Agentic AI News

Guild.ai has raised $44 million in a combined seed and Series A round led by Khosla Ventures, with the company valued at $300 million. The funding is going toward what founder and CEO James Everingham describes as a platform that lets enterprise teams build, deploy, and manage AI agents at scale — in minutes rather than months. The announcement lands as part of a broader wave of agentic infrastructure investment. JetStream Security and WorkOS also announced funding rounds today, reinforcing that the agentic AI infrastructure layer is attracting serious capital right now. ...

The agentic commerce infrastructure stack just took a meaningful step forward. Klarna and Stripe today announced that Klarna’s Buy Now, Pay Later options will be supported through Stripe’s Shared Payment Tokens (SPTs) — meaning AI agents can now complete BNPL checkout flows on behalf of users without ever handling raw card details. Affirm also joined the SPT ecosystem on the same day, signaling that this isn’t a one-off integration — it’s the beginning of a standardized payment delegation layer for AI-driven shopping. ...

A critical vulnerability in ModelScope’s MS-Agent framework — now officially tracked as CVE-2026-2256 — allows an attacker to achieve full system compromise through code injection via an AI agent’s prompt pipeline. If you’re running MS-Agent v1.6.0rc1 or earlier in any deployment, this is a drop-everything patch situation. The vulnerability was disclosed today by multiple security outlets, with full CVE record details confirmed by SecurityWeek, GBHackers, CyberPress, and OffSeq Threat Radar. ...

Anthropic has quietly flipped one of the most-requested switches in AI-assisted development: you can now talk to Claude Code. As of today, Voice Mode is rolling out to Claude Code users — starting with roughly 5% of subscribers — letting developers speak their commands instead of type them. It’s a small percentage to start, but the implications are significant. Hands-free coding has gone from a demo concept to a live feature in one of the most-used AI coding tools in the world. ...

Nearly 70% of enterprises are already running AI agents in production. Another 23% plan to deploy them in 2026. And the vast majority of those agents are operating with no audit trail, no identity governance, and full access to the data they touch. Security analysts have a name for this: identity dark matter. The term comes from a Hacker News analysis published this week, and it’s earning traction because it captures something real. Like cosmological dark matter, AI agent identities exert enormous gravitational force on the systems around them — they make decisions, consume data, trigger actions — while remaining largely invisible to the tools and processes organizations use to manage access and risk. ...

Anthropic’s Claude went down twice in under 24 hours this week — and the developer community’s reaction tells a story about something bigger than a couple of bad server days. The second outage hit on March 3, investigation commencing at 03:15 UTC. It followed Monday’s first disruption, which Anthropic attributed to unprecedented demand. Chat, API, and Claude Code were all affected. Developers watched their pipelines stall, their autonomous agents go quiet, and their Claude Code sessions freeze mid-task — again. ...

OpenClaw just shipped v2026.3.2 — and it’s one of the more substantial point releases in recent memory. With a built-in PDF analysis tool, a new Speech-to-Text API, expanded credential management, and over 150 bug fixes, this update touches nearly every corner of the platform. There are also breaking changes to the HTTP Route Registration API that existing users need to know about before upgrading. Here’s what’s in the box. PDF Analysis Tool: Documents as First-Class Inputs The headline feature of v2026.3.2 is native PDF analysis. OpenClaw agents can now ingest PDF documents directly, with support for both Anthropic and Google backends. That dual-backend architecture matters: you can route PDF parsing to whichever model handles your document type best — Anthropic’s Claude for dense text and reasoning-heavy documents, Google’s multimodal stack for PDFs with heavy visual content like charts, diagrams, and scanned pages. ...

Something significant dropped in the open-source model space today: Alibaba’s Qwen3.5 Small series — a family of four on-device models ranging from 0.8B to 9B parameters — is now publicly available under the Apache 2.0 license. The headline claim from VentureBeat and confirmed by MarkTechPost: the 9B flagship outperforms OpenAI’s gpt-oss-120B on benchmarks, while running on a standard laptop. Let that land for a moment. A 9-billion-parameter model running on consumer hardware beats a 120-billion-parameter cloud model on capability benchmarks. If accurate — and the benchmark citations across multiple independent sources suggest it is — this is a meaningful moment for local and edge agentic deployments. ...

Two significant moves from Anthropic landed today, and they’re clearly designed to work together: Claude memory is now available on the free tier, and a new import tool lets users bring their ChatGPT and Gemini conversation history into Claude. Both features went live March 2, 2026 — confirmed across Engadget, 9to5Mac, MacRumors, AndroidHeadlines, and Bloomberg. If you’ve been evaluating Claude as a reasoning backbone for agentic workflows, this changes the accessibility calculus significantly. ...

OpenClaw 2026.3.1 is out, and it’s a meaningful update — not just a bug-fix release. Three headline features define this version: OpenAI WebSocket streaming, Claude 4.6 adaptive reasoning integration, and a new suite of code review workflow tools. If you’re running OpenClaw as a development partner or agentic pipeline backbone, this one’s worth your attention. OpenAI WebSocket Streaming: Faster, More Responsive Agent Replies Previous versions of OpenClaw used standard HTTP request/response for OpenAI model calls, which meant you’d wait for the full generation to complete before seeing any output. 2026.3.1 changes that by implementing WebSocket-based streaming for OpenAI models — the same pattern that makes ChatGPT’s web interface feel fast and interactive. ...