The Qualcomm Modular deal would value the AI-infrastructure software startup at about $4bn, as the chipmaker builds out a challenge to Nvidia.

Anthropic updated its privacy policy to require some Claude users to submit government ID scans and selfies, raising biometric data concerns.

Tencent is in talks to sell minority stakes in Japanese game studios, including Tokyo-listed Marvelous, sometimes back to founders and at a loss, Bloomberg reports.

Google is building an AI startup incubator drawing on its network of former employees, according to Bloomberg, as ex-staff leave to found their own ventures.

WiseTech founder Richard White denies any involvement in human trafficking after reports of a police probe into a visa-for-sex claim. Shares fell sharply.

TensorX has raised €8m to buy Nvidia Blackwell B300 GPUs and expand its GDPR-compliant AI inference platform, run from Dublin and Helsinki.

Prosus has launched ToqanClaw, a no-code platform that builds business apps from a conversation, bringing OpenClaw-style tools to 5m+ merchants.

The Five Eyes intelligence alliance has warned that frontier AI capable of supercharging cyberattacks is months, not years, from public release.

A US judge ruled Workday must face claims its AI hiring software allegedly screened out applicants in ways that broke California law and a federal disability ban.

A UN University report warns AI’s carbon, water and land footprint is soaring, as Antonio Guterres presses AI firms to disclose costs and switch to clean power.

SoftBank founder Masayoshi Son sees little merit in orbital data centres, the idea Elon Musk has championed, and says the AI race will be won by compute on Earth.

IBM has joined OpenAI’s Daybreak Cyber Partner Program and launched an AI application-security service that hunts software vulnerabilities faster.