OpenAI announced plans to acquire Ona, a startup specializing in secure cloud infrastructure, marking a deliberate move to address a critical gap in enterprise AI deployment: the ability to run long-duration agents with persistent state across workflows. The acquisition expands OpenAI's Codex engine with persistent cloud environments, solving a fundamental problem for organizations attempting to move beyond chatbot-style interactions. Consider a supply chain optimization scenario: an AI agent needs to monitor inventory levels across multiple warehouses, remember historical patterns, access real-time databases, and make decisions over hours or days without restarting. Ona's infrastructure enables exactly this use case—agents that maintain context and state across extended operations rather than resetting on each API call. This addresses what enterprise customers have increasingly demanded: AI that integrates into production systems rather than serving as a conversational tool. The acquisition suggests OpenAI views infrastructure control as essential to competing in enterprise AI deployment, moving beyond its position as a pure model vendor.
Simultaneously, OpenAI launched three Academy courses designed to train workers in practical AI application, shifting focus toward ecosystem enablement. Rather than generic AI literacy, the courses emphasize building repeatable workflows and implementing agents in everyday work. One course, for instance, teaches organizations how to design systems where AI handles routine tasks—like processing customer feedback emails, categorizing issues by urgency, and routing them to appropriate departments—while humans focus on complex judgment calls. This represents a maturation of OpenAI's go-to-market strategy: providing not just models but the knowledge infrastructure for organizations to deploy them effectively. BBVA's scaling of ChatGPT Enterprise to 100,000 employees serves as validation that large organizations are ready to move beyond pilot programs, though it also suggests that widespread adoption still requires significant training and change management.
These moves collectively reframe OpenAI's strategy from licensing model access to building a complete enterprise AI platform. The Ona acquisition addresses the infrastructure layer; the Academy addresses the skills gap; partnership announcements like BBVA demonstrate market validation. However, the framing as 'ecosystem support' warrants scrutiny—this is primarily vertical integration into enterprise workflows rather than ecosystem neutrality. OpenAI is establishing dependencies: organizations trained on OpenAI Academy courses, using persistent Ona infrastructure, running on OpenAI models, face switching costs. Whether this consolidation benefits enterprises through tighter integration or locks them into a single vendor remains an open question. What's clear is that OpenAI has moved decisively beyond model releases to compete on enterprise lock-in, platform effects, and operational integration.