The Informed Layer · Newsscan
Daily briefing

Export ban on Anthropic; agents go enterprise

← All issues

Anthropic / Claude ecosystem

Claude Code v2.1.178 Adds Parameter-Level Permissions and Nested Directory Support

The latest Claude Code update introduces fine-grained permission syntax (Tool(param:value)) and support for nested .claude directories, along with bug fixes for sub-agents and authentication.

Anthropic Rolls Out Multilingual Voice Mode for Claude Apps

Claude now supports voice interactions in 18 new languages with real-time language switching on mobile apps, expanding its accessibility for global users.

Anthropic Pauses Claude Agent SDK Subscription Change on Implementation Day

Anthropic unexpectedly paused a planned billing change for the Claude Agent SDK on the day it was to take effect, affecting third-party integrations and subscriber costs.

Frontier model providers

OpenAI Expands Codex in Europe with Computer Use and Memory Features

OpenAI rolled out Codex desktop capabilities including browser automation via Computer Use and cross-session Memory to European users, powered by GPT-5.3-Codex-Spark.

AI developer tooling & infrastructure

Databricks Open-Sources Omnigent for AI Agent Orchestration with OS-Level Policy Enforcement

Databricks released Omnigent, an open-source control plane that enforces policies at the OS level before AI agent actions, addressing prompt- and model-variability weaknesses.

Cloud & platform providers

Google Cloud Unveils Data Agents Across the Agentic Data Cloud

Google Cloud launched new data agents and tools that ground AI in enterprise data with near-100% accuracy and unified governance, enabling proactive intelligence in analytics.

AWS Summit New York Preview: Kiro AI Agent IDE and AgentCore to Be Announced

Ahead of the summit, AWS is expected to reveal Kiro, a spec-driven AI coding agent replacing Q Developer, along with AgentCore for Bedrock, shifting development to agentic workflows.

AWS Brings Compute, Storage, and AI Services to Athens with New Local Zone

AWS announced an Athens Local Zone opening July 2026, offering in-country data processing and low-latency access for AI and other services, supporting data residency requirements.

AI policy, regulation & governance

Anthropic Cyber Red Team Meets Trump Officials Over Model Weaponization

Anthropic's cybersecurity red team is meeting with Trump administration officials to demonstrate that its AI models cannot be weaponized by foreign threats, following a U.S. export ban on its latest models. The discussions aim to influence the scope of the export controls.

Cohere Sees 'Huge' Inbound After US Blocks Anthropic's Latest AI Models

Cohere reports a significant increase in interest from global customers and governments following US export controls on Anthropic's Fable 5 and Mythos 5, positioning itself as a sovereign alternative.

France Drops Palantir for Local Rival, Deploys Mistral AI Assistant to All Civil Servants

France's security services replaced Palantir with ChapsVision and rolled out a Mistral-powered AI assistant across the entire civil service, marking a shift to sovereign AI.

New Policy Tool Assesses Australia's AI Strengths and Gaps Beyond Sovereignty

A new 'AI Agency Tool' framework was launched to measure national AI capability beyond traditional sovereignty metrics, highlighting areas for strategic investment.

Industry & market moves

Ineffable Intelligence Selects Google Cloud to Power Its Superintelligence Mission

Ineffable Intelligence, a frontier AI lab led by AlphaGo’s David Silver, chose Google Cloud as its infrastructure backbone, deploying one of the largest A5X GPU clusters for a 'superlearner' project.

SpaceX to Acquire AI Coding Platform Cursor for $60 Billion in Stock

SpaceX announced plans to acquire Anysphere (Cursor) for $60B in an all-stock deal, merging its xAI division with the popular AI developer tool to compete with Anthropic and OpenAI in coding.

Informatica Expands AI Data Partnerships Across AWS, Azure, Google Cloud, Databricks, and Snowflake

Informatica extended its headless Intelligent Data Management Cloud to major cloud platforms, enabling unified governance and quality for agentic AI workflows.

Nebius Closes $643 Million Acquisition of Eigen AI to Enhance Token Factory Platform

Nebius completed its acquisition of Eigen AI, adding the company's AI inference optimization stack to its Token Factory cloud platform.

Genesis AI and LG CNS Partner to Scale General-Purpose Robotics Across Enterprises

Genesis AI and LG CNS announced a strategic partnership to deploy general-purpose robotics across enterprise operations with a joint playbook and multi-phase rollout.

AI product & feature launches

NASA and Loft Orbital Run Gemma 3 AI Model in Orbit to Summarize Earth Data

NASA and Loft Orbital successfully deployed Google's Gemma 3 model on a satellite to perform on-orbit summarization of Earth observation data, reducing data transmission needs.

Databricks Summit Day 2: Agentic AI and Catalog Federation Go Enterprise

At its summit, Databricks showcased federation of Unity Catalog with AWS Glue and Bedrock-driven agents, enabling governed, zero-copy analytics across data sources.

Digimarc Adds AI Provenance Tools to LangChain, Copilot, and Other Platforms

Digimarc introduced crypto-backed provenance and verification capabilities natively integrated into major AI agent platforms, enhancing trust and auditability.

Ping Identity Extends Runtime Identity for AI Agents Across AWS, Google Cloud, and Cloudflare

Ping Identity expanded its Runtime Identity solution to major cloud and edge platforms, enabling secure AI agent deployment and operation with continuous authentication.

Tencent Cloud and Inworld AI Partner on Real-Time Voice AI Solution

Tencent Cloud and Inworld AI announced a strategic partnership to deliver a lifelike, real-time voice AI solution integrated with Tencent Cloud RTC for immersive experiences.

Shifters to Debut AI-Native Ground Autonomy Ecosystem at Eurosatory 2026

Shifters will showcase its full autonomous robotics ecosystem (RITA, ARENA, TRUST) for ground operations, designed for defense and security applications.

Alibaba's Qwen-RobotSuite Releases Three Embodied AI Models for Manipulation, World Modeling, and Navigation

Alibaba open-sourced Qwen-RobotManip, Qwen-RobotWorld, and Qwen-RobotNav, a suite of foundation models for robotic manipulation, video world modeling, and navigation.

Verge Labs Unveils vBx Foundation Model for Precision Neurology

Verge Labs released vBx-1.0, a foundation model claimed to dramatically outperform prior methods in precision neurology, and is previewing with partner drug developers.

SiMa.ai Launches Palette Neat, First Agentic Environment for Physical AI

SiMa.ai introduced Palette Neat, an agentic development environment that reduces physical AI model development from months to days, targeting robotics and edge devices.

Genesis AI Unveils Eno, a Non-Humanoid General-Purpose Robot

French startup Genesis AI launched Eno, a wheeled general-purpose robot powered by its GENE foundation model, targeting both industrial and consumer deployments by late 2026.

Nvidia Powers Open-Source Humanoid Robot to Learn Movement in 7 Days via Simulation

Nvidia's Isaac platform enabled the ROBOTIS AI Sapiens humanoid to learn walking from smartphone video in just one week through simulation-first training, democratizing robotic motion data.

Databricks Launches Genie One Agentic Coworker Fueled by Business Data

Databricks released Genie One, an AI agent that grounds responses in enterprise data across tools, enabling automated actions and cross-functional use.

Okta Expands Google Cloud Partnership to Secure AI Agents and the Browser

Okta integrated its identity layer with Google Cloud to provide centralized visibility, policy control, and real-time authentication for AI agents and Chrome Enterprise.

CrowdStrike Launches Identity Controls for AI Agents with Continuous Authorization

CrowdStrike introduced Continuous Identity for AI Agents, providing real-time access governance across on-premises, cloud, and SaaS environments for AI agents.

Sakana AI Commercializes Marlin Agent for 100-Page Research Reports with Slides

Sakana AI launched Sakana Marlin, an enterprise agent that uses AB-MCTS to autonomously generate 60–100 page research reports with slide decks in a single session, acting as a virtual CSO.

Research with immediate practical relevance

ACE Robotics' Kairos World Model Leads Four Global Embodied AI Benchmarks

ACE Robotics' Kairos world model achieved top results on four international embodied intelligence benchmarks, surpassing existing models in physical-world modeling and generalization.