
Microsoft unveiled an AI coding model that runs on personal computers and a security technology to prevent unauthorized data access by AI agents, alongside a high-performance Surface laptop called Ultra powered by Nvidia RTX Spark chips, aiming to shift AI tasks from Azure cloud to Windows devices.
AI-generated summary
Microsoft is seeking to expand AI capabilities beyond its Azure cloud infrastructure by enabling AI agents to run directly on personal computers, leveraging its dominant position in the Windows operating system market.
Microsoft on Wednesday unveiled an AI coding model that can run directly on personal computers, alongside a new security technology designed to prevent AI agents from accessing data without permission.
At an event in San Francisco, the company also showcased a high-performance Surface laptop called Ultra powered by Nvidia’s RTX Spark chips.
Microsoft aims to turn Windows into a platform for AI agents capable of handling tasks ranging from writing computer code to completing complex business projects on desktop and laptop computers.
The strategy reflects Microsoft’s bet that some tasks currently handled by its costly Azure cloud computing data centres can be shifted to high-performance Windows devices in businesses and homes.
The company holds a strong position in that market, where customers bear the hardware costs.
AI outlook — possibilities, not facts
Microsoft will release the AI coding model and security technology to Windows users in the coming months.
Likely · Within months
Demand for Nvidia RTX Spark chips may increase due to their use in the Surface Ultra laptop.
Possible · Within months

The UAE has launched a major digital infrastructure project led by e& to increase international connectivity capacity to over 500 terabits per second by 2030, up from 20 Tbps, supporting AI growth and global competitiveness. Sheikh Mansour bin Zayed Al Nahyan witnessed the launch, emphasizing digital infrastructure as a pillar of the future economy and national sovereignty.

The General Directorate of Public Security in Saudi Arabia launched an upgraded version of the 'Kollona Amn' app on Monday, featuring technical and operational improvements to enhance digital security services and improve report routing to relevant authorities for faster response and accurate processing.

David Robinson, a former OpenAI safety employee who helped draft the company's preparedness framework, resigned and criticized OpenAI's rapid AI development approach in an essay published by The Atlantic, arguing that iterative deployment becomes riskier as AI systems grow more capable and that safeguards must match those in high-risk industries like nuclear power and aviation.
A leak indicates Samsung's Galaxy Z Fold 9 may measure 7.9mm when closed, thinner than the Fold 8's 9.7mm, aiming to reduce the compromise of carrying a foldable by making it feel more like a conventional phone while maintaining tablet-like functionality when opened.

The Future Health Summit 2026 in Abu Dhabi will convene global experts to discuss integrating AI, genomics, and wearable data into healthcare to shift from reactive treatment to proactive, predictive disease prevention.

Meta has launched Muse Gadgets, an open-source project allowing developers to connect physical hardware like Raspberry Pi and ESP32 boards to its Muse personal AI agent.