NerdNews

Microsoft CEO Satya Nadella highlighted the potential of artificial intelligence in Microsoft 365, showcasing how AI agents can be used to build powerful knowledge apps and applications. He discussed the power of AI agents that act as data scientists, using Agent Mode in Excel and PowerPoint to find insights. Nadella also introduced Microsoft Fabric, which ingests data from Microsoft 365, and Microsoft Foundry, a cloud-native platform for building agentic AI systems.

MatX, a chip startup founded by former Google hardware engineers, has raised $500 million to develop AI chips that can train LLMs 10 times better than Nvidia's GPUs. The funding will help MatX produce its chips with TSMC, with plans to start shipping them in 2027.

The Pentagon is escalating a dispute with Anthropic, demanding unrestricted access to its AI model. Anthropic has refused, citing concerns over mass surveillance and autonomous weapons. The Pentagon may invoke the Defense Production Act to force Anthropic to comply, marking a significant expansion of the law's use.

Meta plans to purchase up to $100 billion worth of AMD chips to drive its AI ambitions, including 'personal superintelligence'. The deal includes AMD's MI540 series of GPUs and latest generation of CPUs, and a performance-based warrant for up to 160 million shares of AMD common stock.

Anthropic's Claude Cowork platform now connects with Google Workspace, Docusign, and WordPress, automating tasks in HR, design, engineering, and finance. The update also enables Claude to handle multi-step tasks across Excel and PowerPoint.

Google announced a new 1.9GW clean energy deal, including a 300-megawatt battery made by Form Energy, to power its new data center in Minnesota. The battery can deliver its rated power for 100 hours, making it the largest in the world. The deal also introduces a new utility fee structure to help utilities adopt clean technologies.

Microsoft is expanding data loss prevention (DLP) controls to block the Microsoft 365 Copilot AI assistant from processing confidential documents, regardless of their location. This change will be deployed through the Augmentation Loop (AugLoop) Office component between late March and late April 2026.

CISOs must add intent to the equation for AI security, as traditional identity and access management is no longer sufficient. AI agents require unique identities, defined roles, and intent-based permissions to ensure secure access and prevent privilege inheritance and mission drift.

SolarWinds has released security updates to patch four critical Serv-U remote code execution vulnerabilities that could grant attackers root access to unpatched servers. The most severe flaw, tracked as CVE-2025-40538, allows attackers with high privileges to gain root or admin permissions on vulnerable servers.

Anthropic's Claude Code Security can find software vulnerabilities, but security researchers say that finding bugs is not enough, as validation and patching are still major challenges. The company's AI model found over 500 vulnerabilities, but only a few were fixed, highlighting the need for a more comprehensive approach to cybersecurity.

Google has added a new feature to its Opal app, allowing users to create automated workflows using text prompts. The feature uses the Gemini 3 Flash model and can execute tasks such as maintaining memory across sessions. This update enables non-technical users to build complex workflows within their apps.

Oura has launched a proprietary AI model to deliver personalized insights around women's health, supporting questions from menstrual cycles to menopause. The model is hosted on Oura-controlled infrastructure and conversations are not shared or sold.

Waymo has expanded its robotaxi services to 10 US cities, including Dallas, Houston, San Antonio, and Orlando. The company plans to launch services in several more cities this year, aiming to serve over 1 million rides per week by the end of the year.

A US District Judge ruled in favor of OpenAI, dismissing xAI's trade secrets lawsuit. The judge found that xAI did not provide sufficient evidence of misconduct by OpenAI. The lawsuit alleged that OpenAI had poached xAI employees and stolen trade secrets, but the judge determined that there was no indication of illegal behavior by OpenAI.

SambaNova has raised $350 million to advance its dataflow architecture, an alternative to GPU-based AI systems. Intel Capital is one of the investors, and the two companies will collaborate to provide customers with an alternative to GPUs for generative AI deployments. SambaNova's new RDUs will use Xeon CPUs and feature a three-tier memory hierarchy, allowing for efficient model swapping and key-value cache offloading.

India has become the world's largest market for generative AI app downloads, with a 207% year-over-year growth. Companies like OpenAI, Google, and Perplexity have rolled out free premium offers to accelerate user growth. However, the country still generates a disproportionately small share of AI app revenue, highlighting the monetization challenge.

Multiverse Computing, a Spanish startup, has released a free compressed AI model called HyperNova 60B, which is smaller and more efficient than traditional models. The model is available on Hugging Face and boasts lower memory usage and latency. This development could help companies deploy AI models more affordably and support a wider range of use cases.

Peter Williams, former L3Harris Trenchant boss, sentenced to 7 years for selling hacking tools to Russian broker, potentially accessing millions of computers and devices

Marquis is suing SonicWall, alleging that the firewall provider's security failings led to a ransomware attack on its network. The lawsuit claims that SonicWall's breach exposed critical security information, allowing hackers to steal sensitive data, including customer names, dates of birth, and financial information.

High-performing software teams thrive in a generative culture with trust, learning, and customer closeness. Platform engineering supports engineers by removing repetitive work, and good tech leaders amplify performance by improving systems and providing clear context. Investing in developer experience increases productivity and quality by enabling autonomy, mastery, and purpose.