TrueFoundry Bets Enterprises Will Own Their Agent Runtime

TrueFoundry open sourced TrueForge, an agent harness aimed at Claude Managed Agents. Its published benchmark shows 30% to 75% savings, not the advertised 50% TCO cut.

Agent2Agent Joins The Agentic AI Foundation Alongside MCP

Agent2Agent is moving into the Agentic AI Foundation, joining MCP under one Linux Foundation umbrella. What the governance change does and does not do for enterprises.

Stripe Bets Over $8 Billion On OpenRouter’s AI Model Traffic

Stripe has agreed to buy OpenRouter for more than $8 billion, buying distribution and inference demand rather than metering tech. The neutrality question comes with it.

Open Models Have A Distribution Layer Enterprises Rarely Track

Hugging Face counts 28,531 community GGUF conversions of Qwen models and 54 from Qwen itself. The file that runs is rarely the one the lab tested or signed.

Dynatrace Pays $915 Million To Move AI Evaluation Upstream

Dynatrace already shipped LLM-as-a-judge evaluation. Its $915 million Arize deal buys something else, a position with AI engineers before an app reaches production.

Municipal Water System Attacks Spotlight Insecure Industrial Systems

Cyberattacks are sobering reminders of vulnerabilities common in industrial automation, but there are steps the affected organizations can take to secure their systems.

Is Agentic AI Pricing Getting Better? What’s Coming Next

Pricing for enterprise agentic AI continues to be too complex, but the author has ideas about likely forthcoming improvements — and what enterprises should do now.

Trends In Enterprise AI Success: Proprietary Data And Open Models

In AI deployments, the largest enterprises are interested in protecting their proprietary data and favoring open AI models with proprietary software customization.

Arcade Acquires Smithery To Own The Agent Tool Supply Chain

Arcade.dev has acquired Smithery, the MCP registry developers default to. The deal signals discovery and governance of agent tools consolidating into one procured layer.

OpenAI Paused Astra Over Cybersecurity Fears. AI Hacking Is Here To Stay.

OpenAI’s decision to slow work on Astra over cybersecurity concerns signals a new era of AI hacking, where autonomous systems could discover vulnerabilities, exploit targets and fundamentally change the economics of cyberattacks.