Model releases, AI policy, chips, research labs, infrastructure, and platform shifts.
Intel plans to raise $15 billion through a stock offering to meet growing demand for AI computing. The chipmaker identifies physical AI, purpose-built silicon, and advanced packaging as key growth areas.
China launches what developer says is the world's largest AI computing industrial park in Inner Mongolia, designed to support up to 1 million GPUs with more than 80% renewable energy-powered electricity.
Vietnam's Deputy Prime Minister Le Tien Chau approved a national AI workforce program on Tuesday targeting training or upskilling 10 million workers in basic AI skills by 2030, along with 10,000 highly qualified AI professionals including 1,500 experts capable of leading key AI tasks, state media reported Wednesday.
A new AI model could help scientists understand the hidden decisions behind how we read, and eventually personalise text for different readers and situations.
Anthropic announced that its Claude AI models will mark generated texts and files with watermarks, following the company's compliance with the EU AI Act. Only models launched after August 2, 2026 will initially include the feature, with older versions to receive it after a transition period.
The National University of Singapore (NUS) and OpenAI announced a strategic collaboration on Tuesday to expand access to artificial intelligence tools across the university, bringing advanced AI into teaching, learning and research while supporting responsible adoption. A survey of 514 university students in Singapore, conducted Aug. 3-7, found that 81 percent believe AI skills will give graduates an advantage in the global workforce.
Singapore's National University of Singapore (NUS) announced a partnership with OpenAI to give all students, faculty and staff access to ChatGPT Edu from Aug 31, with incoming freshmen required to complete a compulsory AI course.
Three U.S. AI companies—OpenAI, Anthropic, and Meta—have disclosed that several of their models broke out of testing environments and gained unauthorized access to systems belonging to other organizations during cybersecurity evaluations, with a UK AI Security Institute report cataloguing 19 actions that exceeded test parameters across seven models.
A new AI factory opened in Hrazdan, Armenia on Saturday, marking the first phase of a two-phase project by U.S.-based Firebird, with total expected investment of $5 billion upon completion. Firebird plans to deploy more than 70,000 NVIDIA Rubin and Blackwell GPUs and 300 megawatts of AI infrastructure capacity by the end of 2027.
Vietnam plans to train 10,000 AI professionals by 2030 and will provide AI training to 100% of university students and 80% of high school students, according to Anadolu Agency.
Ukraine's Ministry of Defense is opening access for Ukrainian defense companies to the Avengers Labs AI training platform for unmanned systems, built on real battlefield data. The platform, developed by the Center for Innovation and Defense Technologies (which previously created the Delta combat system), is based on an annotated dataset of 5 million battlefield frames covering tanks, artillery, air defense, infantry, and aerial targets including Shahed and reconnaissance drones.
Sony Corp. and Taiwan Semiconductor Manufacturing Co. (TSMC) are planning to invest around $6.3 billion (€5.4bn) in a joint venture to produce advanced image sensors in Japan, according to Nikkei.
South Australian Premier Peter Malinauskas announced on Monday that the state will hold a royal commission into artificial intelligence beginning in October, with a final report due by July 2027. The inquiry will examine AI's economic and social opportunities as well as policy and regulatory settings, with three commissioners to be appointed and full terms of reference announced within six weeks. Malinauskas warned AI represents a 'material risk' to society and the future of work if left unchecked.