AIINEWS

AIINEWS

Báo cáo chuyên sâu về Trí tuệ nhân tạo và Công nghệ toàn cầu

Nvidia’s Jensen Huang resigns, chipmaker names CFO Colette Kress as interim CEO

Ban biên tập AIINEWS•hardware
12 tháng 11, 2026•5 phút đọc
Nvidia’s Jensen Huang resigns, chipmaker names CFO Colette Kress as interim CEO

Tóm lược nhanh

“Nvidia gets a new chief executive, days after former CEO Jensen Huang admitted to a romantic affair with an alleged Chinese spy.”

Jensen Huang has resigned from his position as president and chief executive of Nvidia, the company said on Thursday, just a few days after he admitted to a romantic affair with an alleged Chinese spy, sparking a U.S. national security crisis.

In a regulatory filing, Nvidia said its board accepted Huang’s resignation on Wednesday.

Nvidia said its current chief financial officer and executive vice president, Collete Kress, will take over as interim CEO until the company finds a permanent successor. Kress will continue to serve as the company’s CFO.

Nvidia’s shares were down nearly 12% in early morning trading on Thursday. Since Friday, the company’s stock has fallen by as much as 19%, resulting in a market cap loss of more than $855 billion.

Huang founded Nvidia in 1993 and has led it as CEO ever since. Under his leadership, Nvidia played a crucial role in developing the the graphics processing unit (GPU), a specialized chip that does graphics processing for PCs, as we know it today. The chips were not an immediate hit, and the early days of Nvidia were characterized by struggle and layoffs, but eventually, it struck on a GPU design that performed so well, it became the standard for graphics processing.

These chips also turned out to be extremely good at doing large numbers of calculations simultaneously, which has made them popular for computation-intensive tasks like crypto mining, and more recently, training and running neural networks, which are what AI models are at their core.

Nvidia was initially focused on making GPUs aimed at gaming, but it had already moved into general-purpose computing by 2006 with the launch of the CUDA architecture, which lets developers use the chips for other tasks, such as doing calculations and research. After AI researchers began using the company’s GPUs to train neural networks over a decade ago, Huang leaned in towards developing chips for AI and then data centers, and grew it from there into its dominant position.

Repercussions galore

The potential implications of Huang’s departure for the AI and data center industries are vast, given how active Huang, and by extension Nvidia, has been in furthering the AI boom. In particular, it raises questions about the future of Nvidia’s sprawling AI investments, which have helped bankroll everything from startups developing new technologies to the enormous data centers needed to run them.

Under Huang, Nvidia deftly used its financial muscle to support companies buying its chips, effectively helping to finance the market for its own products, as well as lobbying Washington to relax potential rules around AI and data center regulation.

As of the end of its second quarter, Nvidia reported $99 billion worth of investments and another $25 billion of commitments. It also has $36 billion in cloud service commitments, which are arrangements that help AI cloud operators finance the GPUs and data center equipment they buy from Nvidia, with the chipmaker saying it will buy compute from them and potentially earn a share of their operators’ revenue.

Nvidia has also been investing directly in neoclouds, the companies serving a lot of the AI processing needs. In January, the company invested $2 billion in CoreWeave, and another $2 billion in Nebius in March. The chipmaker was said to be considering being a guarantor for Nscale’s planned West Virginia data center project. It’s also signed up to provide up to $105 billion worth of credit guarantees for an OpenAI data center campus in Ohio.

Nvidia’s also a major investor in the labs building the AI models that run on its GPUs, putting $30 billion into OpenAI, and $10 billion into Anthropic. It’s also backed Mira Murati’s Thinking Machines Lab as part of a deal involving at least 1 GW of its Vera Rubin GPU systems. The chipmaker has also invested $3.5 billion in MediaTek to help the Taiwanese company develop custom AI chips that work with Nvidia’s systems.

Many investors have grown uneasy with the enormous amounts of debt companies in the space were raising to sustain the AI infrastructure buildout, and Huang’s exit throws another uncertainty-shaped wrench into the works. CoreWeave’s shares were down about 9% on Thursday, reversing part of a 25% rally on Wednesday, while Nebius is down about 16%. IREN fell 6.6%.

Nvidia’s board has also appointed Harvey C. Jones, a managing partner of Square Wave Ventures and former CEO of Synopsys, as the company’s first-ever chair of the board of directors. Jones has been on Nvidia’s board since 1993. As chair, Jones is likely to play a central role in Nvidia’s search for a permanent CEO to replace Huang.

Huang is the single largest personal shareholder at Nvidia, but he owns only about 3.5% of the company’s shares. Unlike many of Silicon Valley’s tech giants, he does not hold supervoting rights, and the board had the power to vote him out at any time.

Huang famously ran a very flat organization, with around 60 direct reports. There was no heir apparent or succession plan either (although two of his children also worked in executive roles at the company). Huang famously maintained tight control over the company he founded, making it challenging for Nvidia to find a path forward without him.

The industry will get its first glimpse of that next week during the quarterly conference call. Nvidia reports earnings on Tuesday.

“Jensen’s contributions to NVIDIA are profound. NVIDIA is one of the most consequential technology companies in the world. The Board and the company thank Jensen for his immeasurable contributions to the company and the industry,” the company said in a statement.

This story is developing…

Bản quyền nội dung và biên tập thuộc về Ban biên tập AIINEWS
Chia sẻ bài phân tích này:
Góc nhìn mở rộng & Đột phá công nghệ

Các phân tích & Tiêu điểm liên quan

Hiển thị 8 bài phân tích tuyển chọn
Alibaba ra mắt mô hình ngôn ngữ lớn Qwen4-Max với kiến trúc huấn luyện tối ưu
Mô hình & Đột phá AIkhoảng 9 giờ trước

Alibaba ra mắt mô hình ngôn ngữ lớn Qwen4-Max với kiến trúc huấn luyện tối ưu

Alibaba vừa chính thức giới thiệu Qwen4-Max, mô hình ngôn ngữ lớn mạnh mẽ nhất của hãng với tổng số tham số lên tới 2.4 nghìn tỷ nhưng chỉ kích hoạt 71 tỷ tham số tại mỗi thời điểm hoạt động. Nền tảng này tích hợp các cải tiến chuyên sâu về cơ chế chú ý và quy trình huấn luyện giúp giảm đáng kể chi phí vận hành.

2 phút đọc2,071
Startup chip bán dẫn Bharat Semi huy động 24 triệu USD nhằm phát triển cảm biến AI

Startup chip bán dẫn Bharat Semi huy động 24 triệu USD nhằm phát triển cảm biến AI

Công ty thiết kế vi mạch Bharat Semi vừa huy động thành công khoản vốn hạt giống 24 triệu USD để chế tạo dòng cảm biến hình ảnh tích hợp trí tuệ nhân tạo. Doanh nghiệp này đang nhắm tới các thị trường quốc tế như Mỹ với mục tiêu thương mại hóa sản phẩm vào năm 2027.

2 phút đọc4,911
Sam Altman làm rõ phát biểu về dung lượng RAM và chuỗi cung ứng trí tuệ nhân tạo
Lãnh đạo & Nhân vật AIkhoảng 10 giờ trước

Sam Altman làm rõ phát biểu về dung lượng RAM và chuỗi cung ứng trí tuệ nhân tạo

Giám đốc điều hành OpenAI Sam Altman vừa giải thích lại những phát ngôn gây tranh cãi về bộ nhớ máy tính trong bối cảnh các nhà sản xuất trí tuệ nhân tạo đẩy giá linh kiện tăng vọt. Ông cho biết mình hoàn toàn ủng hộ người dùng cá nhân sở hữu nhiều RAM để tự mày mò công nghệ.

2 phút đọc2,316
Mỹ chuẩn bị siết xuất khẩu chip AI để hạn chế truy cập năng lực tính toán từ xa
Chính sách & Pháp lý AIkhoảng 8 giờ trước

Mỹ chuẩn bị siết xuất khẩu chip AI để hạn chế truy cập năng lực tính toán từ xa

Chính quyền Mỹ đang chuẩn bị mở rộng kiểm soát xuất khẩu chip AI nhằm ngăn doanh nghiệp ở Trung Quốc và các nước bị hạn chế tiếp cận năng lực tính toán đặt tại nước thứ ba. Theo bản tin nguồn, quy định dự kiến sẽ yêu cầu giấy phép đối với một số lô hàng phục vụ trung tâm dữ liệu.

3 phút đọc1,474