Latest AI (Artificial Intelligence) News
Anthropic’s Claude Mythos Clears UK AI Security Institute Cyber Threshold
The UK’s AI Security Institute revealed Anthropic’s Claude Mythos Preview as the first model to pass the 32-step 'The Last Ones' corporate-network simulation, enabling full domain takeover in offensive cyber operations. This marks frontier AI crossing into real-world cyber threats, typically requiring 20 hours of human red-teaming.
Concerns over such models are prompting U.S. CISA to consider slashing vulnerability fix deadlines from weeks to three days.
China Releases Four Open-Weights Coding Models Achieving Parity
Chinese labs including Z.ai, MiniMax, Moonshot, and DeepSeek launched GLM-5.1, M2.7, Kimi K2.6, and V4 within 12 days, reaching agentic engineering parity at lower inference costs than Western frontiers. DeepSeek V4-Pro claims parity with Opus 4.6 and GPT-5.4 on their benchmarks, though NIST CAISI shows an eight-month lag on cross-domain tests.
This sprint breaks China's previous lag in coding AI development.
SpaceX Places $60B Buyout Option on AI Coding Firm Cursor
SpaceX preempted Cursor’s $2B fundraise with a $60B buyout option or $10B for an AI collaboration, deferred until after its summer IPO. This massive deal highlights AI's strategic value in space and engineering applications.
Cursor's AI coding tools are seen as critical for SpaceX's ambitions.
CISA Considers Three-Day Deadlines for Fixing AI-Exploitable Vulnerabilities
U.S. CISA is weighing cuts to vulnerability fix deadlines from weeks to three days due to AI models like Anthropic’s Mythos and OpenAI’s GPT-5.4-Cyber enabling rapid hacker exploits. Previously taking months, AI compresses exploit timelines to hours.
This could model responses for states, businesses, and beyond.
Profluent Partners with Lilly on $2.25B AI Bio Therapeutics
Frontier AI for biology firm Profluent announced a $2.25B partnership with Lilly for large-gene insertion therapeutics. This deal underscores AI's growing role in biotech drug development.
It highlights accelerating investments in AI-driven medical innovations.
Skild AI Acquires Zebra’s Robotics for End-to-End Warehouse Stack
Skild AI bought Zebra Technologies’ Robotics Automation business, including Symmetry Fulfillment and Fetch Robotics. The acquisition creates the first AI-native end-to-end warehouse automation with humanoids, AMRs, arms, and orchestration.
This positions Skild as a leader in embodied AI for logistics.
Sereact Closes $110M Series B for Embodied AI
Embodied AI startup Sereact secured a $110M Series B funding round. The investment supports advancements in physical AI applications.
It reflects surging venture interest in robotics and real-world AI agents.
Microsoft and OpenAI Reportedly Reset Partnership
April 2026 saw a reset in the Microsoft-OpenAI relationship amid evolving AI industry dynamics. Details point to strategic realignments in compute, models, and investments.
This follows $122B funding pursuits for OpenAI.
AI Agents Begin Interacting with Real Markets
AI agents are now meeting real markets, transitioning from simulations to live economic interactions. This development signals practical deployment in trading and automation.
It raises implications for market efficiency and regulation.
Singapore Chip Firms Expand to US Amid AI Boom
Singaporean chip companies are moving operations to the U.S. driven by explosive AI demand. This shift aims to tap into North American markets and supply chains.
It reflects global realignment in AI hardware production.
China Court Rules AI Cannot Replace Workers Without Consequence
A Chinese court decision mandates consequences for AI replacing human workers, addressing job displacement. The ruling sets precedents for labor protections in the AI era.
It balances technological adoption with social impacts.