AI policy advances globally; alliances surge

Anthropic / Claude ecosystem

No significant new developments.

Frontier model providers

No significant new developments.

AI developer tooling & infrastructure

AgentSwarms.fyi Offers Free Browser‑Based Multi‑Agent Orchestration Sandbox

A new platform provides a visual environment for designing and testing multi‑agent systems, complete with 60+ interactive notebooks. It aims to lower the barrier to experimenting with complex agentic workflows.

Cloud & platform providers

No significant new developments.

AI policy, regulation & governance

AI CEOs Urge Congress to Mandate DNA Screening for Synthetic Nucleic Acids

Sam Altman, Dario Amodei, Demis Hassabis and other AI leaders signed an open letter asking US lawmakers to require screening and recordkeeping for synthetic DNA/RNA purchases. The move aims to mitigate biosecurity risks as AI capabilities grow.

Italy Regulator Drops Investigation into Meta’s WhatsApp AI Bot

The Italian Competition Authority (AGCM) has closed its probe into Meta’s WhatsApp chatbot after the European Commission extended its investigation to include Italy, effectively resolving the case nationally.

Greece Prepares National AI Policy, Integrating EU AI Act

Greece plans to enact a domestic AI policy while implementing the EU AI Act, and will launch an AI innovation hub called Pharos. The move reflects a whole‑of‑government approach to AI governance.

Australian Employment Services Providers Submit AI Plans to DEWR

Multiple suppliers of Workforce Australia have filed AI deployment plans to the Department of Employment and Workplace Relations. No approvals have been granted after six months, highlighting regulatory review delays.

Industry & market moves

Google DeepMind and Wellcome Sanger Institute Launch AI Genomics Consortium

The two organisations have formed a consortium to create large genomic datasets for AI-driven biology research, aiming to accelerate the discovery of disease mechanisms and drug targets.

NTT DATA and Google Cloud Team Up to Scale Enterprise AI from Pilots to Production

The expanded alliance will use Google Cloud’s Gemini Enterprise to help businesses move AI projects from proof‑of‑concept to full production, combining NTT DATA’s system integration expertise.

Cegid Acquires Shine to Create Europe’s First AI‑Driven SMB Platform

The acquisition combines Cegid’s cloud‑native fintech with Shine’s financial services to build an AI‑powered platform for small businesses and accountants. The result is marketed as Europe’s first 100% cloud AI platform for SMBs.

OKI Partners with Lazarus AI for Defense‑Focused, High‑Reliability AI

Japanese manufacturer OKI and US‑based Lazarus AI will jointly develop on‑premises AI for defense and critical infrastructure. The solution leverages Lazarus’s source‑cited, document‑understanding technology.

VinDynamics and Skild AI Partner on Humanoid Robots

Vietnamese manufacturer VinDynamics and US startup Skild AI signed an MOU to co‑develop humanoid robots, combining Skild’s AI platform models with VinDynamics’ hardware expertise.

Temenos Acquires additiv to Bolster AI‑Powered Digital Banking

Temenos will absorb additiv’s wealth‑management orchestration platform, enhancing its AI‑driven banking and compliance services for financial institutions.

AI product & feature launches

Unisound U2 Agentic LLM Can Autonomously Execute 100‑Step Workflows

U2 is a native agentic large model designed for execution, able to decompose and complete complex real‑world tasks with over 100 steps without human intervention.

Eros Innovation Launches Cultural AI Platform in 34 Languages

The platform combines a Large Cultural Voice Model and Persona AI to deliver culturally nuanced interactions across 34 languages, targeting global content and communication industries.

Research with immediate practical relevance

DeepSeek‑V4‑Pro Post‑Trained on Huawei Ascend 910C, Boosting AI Independence

For the first time, a 1.6‑trillion‑parameter model, DeepSeek‑V4‑Pro, was post‑trained using Huawei’s Ascend 910C chips instead of NVIDIA GPUs. This marks a milestone in China’s pursuit of self‑reliant AI infrastructure.

Xiaomi MiMo Trillion‑Parameter Model Hits 1200 Tokens/s on Commodity GPUs

Xiaomi’s MiMo‑V2.5‑Pro‑UltraSpeed, a 1‑trillion‑parameter model, achieved inference speeds exceeding 1000 tokens per second on commodity GPUs, demonstrating hardware‑efficient scaling.