黃仁勳發全員信擁抱 OpenAI Codex：1 萬多名 NVIDIA 員工已上手，GPT-5.5 跑在 GB200 上

📄完整原文· 由 trafilatura 自動擷取Gemini 翻譯2056 字

NVIDIA CEO Jensen Huang personally sent an all-hands email, calling on employees to fully adopt OpenAI's latest tool, Codex, and publicly disclosed that over 10,000 employees across 9 departments are already using it. This move not only symbolizes the deepening cooperation between the two tech giants, but more intriguingly: the GPT-5.5 that powers Codex is running on NVIDIA's own GB200 NVL72 rack-scale systems. (Previous coverage: OpenAI launches new engineer agent Codex! AI can write functions, fix bugs, run tests... limited to 3 types of users for early access) (Background: Massive! NVIDIA invests $100 billion to partner with OpenAI, co-build 10GW AI data center, NVDA jumps 3.9%) An internal email from Jensen Huang caused quite a stir in the tech circle this week—not because of any earth-shattering secret, but because Sam Altman posted it directly on the X platform. According to the official NVIDIA blog, the core of the letter is to announce that the GPT-5.5-powered OpenAI Codex has been fully opened to employees, explicitly naming 9 departments: Engineering, Product, Legal, Marketing, Finance, Sales, HR, Operations, and Developer Programs—with over 10,000 NVIDIANs currently using the tool. The letter ends in his signature style—the official NVIDIA blog quoted the original text verbatim: "Let’s jump to lightspeed. Welcome to the age of AI." NVIDIA noted on its official blog that Codex is now powered by GPT-5.5—OpenAI's latest frontier model—and the underlying computing infrastructure for this model is NVIDIA's own GB200 NVL72 rack-scale system. OpenAI officially released GPT-5.5 to paid subscribers on Thursday (4/23), just 6 weeks after the previous generation GPT-5.4, showing that the pace at which frontier AI labs are competing for enterprise clients has become breathtakingly fast. For NVIDIA, the signal from this all-hands email is clear: the best AI tools run on the best AI chips. Jensen Huang doesn't need to say much; this logic itself is the most powerful advertisement. Jensen Huang specifically named 9 departments in the letter: Engineering, Product, Legal, Marketing, Finance, Sales, HR, Operations, and Developer Programs. This list itself is a manifesto—Codex is not just for those who write code. The letter positions Codex as "not just for coders"—the screenshot of the letter posted by Sam Altman shows that Jensen Huang describes the AI agent as a "teammate" for every employee, emphasizing that it can make work "better, smarter, and faster." However, it should be noted: except for the closing line "Let’s jump to lightspeed" which was quoted verbatim by the official NVIDIA blog, the rest of the letter's content is based on the screenshot posted by Sam Altman on the X platform and reports from foreign media. It is worth noting that the official NVIDIA blog describes this rollout as "already using"—describing the current situation rather than a top-down mandate. After Sam Altman posted the screenshot, foreign media outlets like Gurufocus and Futunn also used words like "encouraged" and "urging everyone to use" to describe Jensen Huang's actions—which are closer to "encouragement" and "call to action" in Chinese, rather than "mandatory requirement." To understand the background of this letter, one cannot ignore the complex interests between NVIDIA and OpenAI. On the surface, this is a deep partnership: NVIDIA previously announced a $100 billion investment to partner with OpenAI to co-build a 10GW AI data center, and foreign media even reported on a $30 billion stake negotiation—if it comes to fruition, NVIDIA will become one of OpenAI's major shareholders. Jensen Huang's personal call for all employees to adopt Codex is merely pushing this partnership deeper. But cracks also exist. OpenAI is actively pushing its own AI chip development project, expecting to complete the design this year and hand it over to TSMC for trial production, aiming to reduce its reliance on NVIDIA's computing power; on the other hand, AMD has also secured a 10% stake in OpenAI, making both AI chip manufacturers hardware partners of OpenAI, with an increasingly subtle competitive and cooperative relationship. From this perspective, the significance of Jensen Huang's letter is not just "promoting an AI tool"—he has explicitly tied the daily productivity of over 10,000 NVIDIA employees to OpenAI's model ecosystem. This is both a public endorsement of the depth of the partnership and can be interpreted as a gesture of goodwill: we are your most loyal users, there is no need for you to build your own chips. Jensen Huang loves using the

資料狀態✓ 已擷取全文閱讀原文（動區 BlockTempo）

🔍歷史類似事件· 關鍵字 + 標的比對6 則

2026-04-24

黃仁勳發內部信：強制 1 萬名 NVIDIA 員工改用 OpenAI Codex，GPT-5.5 用 GB200 晶片跑

相似度 270%關鍵字 openai/codex/nvidia同分類 zh

2026-04-22

OpenAI 祭出 GPT-Image-2：視覺生成迎來碾壓，設計師這次真的要失業了

相似度 170%關鍵字 openai/gpt同分類 zh

2026-04-23

OpenAI 發布 GPT-5.5：速度更快、更聰明——且價格更高

相似度 130%關鍵字 openai/gpt

2026-04-23

OpenAI 發布 GPT-5.5 以挑戰 Anthropic 的 Claude Opus 4.7

相似度 130%關鍵字 openai/gpt

2026-04-16

OpenAI Super App 成形：Codex 具備 Computer Use、瀏覽器與圖像生成功能

相似度 130%關鍵字 openai/codex

2026-04-24

DeepSeek V4 發布，程式能力可打敗 GPT 和 Claude 嗎？費用再次屠榜

相似度 120%關鍵字 gpt同分類 zh

💡 目前用關鍵字 + 標的比對（MVP）· 之後會升級為 embedding 語意搜尋

原始資訊

ID：2d2c8436eb

來源：動區 BlockTempo

發佈：2026-04-24 04:04:14

分類：zh_news · 導出分類 zh

標的：未指定

社群投票：+0 / −0 · ⭐ 0 重要 · 💬 0 留言