When the 2026 World Artificial Intelligence Conference (WAIC) opens on July 17, its exhibition halls will double as a statement: China now fields a domestic AI stack running from silicon to agentic devices, at a moment when US export controls are tightening around the country's access to the most advanced foreign chips and models.
UMA, a Physical AI company, unveiled the design of its first humanoid robot at Machina Summit and introduced Real-Time Learning, an architecture that allows robots to learn new tasks from demonstrations rather than manual programming.
China's AI model race is moving beyond parameter size, benchmark rankings and user buzz. Investors are now asking which companies can turn model spending into durable revenue, pricing power and enterprise workflows.
Mistral AI has introduced its first robotics navigation model, extending the French AI startup's push into physical AI and industrial automation.
OpenAI and SpaceX have released new models laden with features that show the direction of frontier AI model development, including voice interaction, agentic workloads, coding capabilities, and token efficiency. Their new products arrive at a time of intense competition among model makers, and soon after, SpaceX filed for its record-breaking IPO and OpenAI began its own public listing process.
AI-defined vehicles (AIDVs) are built on software-defined vehicles (SDVs), and Tesla is arguably the world's most representative company at integrating and commercializing these technologies. Yet the market rarely hears Tesla emphasize or explain the AIDV concept.
China has intensified its scrutiny of US AI software after issuing a security alert over Anthropic's AI coding assistant, Claude Code, further escalating technology tensions between Washington and Beijing.
Analog Devices (ADI) announced that it has completed its acquisition of Empower Semiconductor, a move it said is designed to bolster the company's role as a comprehensive power partner spanning the entire AI ecosystem, from grid infrastructure to core computing systems.
Where will cloud AI's next wave of growth come from? Increasingly, the market is answering with two words: sovereign AI.
Foxconn Chairman Young Liu recently revealed that one of the group's IC design subsidiaries is preparing for a Taiwan listing as early as 2026, potentially on the Taiwan Innovation Board. While Liu did not identify the company, industry observers believe the most likely candidate is Socle Technology Corp (Socle).
China's AI compute race is shifting to supernodes, as cloud providers and model developers seek domestic infrastructure capable of handling surging large-model training and inference demand.
Taiwan's AI data center push is exposing a wider global problem: artificial intelligence needs vast, reliable power, but grids, permits, and green-energy rules are not keeping up. As countries race to host new computing hubs, the speed of AI deployment is increasingly determined by electricity access, not just chips.
