OpenAI acquires AI agent orchestration startup Ona

· Source: AI – SiliconANGLE · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Software Development & Engineering · Depth: Intermediate, quick

Summary

OpenAI Group PBC announced its acquisition of Ona, a startup providing a cloud-based platform for managing long-running artificial intelligence agents. Ona, officially Gitpod GmbH, addresses the limitation of local AI agents that cease operation when developer workstations are turned off, by enabling agents to run continuously in cloud-based sandboxes. This platform offers enhanced computing resources for faster task completion and improves security by promptly deleting sandboxes and preventing data theft risks associated with persistent local environments. Ona's software includes security guardrails, such as detecting obfuscated malicious applications using hashing and blocking access to sensitive file system credentials or malicious outbound connections. OpenAI plans to integrate Ona's technology to enhance its Codex AI assistant, which currently serves over 5 million weekly users, specifically improving its capability for long-running tasks that span hours or days and allowing user input during these processes. This acquisition follows OpenAI's purchase of Promptfoo Inc. three months prior, aimed at bolstering cybersecurity for its OpenAI Frontier platform.

Key takeaway

For AI Engineers developing long-running agent-based applications, OpenAI's acquisition of Ona signals a shift towards cloud-native, persistent execution environments. You should evaluate cloud sandbox solutions to ensure your agents can reliably complete multi-day tasks and benefit from enhanced security features like obfuscation detection and restricted file system access. This move underscores the importance of robust, secure infrastructure for scalable AI agent deployment.

Key insights

Cloud-based sandboxes enable persistent, secure, and resource-rich execution for long-running AI agents, overcoming local machine limitations.

Principles

Method

Ona's platform runs AI agents in cloud-based sandboxes, which remain online independently of developer workstations. It uses hashing for obfuscation detection and implements file system and outbound connection security blocks.

In practice

Topics

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, MLOps Engineer, Director of AI/ML

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI – SiliconANGLE.