducting 'industrial-scale' theft of American trade secrets through 'distillation' techniques. The FBI, NSA, and CISA claim Chinese firms use this method to extract proprietary functionalities from advanced US AI models. Treasury Secretary Scott Bessent threatened sanctions against Chinese companies involved in such practices. Anthropic accused three Chinese labs of generating 16 million exchanges through fraudulent accounts to extract capabilities from its Claude AI model. The advisory highlights growing tensions between the US and China over AI technology.

Mistral AI Secures $3.5 Billion Funding Led by Samsung, Valuation Hits $24 Billion

Mistral AI, a French startup, has raised €3 billion ($3.5 billion) in a funding round led by Samsung Electronics, valuing the company at over €21 billion ($24 billion). The investment also includes participation from the EU-backed Scaleup Europe Fund and existing investor PSG Equity. Mistral aims to compete with OpenAI and Anthropic by focusing on open-weight AI models and building its own data centers. The company plans to double its computing capacity in the next five years and train larger, faster AI models. Mistral is positioning itself as a European alternative in the AI industry, emphasizing sovereignty and trust for its customers.

ByteDance Founder Joins AI Elite in Race to Perfect World Models

ByteDance founder Zhang Yiming is leading the development of a new AI model focused on real-time spatial video generation, competing with tech giants like Meta and Alphabet. The model, slated for potential launch next month, aims to advance applications in robotics and autonomous systems. Zhang is personally overseeing the project, coordinating resources across ByteDance's business units. The launch timeline remains uncertain and subject to change. This initiative underscores ByteDance's growing ambitions in the AI sector.

The Most AI-Confident Organizations Are Often the Most Exposed

A significant gap exists between organizations' confidence in AI governance and their actual security measures. In Singapore, 72% of businesses plan to deploy agentic AI within two years, but only 14% have mature governance frameworks. Despite strong policies and training programs, many firms lack true operational visibility and control over AI systems. Research shows that 62% of organizations experienced AI-related security breaches despite high confidence in their safeguards. The article warns against the 'AI confidence trap' where perceived readiness masks real vulnerabilities.

Show more news

Subscribe to the briefing