Nvidia has rolled out the OpenAI Codex agent tool, based on the latest GPT-5.5 model, for nearly 10,000 of its employees. This decision was made after China's DeepSeek laboratory refused to grant Nvidia and AMD early access to its new V4 model. The American technology giant responded by accelerating the integration of in-house developments into its workflows. Shazoo.ru reports this.
Codex is an agent application designed for programming, with GPT-5.5 model support currently being added. According to Nvidia's corporate blog, this tool has already enabled significant efficiency gains. Bug-fixing processes that previously took several days are now completed in just a few hours. Also, experiments on complex code bases that required weeks of time now find their solutions overnight.
For security purposes, each employee was assigned a separate cloud virtual machine to run the agent. Although Codex has the ability to read corporate data, it does not have the right to directly edit or delete it. This ensures the system operates safely in an isolated environment. The tool is currently used not only by engineers but also by employees of legal, marketing, finance and HR departments.
Such close cooperation between Nvidia and OpenAI was an expected development. The company led by Sam Altman relies on Nvidia investments both technologically and financially. According to experts, this step further strengthens the consolidation of US companies in the global race in the field of artificial intelligence.
Read “Zamin” on Telegram!
