A curated bulletin of AI development stories, sorted onto the AI Kaleido AI Intelligence Radar. Filter by vector to see where the field is moving.
The five vectors
Frontier Intelligence
LLMs → reasoning → coding → autonomous agents.
Agentic Enterprise
AI doesn't merely answer — it plans, executes and collaborates with other agents.
Physical AI
AI leaves the screen — robots, factories and autonomous systems.
AI Infrastructure
Compute → inference → simulation → AI factories → data centers.
AI Governance
The more autonomously agents act, the more containment and monitoring matter.
The AI story is shifting from “Who has the smartest model?” to “Who can build the most capable AI system around the model?”
- 01Physical AI
AI's Next Big Leap Is Into the Real World
WSJ
World models are pushing AI from language into robots, physical environments and embodied intelligence.
Read the WSJ article - 02Physical AI
Robots Poised for a "ChatGPT Moment"
Reuters
Unitree's CEO predicts a major breakthrough in embodied AI, where robots understand unfamiliar environments and execute tasks.
Read Reuters - 03Physical AI
ACE Robotics: Robot Brains Could Have a "ChatGPT Moment" by 2027
Reuters
Another signal that physical AI is the next frontier after LLMs — driven by multimodal perception, simulation and action planning.
Read Reuters - 04Physical AI
China's Robots Rock, Box and Mix Drinks
Financial Times
A strong look at China's accelerating humanoid robotics ecosystem and its manufacturing advantage.
Read the Financial Times article - 05Physical AI
Silicon Valley's Robot Boom
Business Insider
Robotics investment is exploding, but hardware and real-world training data remain the binding constraints.
Read Business Insider - 06AI Governance
AI Hasn't Gone Rogue. It's Worse Than That
Financial Times
AI agents are increasingly capable of autonomous cyber activity, including discovering and exploiting vulnerabilities during tests.
Read the Financial Times article - 07AI Governance
AI Firms Can't Yet Contain What They've Built
Reuters
A study of OpenAI, Anthropic, Meta, Google and xAI highlights weaknesses in containment, monitoring and safety protocols.
Read Reuters - 08AI Governance
OpenAI Hit the Brakes. Now What?
The Verge
A temporary deployment slowdown illustrates the tension between the frontier capability race and safety engineering.
Read The Verge - 09AI Governance
AI Safety Clock Moves Closer to Midnight
IMD
How agentic AI, autonomous systems and AI embedded in critical infrastructure are changing the risk profile.
Read IMD - 10Frontier Intelligence
Google's Latest AI Developments
Google
Three Gemini models designed for building AI agents at scale, plus Gemini Robotics ER 2.
Read Google's AI update - 11Agentic Enterprise
Google-Backed Agent2Agent Protocol Gets a New Home
Axios
A2A moves toward the Agentic AI Foundation, strengthening interoperability between agents — key for enterprise architecture.
Read Axios - 12Agentic Enterprise
How Enterprises Can Catch Up With the Rapid Pace of AI Advances
AI Business
Capability is advancing faster than most organisations can absorb it, creating an adoption and value-realisation gap.
Read AI Business - 13Agentic Enterprise
Build vs. Buy: The AI Agent Landscape for Businesses
AI Business
The emerging strategic decision: proprietary agents you build versus agent platforms you buy.
Read AI Business - 14Agentic Enterprise
IBM + OpenAI Partnership to Accelerate Enterprise AI
AI Business
An enterprise signal on bringing frontier models into finance, government and other large-scale environments.
Read AI Business - 15AI Infrastructure
NVIDIA GTC 2026: What's Next in AI
NVIDIA
NVIDIA's major developments across infrastructure, models, robotics and the full AI stack.
Read NVIDIA's GTC coverage - 16AI Infrastructure
NVIDIA Advances Agentic and Physical AI
NVIDIA
The convergence of simulation, generative AI, agents and robotics into one stack.
Read NVIDIA - 17AI Infrastructure
NVIDIA and Japan Bring Full-Stack AI and Robotics to Industry
NVIDIA
AI moving down the stack into manufacturing, robotics and industrial systems.
Read NVIDIA - 18Frontier Intelligence
OpenAI: Latest News and Insights
Computerworld
A rolling source for tracking OpenAI's product and organisational developments.
Follow Computerworld's OpenAI coverage - 19Frontier Intelligence
AI Scientists: The Next Frontier
arXiv
AI systems that form hypotheses, design and run experiments, interpret results and revise beliefs.
Read the research paper - 20Frontier Intelligence
From BERT to Frontier Agents
arXiv
A research overview of the progression from language models to task-solving agents — including the collapse in capability-cost ratios.
Read the research paper - 21AI Infrastructure
The Inference Cost Curve Keeps Falling
Industry analysis
Cost per unit of capability keeps dropping, which quietly turns yesterday's pilot economics into today's production business case.
Explore the analysis - 22Agentic Enterprise
Model Context Protocol Becomes the Plumbing of Agents
Ecosystem
MCP is emerging as the standard way agents reach tools and data, complementing A2A on the agent-to-agent side.
Explore MCP - 23AI Governance
EU AI Act Implementation Enters the Hard Phase
Policy
Obligations for general-purpose and high-risk systems turn governance from a slide into an operating requirement.
Track the rollout