Prime Intellect raises a $130M Series A at a $1B valuation, with NVIDIA, Intel, and Dell on the cap table and $100M in annualized revenue across 6,000 customers.

Wiz discloses GhostApproval, a trust-boundary flaw in six AI coding assistants that lets a malicious repo write outside the sandbox via symlinks. AWS, Cursor, Google patched; Anthropic, Augment, Windsurf have not.

Codenotary shipped AgentMon 3, an enterprise AI runtime security platform that replaces static allow-lists with self-refining policies learned from more than five million AI agent interactions per day.

OpenAI and Anthropic are pulling in opposite directions, with hiring data framing the gap as a new agent-native security risk called belief injection that current SIEM and EDR tools cannot see.

Samsung began mass production of the PM1763, a PCIe 6.0 SSD with 28.4 GB/s read speeds and 1.4-second LLM load times, built on 9th-gen V-NAND and a 4-nanometer controller, with PQC and TDISP support.

Keeper Security extended agentic AI governance to Keeper Endpoint Privilege Manager on July 7, 2026, enforcing the same OS-level policy framework the company already applies to human users.

Nscale closed a $900M revolving credit facility led by J.P. Morgan, Goldman Sachs, and Morgan Stanley. The revolver funds AI data center expansion across the US, Europe, and APAC for the Nvidia-backed UK operator.

Couchbase launched its AI Data Plane today, framing the product as the operational data layer for production AI agents, with Agent Memory, an Agent Catalog, an MCP server, and Iceberg federation.

Radware extends Agentic AI Protection to cover Anthropic's Claude Code on developer endpoints, with audit-ready reports mapped to ISO 42001, the EU AI Act, and the NIST AI Risk Management Framework.