NVIDIA Reports 10x Research Speed Gain With Codex, Expands Access to 40,000 Staff
Updated
Updated · OpenAI · May 12
NVIDIA Reports 10x Research Speed Gain With Codex, Expands Access to 40,000 Staff
1 articles · Updated · OpenAI · May 12
NVIDIA said Codex with GPT-5.5 has become a default engineering and research tool, cutting end-to-end machine-learning research workflows by 10x and helping teams ship production systems faster.
GPT-5.5’s longer autonomous sessions let Codex identify bugs, write scripts, run experiments on remote machines via SSH, and test software with minimal handholding, according to NVIDIA engineers and researchers.
40,000 NVIDIANs now have access to Codex, which runs in production on NVIDIA GB200 and GB300 infrastructure and is being used across enterprise development workflows in North America.
Engineers said the tool helped turn an internal platform from MVP to production-ready, build a podcast-recording app in hours instead of weeks, and rewrite older Python code into Rust with reported 20x efficiency gains.
With AI automating entire workflows, what is the new role for human engineers beyond simply managing their digital colleagues?
As AI agents become digital employees, what happens when one goes rogue and causes a catastrophic system failure?
Is NVIDIA's success creating an unbridgeable competitive gap, leading to an AI monopoly instead of democratizing innovation?
Inside NVIDIA’s Landmark 100,000-GPU AI Agent Deployment: The GPT-5.5 Codex Revolution
Overview
In May 2026, NVIDIA launched a global rollout of OpenAI’s advanced AI agent, Codex, powered by the new GPT-5.5 model. This milestone was made possible by a decade of collaboration between NVIDIA and OpenAI, leading to the development of the powerful GB200 NVL72 systems and the massive 100,000-GPU cluster. Codex was specifically engineered to run on this infrastructure, enabling NVIDIA’s worldwide workforce to achieve dramatic gains in productivity and efficiency. This deployment marks a turning point in enterprise AI, showing how deep partnerships and cutting-edge hardware can transform how companies work and innovate.