SearchLeak Shows How Microsoft 365 Copilot Search Can Become a Data Leak
Varonis disclosed SearchLeak, a patched Microsoft 365 Copilot Enterprise Search vulnerability chain that could turn one trusted-looking Microsoft link into a path for stealing emails, files, calendar data, and MFA codes.
Dream’s $260M Round Turns Sovereign AI Into Cyber Infrastructure
Dream raised $260 million at a $3 billion valuation for sovereign AI and cyber defense systems built for governments. The deal shows how national AI infrastructure and critical-infrastructure security are becoming the same market.
SandboxAQ’s $500M CHIPS Award Moves AI Into Chip Materials
The Commerce Department awarded SandboxAQ $500 million to use physics-based AI for semiconductor materials discovery. The target is not a new chip, but the chemicals, magnets, catalysts, and backup-power systems that fabs need to run.
Apple-Intel Chip Claim Puts Intel Foundry’s 18A-P Test in Public View
President Trump says Apple has agreed to work with Intel on U.S.-made chips, but Apple and Intel have not announced a signed foundry deal. The real test is whether Intel’s 18A-P process can move from risk production to Apple-grade yield, volume, and product trust.
ASML Denies EUV Machine Reached China as U.S. Chip Controls Tighten
U.S. officials reportedly raised concerns that ASML EUV-related equipment may have reached China, a claim ASML denies. The dispute shows why chipmaking tools, servicing, and allied export controls are becoming central to the AI hardware race.
Meta Expands Teen Accounts With AI Age Checks and Parent Alerts
Meta is rolling out 13+ Teen Account defaults across Instagram, Facebook, and Messenger while using AI age assurance to find younger users who list adult birthdays. The update shows how teen safety is becoming a platform enforcement system, not just a parental-control setting.
Microsoft AutoJack Research Shows How AI Browsing Agents Can Break Localhost Trust
Microsoft’s AutoJack research shows how an AI browsing agent could turn a malicious webpage into a local remote-code-execution path through AutoGen Studio’s MCP WebSocket surface. The specific issue was fixed before a PyPI release, but the localhost trust problem is bigger than one tool.
Google DeepMind’s AI Control Roadmap Treats Agents Like Insider Threats
Google DeepMind released an AI Control Roadmap for securing powerful internal AI agents. The plan borrows from cybersecurity, maps rogue-agent tactics to a MITRE ATT&CK-style taxonomy, and lays out detection and response tiers for systems that may soon act faster than human reviewers can supervise.
FERC Gives AI Data Centers a Faster Path to the Grid
FERC ordered six regional grid operators to justify or revise their rules for connecting data centers and other large power users. The move could speed AI infrastructure projects, but it puts cost allocation, flexible loads, and state oversight under a sharper deadline.
OpenAI o3 Helps Doctors Revisit Rare Disease Cases in NEJM AI Study
Researchers at Boston Children’s, Harvard, and OpenAI used o3 Deep Research to reanalyze 376 previously unsolved rare disease cases. The model surfaced evidence-linked leads that helped specialists confirm 18 diagnoses, but the study is careful about what AI did and did not decide.