OpenAI's Codex has crossed a major milestone that signals AI agents are no longer experimental tools, but operational infrastructure for enterprises. The platform now serves more than 3 million weekly active users, a figure that grew more than fivefold since January 2026, according to recent reporting. What started as an autocomplete tool for developers has evolved into a system that can run multiple AI agents simultaneously, each handling different parts of complex projects and handing work off to one another.

What Changed in Codex's Latest Update?

In April 2026, OpenAI introduced capabilities that fundamentally transformed how Codex operates. The platform gained the ability to execute multiple AI agents in parallel, control computer interfaces directly, and delegate tasks across entire software development lifecycles. This means Codex can now orchestrate work across different stages of a project without human intervention between steps.

The practical impact is significant. Tasks running in Codex's sandbox environments typically complete within minutes, while more complex jobs wrap up in about 30 minutes. The platform supports both cloud-based and local execution through command-line interfaces (CLI) and integrated development environment (IDE) integrations, giving enterprises flexibility in how they deploy the technology.

How Is Codex Being Used Beyond Software Development?

The expansion beyond coding represents the most striking shift in Codex's trajectory. In June 2026, OpenAI released a report highlighting Codex's application to automating routine knowledge work across various professions, not just software engineering. This includes tasks in finance, operations, customer service, and administrative work, anywhere repetitive decision-making or data processing occurs.

Enterprise adoption is accelerating. Major corporations including Goldman Sachs, DoorDash, Thermo Fisher, State Farm, and LY Corporation are all actively expanding their use of Codex in 2026. These are not small pilots or experimental deployments, but strategic integrations into core business operations. The fact that financial services firms and logistics companies are betting on Codex suggests the technology has moved past the proof-of-concept phase.

Steps to Understanding Codex's Enterprise Impact

• Multi-Agent Execution: Codex can run multiple AI agents simultaneously, each handling different parts of a project and coordinating with other agents, enabling complex workflows without human oversight between steps.
• Cross-Industry Application: While originally designed for software development, Codex now automates routine knowledge work across finance, logistics, insurance, and other sectors where repetitive decision-making is common.
• Speed and Efficiency: Tasks typically complete within minutes to 30 minutes depending on complexity, allowing enterprises to process work at scale that would previously require human teams.
• Flexible Deployment: The platform supports both cloud-based and local execution through CLI and IDE integrations, giving enterprises control over where and how their AI agents run.

What Infrastructure Powers This Growth?

The explosion of Codex usage has infrastructure implications. NVIDIA has noted outstanding performance improvements when pairing its hardware with GPT-5.5, the language model powering Codex's latest agentic workflows. This partnership suggests that as Codex workloads scale, the computational demands will continue to grow, benefiting hardware providers and cloud infrastructure companies.

OpenAI has emphasized that these agentic capabilities represent a fundamental shift toward AI-driven operational efficiency at scale. The company is positioning Codex not as a tool that assists humans, but as a system that can independently manage tasks across an organization's workflows. This distinction matters because it changes how enterprises think about AI investment, moving from augmentation to automation.

The growth from January to June 2026 demonstrates that enterprises are not waiting for perfect AI agents. They are deploying Codex today, learning from real-world usage, and expanding deployments as they identify high-value use cases. This adoption curve suggests that AI agents will become as foundational to enterprise infrastructure as cloud computing and databases are today.