Sequoia-backed Empirik launches to predict infrastructure outages

AIR has raised $50 million to build a platform that discovers AI agents operating within companies, continuously monitors the skills and add-ons they use, and blocks unwanted behavior. The funding addresses a growing operational challenge as enterprises deploy multiple AI agents without full visibility into their capabilities and actions. The platform serves companies seeking to maintain control and security over AI agent deployments.

OpenAI has integrated ChatGPT Health with Epic, a major electronic health records system, allowing clinicians to import patient data directly into the platform. The integration provides read-only access to health records, enabling doctors to reference patient information within ChatGPT Health. This move expands the practical utility of ChatGPT Health in clinical workflows by connecting it to one of the most widely used EHR systems in healthcare.

Perplexity launched hybrid compute for its Computer platform, allowing a single AI agent to split work between cloud-based frontier models and locally-running open-weight models on Apple silicon Macs. Sensitive data is routed to the local machine via a trained PII classifier called a Privacy Gate, ensuring confidential information never leaves the device while the agent maintains task context. The feature is available today for enterprise customers and Pro/Max subscribers on macOS 15 or later.

NVIDIA and CrowdStrike announced SafeMind, an agentic cybersecurity system that pairs CrowdStrike's defensive models built on NVIDIA Nemotron with proprietary harnesses designed to automate threat response. The system operates in a continuous coevolution loop where offensive and defensive AI challenge each other to improve security posture. CrowdStrike also launched Falcon IQ for agentic workload automation and expanded its Guardian AI safety solution, addressing what CEO George Kurtz called a critical gap where attackers had frontier AI capabilities but defenders did not.