NVIDIA CEO Huang Renxun recently sent a company-wide instruction to the company's ten thousand employees through an internal email, requiring all departments, from engineering to human resources, to fully adopt the OpenAI Codex programming and automation tool based on the GPT-5.5 architecture. This move marks a deep integration between a global semiconductor giant and a top AI laboratory at the application level, aiming to significantly improve the company's global operational efficiency by viewing AI agents as "teammates."
According to internal details disclosed by Sam Altman, the Codex based on the latest GPT-5.5 core has not only made a leap in code generation but has also expanded its versatility into non-technical areas.
Currently, the 10,000 employees across departments such as engineering, products, legal, finance, and operations have completed the initial testing. Huang Renxun emphasized that Codex is not just a simple tool but a "superpower" assistant that allows employees to surpass their previous capabilities. To accelerate the implementation of the technology, the two companies have established a joint "Codex Lab" in NVIDIA's Endeavor building and plan to assist all employees in transitioning through a series of workshops.
From an industry perspective, the release of Codex under the GPT-5.5 architecture is not only an iteration of model capabilities but also a key turning point for OpenAI's transformation from a single model supplier to an enterprise-level solution provider. As an upstream provider of computing power, NVIDIA's decision to deeply integrate its core business processes into the OpenAI ecosystem not only verifies the reliability of GPT-5.5 in complex industrial and commercial scenarios but also signals that "AI for all" will become a standard configuration for tech giants. This vertical integration between software and hardware giants is accelerating AI's evolution from an experimental tool to a core productivity operating system for enterprises.
