OpenAI acquires Ona to complement the intelligent cloud base for enterprise-level AI applications, shifting the narrative from large models to AI agents in execution.

Global AI leader OpenAI has agreed to acquire Ona, a well-known startup that provides cloud computing services to support the deployment of artificial intelligence agents for enterprises. This move can be seen as the top global AI developer with ChatGPT making efforts to enhance the efficiency of enterprise operations with cutting-edge AI technology and help enterprises deploy AI agents more conveniently. The key to OpenAI's acquisition of Ona is not just to expand the Codex team, but to completely fill the gap in the safe, reliable, and governable cloud execution capability required to advance AI agents into enterprise production and operational environments. From AI chatbots like Siasun Robot&Automation to enterprise AI agent execution platforms, OpenAI's decision to acquire Ona at this moment undoubtedly bets on AI agents becoming the next generation productivity platform. This deal has not been completed yet and will formally integrate the Ona team into OpenAI's Codex project; this ChatGPT manufacturer announced this news in a major way on Thursday EST. OpenAI stated in a statement that currently over 5 million people are using its Codex AI programming tool every week. For OpenAI's ambitious growth plan, one of the significances of acquiring Ona is to expand Codex from a developer tool to an essential enterprise knowledge work operation process. OpenAI disclosed that Codex's weekly users have exceeded 5 million, a 400% increase from earlier this year; at the same time, Codex is no longer just a software development tool, non-software platform developer users including analysts, marketers, operations personnel, designers, researchers, institutional investors, and senior bankers, account for about 20% of the overall users and their growth rate is more than three times that of the developer community. OpenAI stated that Ona's cloud computing-related services provide a "secure, reliable developer environment where AI agents can securely and efficiently access the essential tools, systems, and context needed as they progress over time, to carry out the agent-style AI workflows". The terms of the Ona acquisition deal were not disclosed. From "How powerful are AI large models" to "Can AI agents really work tasks" For the valuation prospects of the global AI leader OpenAI, this acquisition strengthens OpenAI's narrative of accelerating the transformation from "consumer-level AI entry" to "enterprise-level AI productivity platform". OpenAI is packaging its "AI large model capabilities' lead" as "AI large models' landing ability in enterprise production and operational systems". In the competition with rivals like Anthropic for the enterprise AI entry, the one who can safely run AI agents in enterprise production environments for the long term is more likely to gain higher enterprise valuation multiples and stronger revenue visibility. OpenAI confirmed the submission of a confidential S-1 filing on June 8th, but has not yet decided on the timing of the IPO, which may take some time because some things are easier to advance as a private company. OpenAI is currently engaged in fierce competition with long-time competitor Anthropic PBC to accelerate the development and sale of more advanced AI agent systems that help enterprise customers simplify workflows significantly and significantly improve operational efficiency. Both of these top global AI startups focusing on cutting-edge AI application technology have submitted IPO draft registration statements and are aiming to make their public debut on Wall Street as early as this fall. For investors, OpenAI swallowing Ona is enough to indicate that the main investment theme of AI applications is moving from "whose AI large model is smarter" to a new stage of "whose AI agents can truly enter enterprises, perform tasks, interact with systems, be regulated, and be auditable"; This means that in the future, the valuation premium of the AI industry chain will increasingly flow towards AI agent cloud computing infrastructure, enterprise workflow orchestration systems, secure cloud execution, developer tools, and high-compliance industry landing capabilities. OpenAI acquires Ona: Filling in the "AI agent cloud pedestal" before the major IPO The core logic of OpenAI's acquisition of Ona is not simply "buying a cloud service startup", but filling in the most critical infrastructure for enterprise AI agents before the IPO window: a secure, reliable, and governable cloud execution environment. OpenAI officially stated that Ona will integrate secure cloud execution and orchestration technologies into the rapidly expanding Codex ecosystem, with the goal of upgrading Codex from a "single-session AI programming assistant on a single device" to an AI agent-style workflow platform that can run continuously for several hours or even days in enterprise cloud environments; public information shows that Ona's exclusive technology can provide a secure, persistent environment that allows AI agents to continuously access the tools, systems, and context needed to complete long-term tasks. From a product engineering perspective, this acquisition solves the "last mile" problem of commercializing AI agents: it's not enough for models to reason and write code, what enterprises really need is for AI agents to execute real tasks in an environment with controlled permissions, isolated credentials, audit trails, reversible processes, and auditable results. OpenAI explicitly mentioned that Ona's controllable execution models will allow AI agents to run in the enterprise's own cloud environment while OpenAI provides intelligence and orchestration capabilities, allowing enterprises to deploy Codex without sacrificing infrastructure, data, and control over security boundaries. This is particularly crucial for highly compliant industries such as finance, manufacturing, healthcare, software engineering, legal, investment research, as these clients are not purchasing "AI large models that can chat," but AI agents that can enter production processes, carry out continuous work, and meet governance requirements. In other words, Ona does not simply enhance "coding", but pushes Codex towards more generalized enterprise automation: running tests, fixing bugs, handling work orders, generating reports, building dashboards, modernizing legacy systems, reviewing code, and connecting with enterprise high-frequency applications like Slack/Google Docs/Snowflake/Tableau. OpenAI has also strengthened its cooperation with Oracle recently, allowing OCI customers to use OpenAI models and Codex through existing cloud commitments, further demonstrating its strategy to reduce friction in enterprise procurement, deployment, and governance. OpenAI is currently preparing for a major IPO with a potential valuation of up to $1 trillion. Although the company is expected to achieve an annualized revenue run rate of around $20 billion by the end of the year, losses are also widening; this means that the public market will not only be concerned with revenue growth rates in the future, but also revenue quality, enterprise customer stickiness, unit economic models, compute costs control, and sustainable gross profit margins. Ona serves these valuation variables: allowing Codex to penetrate deeper into enterprise workflows, enhancing retention rates, customer value per transaction, and long-term contract value, and shifting OpenAI's revenue structure from "subscription/API calls" to a "continuous execution-based AI agent platform".