OpenAI Codex Reaches 1.6 Million Users as Agentic Coding Takes Hold

In just eight weeks, the landscape of software development has shifted beneath our feet. OpenAI recently revealed that Codex now supports 1.6 million weekly users, a staggering threefold increase since the start of the year. This surge represents the dawn of "agentic" engineering, where AI moves beyond simple autocomplete to plan and execute complex tasks independently.

The Infrastructure of an AI Giant

The rapid adoption of Codex is closely tied to OpenAI’s recent $110 billion funding round, which valued the company at a historic $840 billion. This capital injection is being immediately funneled into physical infrastructure. Strategic partnerships with Amazon and NVIDIA ensure that OpenAI has the chips and server capacity required to maintain its lead. Specifically, a $50 billion investment from Amazon will allow Codex to run on custom AWS Trainium chips, while NVIDIA provides next-generation inference hardware.

Inside OpenAI, the tool has become the standard for development. The company reported that 95% of its own engineers use Codex weekly, leading to a 70% increase in pull request volume. This internal success mirrors the broader market, where a standalone Codex Mac app surpassed one million downloads within its first week this February. For most developers, the tool is no longer an optional add-on but a fundamental part of the environment.

Enterprise adoption has reached a saturation point that few expected so early in the decade. Currently, 90% of Fortune 100 companies have integrated Codex-powered tools into their workflows. As Sam Altman noted in a recent blog post, leadership in this era is defined by who can scale infrastructure fast enough to meet this unprecedented demand.

From Assistant to Autonomous Teammate

The true catalyst for this growth was the February launch of GPT-5.3-Codex. This model introduced "agentic" capabilities, allowing the AI to act as a proactive teammate rather than a passive assistant. Instead of simply predicting the next line of code, Codex can now run its own tests, conduct research across libraries, and manage the deployment infrastructure. Alexander Embiricos, OpenAI’s Codex Product Lead, argues that the bottleneck is no longer the AI's capability, but rather human typing speed.

While productivity gains are massive—with developers reporting they complete tasks 55% faster—the economic transition is proving painful for many. In late February, the fintech giant Block announced it was laying off 4,000 employees, nearly 40% of its workforce. The company explicitly cited AI-driven productivity gains as the reason for the reduction. This has sparked a complicated debate about the future of entry-level engineering roles in an age where AI handles the majority of boilerplate work.

Despite its dominance, OpenAI is not without competition. Anthropic’s Claude Code remains a favorite for developers who prioritize safety and rigorous reasoning. Furthermore, the "QuitGPT" movement has gained some traction among users concerned about OpenAI’s deepening ties to military infrastructure and its shift toward ad-supported models. However, with token usage increasing fivefold in two months, the momentum of Codex seems difficult to stall.