No longer directly challenging AWS, NVIDIA reorganizes cloud team

Media reports indicate that more than two years after NVIDIA CEO Jensen Huang proposed the vision of building a cloud service to compete with Amazon Web Services (AWS), he has abandoned the idea of directly competing with these companies and reorganized the company's cloud team last week.

Several individuals familiar with the organizational changes and an internal memo disclosed this shift. These individuals told the media that the restructuring memo shared with some employees last week involved a reassignment of responsibilities for the head of NVIDIA's cloud business unit and several related executives, while some individuals also left the company.

According to the memo, one executive, Alexis Black Bjorlin, joined NVIDIA from Meta in 2023 and reports directly to Jensen Huang, is currently looking for a new position within the company. NVIDIA has merged its cloud team, which consists of hundreds of people, into the engineering and operations organization, led by Senior Vice President Dwight Diercks, who also reports to Jensen Huang.

The memo and several individuals who previously worked in the department told the media that the cloud team, known as DGX Cloud, will primarily serve the needs of NVIDIA engineers for NVIDIA chips, which are used to develop open-source AI models, rather than focusing on selling cloud services to external corporate clients.

As part of the reorganization, NVIDIA's newer cloud service DGX Cloud Lepton will also be integrated into the engineering team. This service allows cloud service providers to list their unused NVIDIA server computing power in a market operated by NVIDIA, but the project has not started smoothly.

Jensen Huang first launched the DGX Cloud service at NVIDIA's annual flagship conference for developers and customers in March 2023. This initiative aimed to create new revenue streams and help the company establish direct connections with AI developers renting NVIDIA chips from cloud service providers like AWS, Google, and Microsoft.

NVIDIA's selling point at the time was that using chips through DGX Cloud would outperform the configurations provided by the cloud service providers themselves.

However, in reality, NVIDIA is concerned that as Google, Microsoft, and Amazon each develop their own AI chips, they will reduce their reliance on NVIDIA chips and steer customers toward these alternatives. Establishing direct relationships with AI developers is seen by NVIDIA as an important means to hedge against this risk.

This plan has significant potential at least on paper, and NVIDIA has also promoted its early customers, including ServiceNow, SAP, and Amdocs. To build its cloud service, NVIDIA rents servers from large cloud service providers, customizes them according to its extremely strict standards, and then rents these servers to AI developers.

Conflicts of Priority in Competition

However, several individuals who previously worked in the department told the media that the DGX team has always struggled to attract enough customers. One individual pointed out to the media that since DGX Cloud actually operates in the data centers of different cloud service providers like AWS, NVIDIA also faces difficulties in providing troubleshooting support to customers, as repairs made in one cloud service provider's data center may not be applicable to another company's facilities One person told the media that Jensen Huang is also unwilling to significantly expand this business to avoid upsetting cloud service providers who are already the largest chip customers of NVIDIA. At the same time, NVIDIA has taken several measures to provide financial support to emerging cloud service providers such as CoreWeave and Lambda, whose businesses are actually in competition with DGX Cloud.

Earlier this year, NVIDIA began to scale back its cloud business, which is still in its early stages. The company had previously told investors that this business is expected to generate $150 billion in revenue in the future, exceeding the current annual revenue scale of AWS.

Despite companies like Google trying to capture some market share, NVIDIA's absolute dominance in AI chip sales remains solid. AWS has significantly reduced the price of its Trainium AI chips and is in discussions with OpenAI about a collaboration plan to allow this major AI developer, which primarily uses NVIDIA chips, to also start using Trainium chips. Additionally, Meta Platforms is considering spending billions of dollars to purchase Google's AI chips, specifically Tensor Processing Units (TPUs).

In the highly intertwined relationships among these tech companies, NVIDIA itself has become one of the largest lessees of NVIDIA servers, which are procured by cloud service providers like AWS and Google. On one hand, NVIDIA is trying to rent some of these servers to DGX Cloud customers, while on the other hand, it is using these servers to develop various AI models, including those related to robotics and autonomous driving technology.

NVIDIA stated that it plans to invest $26 billion in renting such servers over the next few years.

A spokesperson for NVIDIA told the media:

"We will continue to invest in DGX Cloud, providing world-class infrastructure for cutting-edge research and the software capabilities needed for cloud partners to succeed. Our goal has always been to cultivate DGX Cloud as a pilot project, learning how to better build systems for ecosystem partners, and that has not changed."