OpenAI acquires Ona: Codex enters the enterprise

The acquisition timeline: what we know so far

On June 11, 2026, OpenAI published on its official website the announcement of the Ona acquisition . The financial terms of the deal were not disclosed. However, the communication was clear on the technology and strategy front.

Ona is a startup focused on creating secure, isolated, and persistent cloud environments. Basically, they are computational sandboxes capable of maintaining state across different sessions. This is exactly what is needed to run AI agents on long-running enterprise tasks.

The integration is aimed at Codex, OpenAI's model built for generating and reviewing code. Still, the impact goes way beyond coding. So, it's definitely worth looking at this move with the right strategic mindset.

The technical sticking point: why persistence changes everything

Until now, one of the structural limitations of AI agents was the lack of continuity. An agent could execute a task, but it lost the context in the next session. As a result, complex workflows required constant human supervision or expensive workaround architectures.

Ona's persistent cloud environments solve this problem at the root. The agent keeps operational memory, files, state variables, and logs between runs. Plus, secure isolation ensures every company workspace stays totally separate.

In practical terms, this means an AI agent can now start a data analysis process, pause it, and resume it hours later right where it left off. For enterprise companies, this is a game-changer. In fact, it paves the way for operational delegations that were simply unfeasible before.

According to the analyses of Gartner on Agentic AI , by 2028 over 30% of complex business workflows will be handled by autonomous agents. Environment persistence is one of the key technical prerequisites to reach that threshold.

Winners and losers in the enterprise AI ecosystem

Every acquisition reshapes market balances. This one is no exception. Therefore, it's helpful to spot who wins and who loses ground.

OpenAI reinforces its competitive advantage. With Ona integrated into Codex, OpenAI offers an end-to-end infrastructure for enterprise agents. This reduces the need for third-party solutions for environment management. Furthermore, it strengthens its positioning against Microsoft Copilot, Google Gemini, and Anthropic Claude, all active in the same segment.

Cloud infrastructure vendors are watching closely. AWS, Google Cloud, and Azure currently provide the foundations these architectures run on. However, if OpenAI brings the management of persistent environments in-house, it could reduce its reliance on certain third-party cloud services. On the other hand, it might also increase compute consumption, generating more revenue for cloud providers.

Competing startups in the agentic AI segment are feeling the pressure. Replit, E2B, and other players offering execution environments for agents see their competitive space shrinking. Despite this, they remain relevant for those who do not want to depend on the OpenAI ecosystem.

For a deeper dive into the competitive dynamics in the sector, it is worth checking out the analysis by MIT Technology Review on AI agents in enterprise settings .

What the press releases don't say

The official announcement is measured and technical. However, some interesting tensions emerge between the lines.

First and foremost, the security issue. Persistent cloud environments hosting AI agents with access to business systems are an attractive target for cyberattacks. OpenAI explicitly cites security as a priority, but the implementation details are not yet public. Therefore, companies considering adoption will need to demand transparency on this point.

Next up is the governance issue. Who checks what a persistent agent is doing? How do you track actions done on their own? These things are super important for compliance, especially regarding the European AI Act. In fact, Italian small and medium-sized businesses working in regulated fields will have to deal with these hurdles before they start using agentic setups.

Finally, there is the issue of vendor lock-in. Integrating critical workflows into OpenAI proprietary environments creates a structural dependency. Therefore, a multi-vendor strategy remains advisable even in this scenario.

SHM Studio's take: what this means for Italian SMEs

We at SHM Studio we work daily with Italian B2B and retail SMEs. The question we get most often lately is about how practical AI agents really are: when will they actually be usable in real business settings?

The acquisition of Ona is a clear signal: that threshold is getting closer. However, it is not today yet. The technology is maturing, and the integration between Ona and Codex will require months of work before being stably available for enterprise customers.

That being said, the time to prepare is now. Companies that reach the adoption of persistent agents without a solid digital foundation will find it much harder. In particular, those who haven't optimized their own yet web asset or has not invested in a strategy structured SEO will find itself competing at a disadvantage.

Similarly, companies that currently use tools from Digital marketing in a fragmented way will struggle to integrate AI agents into their processes. Agents work best when they operate on clean data, defined flows, and measurable goals.

Next moves: strategic actions to consider

Based on this acquisition, it is possible to identify some concrete operational directions for SMEs.

• Map out workflow candidates for agentic automation. Not all processes benefit from AI agents. The best candidates are repetitive, multi-step tasks with structured inputs. For example, ad campaign management, SEO monitoring, or periodic reporting.
• Assess the maturity of your data. Persistent agents operate on company data. Therefore, the quality and organization of internal data is a prerequisite. Those who have already implemented a structured CRM or an advanced analytics system have the upper hand.
• Monitor the evolution of Codex enterprise. OpenAI will roll out progressive updates. Following the official roadmap lets you plan adoption ahead of time, avoiding rushed decisions.
• Consider a multivendor strategy. Relying exclusively on the OpenAI ecosystem brings lock-in risks. A comparative evaluation with alternative solutions is still recommended.
• Invest in internal training. AI agents do not replace human skills: they amplify them. SMEs that train their teams on using AI tools will achieve better results than those relying solely on technology.

For those managing acquisition campaigns, it is worth exploring how AI agents will integrate with tools like Google Ads and LinkedIn Ads . Furthermore, the SEO copywriting is one of the fields where agentic automation is already producing measurable results.

Anyone wanting to explore the implications of this evolution for their business can check out our Blog or contact us directly for a personalized evaluation. We also offer a complete overview of our AI services for Italian SMEs.

In short, OpenAI's acquisition of Ona isn't just industry news to file away. It's a concrete indicator of where enterprise AI is heading. SMEs that start positioning themselves today will have a major edge over the next 18-24 months.