OpenAI's Enterprise Pivot: Why 40% of Revenue Now Comes From "Teams of Agents"

OpenAI is no longer primarily a consumer AI company. Enterprise revenue now accounts for more than 40% of total revenue, and is on pace to reach parity with consumer revenue by the end of 2026. For a company that built its brand on a viral chatbot, that shift in revenue composition tells a bigger story about where AI value creation is actually heading.

OpenAI hit $25 billion in annualized revenue in February, up from $20 billion at the end of 2025. Weekly active users across all products reached 910 million. Paying business users hit 9 million in February, up from 5 million in August. And Codex, the company's AI coding agent, crossed 3 million users from what CFO Sarah Friar described as "almost zero" at the start of the quarter.

The numbers are impressive. But the strategic signal underneath them matters more.

From Productivity Tools to Agent Teams

Chief Revenue Officer Denise Dresser, who spent more than a decade at Salesforce before running Slack, framed the enterprise shift in revealing terms: the companies furthest ahead have moved past "traditional productivity" and are now "managing teams of agents to do tasks for them."

That language is precise and deliberate. Not a single agent assisting a single worker. Teams of agents, coordinating with each other, holding context across sessions, and taking action inside business tools without constant human oversight.

This represents a fundamental change in how enterprises consume AI. Phase one was augmentation: AI helps a person write emails, summarize documents, or draft code faster. Phase two is delegation: AI systems execute multi-step workflows autonomously, with humans providing oversight rather than input at every step.

OpenAI's ChatGPT Agent can now plan trips, book hotel rooms, research competitors, generate slide decks, and place online orders without a human in the loop. That's not a productivity tool. That's an autonomous business process engine.

The Enterprise Platform Play

Dresser's diagnosis of the current bottleneck is worth paying attention to: "What's really missing still for most companies is just a simple way to unleash the power of agents as teammates that can operate inside the business without the need to rework everything."

That's a platform positioning statement. OpenAI is betting that most enterprises don't want to rebuild their technology stack to accommodate AI agents. They want a layer that sits on top of existing tools and lets agents operate within them. The company's enterprise agent platform is designed to fill that gap, giving OpenAI a defensible position beyond the consumer subscription business that still drives the majority of revenue.

The competitive implications are significant. If OpenAI succeeds in becoming the default agent orchestration layer for enterprise workflows, it competes directly with Salesforce, ServiceNow, Microsoft, and every horizontal SaaS vendor whose value proposition depends on human users interacting with purpose-built software. When agents can navigate business tools autonomously, the premium shifts from the application layer to the intelligence layer.

The Talent Signal

OpenAI recently hired Peter Steinberger, founder of OpenClaw, described as the world's most popular open-source agentic AI platform, to lead its push into personal AI agents. That hire signals ambition beyond the enterprise: OpenAI wants to own the agent experience for both businesses and individuals.

CEO Sam Altman has positioned multi-agent systems at the center of OpenAI's next product phase. The company's product roadmap increasingly treats the chat interface as a transitional form factor, with agents as the long-term default.

The $85 Billion Question

OpenAI projects reaching $85 billion in revenue by 2030. That figure only makes sense if agents become the primary way businesses interact with AI, not an optional feature layered on top of a chatbot.

The math is straightforward. At $25 billion annualized today with roughly 25% growth implied per year, the path to $85 billion requires enterprise revenue to scale dramatically while consumer revenue continues compounding. The 40% enterprise share reaching 50% by year-end is the first milestone. But the trajectory beyond that depends on whether multi-agent workflows become sticky, recurring enterprise infrastructure or remain experimental deployments that companies test but don't fully commit to.

The IPO adds urgency to this question. CFO Sarah Friar confirmed this week that retail investors will receive a share of the allocation, a move designed to broaden the investor base ahead of what would be one of the largest technology IPOs in history. The enterprise revenue trajectory will be central to how public market investors value the business.

What This Means for the Market

OpenAI's enterprise pivot validates a thesis that has been building across the technology landscape: the next wave of enterprise software value will be captured by companies that control the intelligence layer, not the application layer.

For incumbent SaaS vendors, the threat is structural. If an AI agent can navigate Salesforce, process contracts, manage procurement workflows, and coordinate across business systems autonomously, the differentiation shifts from software features to agent capabilities. The vendor that provides the agent has more leverage than the vendor that provides the tool the agent operates.

For institutional investors, the key metric to watch is enterprise retention and expansion. Consumer AI subscriptions churn. Enterprise contracts, especially those embedded in mission-critical workflows, don't. OpenAI's ability to convert the current 40% enterprise share into durable, expanding enterprise relationships will determine whether the company's $85 billion revenue target is achievable or aspirational.

Dresser's observation captures the moment: "I have never seen this level of conviction spread so quickly and consistently within the industries." Whether that conviction translates into sustained enterprise spending, or proves to be another cycle of technology enthusiasm that fades when ROI doesn't materialize, is the defining question for AI's commercial trajectory in 2026.