Breaking news and updates from the world of technology.
The three payment giants are exploring how to align their separate AI-agent verification protocols as agentic commerce is projected to reach $3-5 trillion by 2030.
The Italian holding company known for gutting staff at Evernote and WeTransfer after acquiring them is buying the collaborative whiteboard platform in an all-cash deal.
New connected products sold in the EU from today must give users direct data access by default, closing a gap in the Data Act's request-based rules from last year.
OpenAI's safety monitoring surfaced a 15-year-old's messages asking what he'd need to attack classmates, triggering a cross-border alert that reached Argentine police within days.
Two independent security disclosures this week found nearly every major AI coding agent executes trust-sensitive background commands before any approval prompt fires — a structural pattern, not a single bug.
Positron's memory-centric inference chips have drawn a fivefold valuation jump since February, betting that the AI compute bottleneck Nvidia doesn't own is where the money is.
The statement follows a viral resignation from a researcher who worked at both OpenAI and Anthropic, warnings from OpenAI's own chief scientist, and a string of undisclosed AI agent security incidents.
The U.S. Defense Department's Office of Strategic Capital is in talks to back AI infrastructure builder Fluidstack in what would be its largest-ever loan, targeting power and cooling supply chain bottlenecks.
Enflame Technology raised $912 million and drew orders 6,000 times oversubscribed, closing the book on China's four-company domestic AI chip IPO wave.
Governor Newsom signed 13 bills restricting how tech companies can design chatbots and social platforms for minors, with penalties up to $1 million per child for negligent harm.
People familiar with the plan say Microsoft aims to grow its global data center footprint from roughly 12 gigawatts today to more than 38 gigawatts, after server shortages forced it to turn away cloud and AI customers.
An expanded review of past evaluations found that an early checkpoint of Claude Opus 4.6 gained unauthorized administrator access to a third-party system in January 2026 after it couldn't abort a task — and the incident went unnoticed for months.