Jensen Huang invested in the "Russian Jack Ma".

On March 12, TechNode reported that on March 11 local time, NVIDIA announced an investment of $2 billion (approximately RMB 13.7 billion) in Dutch AI cloud provider Nebius and established a strategic partnership with it. The plan is to deploy more than 5 gigawatts (GW) of NVIDIA-based computing power centers by the end of 2030. 

Nebius was founded in 2024 and was formerly the Dutch holding company Yandex N.V. of Russian tech giant Yandex. Its founder and CEO, Arkady Volozh, is a co-founder of Yandex. Many industry insiders also call him the "Jack Ma of Russia." 

In 2024, he led Yandex N.V. to sell its Russian business and then underwent a strategic restructuring, renaming it Nebius Group N.V. The company is headquartered in Amsterdam, the Netherlands, and focuses on one-stop AI cloud services. 

▲Arkady Volozh, founder and CEO of Nebius

As of the close on March 11, Nebius' stock price on the NASDAQ rose 16.14% to $112 (approximately RMB 769) per share, with a total market value of $28.337 billion (approximately RMB 194.766 billion). 

▲Nebius' stock price trend (Source: Tencent Stock Choice)

This is another new AI cloud force supported by NVIDIA in the wave of AI Agents. NVIDIA founder and CEO Jensen Huang mentioned in the announcement: "Nebius is building an AI cloud specifically designed for the Agent era." 

Before investing in Nebius, NVIDIA had already invested in well-known AI cloud providers Lambda and CoreWeave. In January 2026, NVIDIA just added an investment of $2 billion in CoreWeave, aiming to help CoreWeave build more than 5 gigawatts of AI computing power by 2030. Deeply binding with emerging AI cloud providers has become one of the important measures for NVIDIA to consolidate its leading position in the AI field. 

According to the document submitted by Nebius to the U.S. Securities and Exchange Commission (SEC) on March 11, Nebius reached a private placement agreement with NVIDIA, issuing prepaid warrants to raise a total gross proceeds of approximately $2 billion. The exercise price of the warrants is $0.0001 per Class A common stock, and the number of exercisable shares is 21,065,936 Class A common stocks. The relevant shares will be delivered from treasury stocks or newly issued shares. 

The company plans to use the net funds raised for the R & D of a full-stack AI cloud platform and build new green data centers. 

▲A screenshot of part of the document submitted by Nebius to the SEC

01 A 62-year-old "Jack Ma of Russia" at the helm, with over $20 billion in large orders from Microsoft and Meta

Arkady Volozh, the founder of Nebius, was born in Kazakhstan in 1964 and is now 62 years old. He is a co-founder of Yandex, a Russian tech giant known as the "Russian Google." 

His career has spanned the market-oriented exploration after the disintegration of the Soviet Union, the rise of the Russian Internet, and now the second entrepreneurial wave in the global AI era. 

▲Arkady Volozh, founder and CEO of Nebius

In 1997, Volozh led the launch of the Yandex search engine, which has ranked first in Russia since 2001 and gradually expanded to nearly a hundred online services such as payment and ride-hailing. In 2013, it became the fourth-largest search engine in the world. 

In 2024, affected by the Russia-Ukraine conflict, Volozh led about 1,300 former Yandex engineers to start a second business and founded the AI cloud platform Nebius, providing one-stop services from data centers to computing power. In October of the same year, Nebius inherited Yandex N.V.'s listing qualification on the NASDAQ, completed an IPO, and resumed stock trading. 

As of the end of 2025, Nebius had signed AI large orders worth more than $20 billion with companies such as Microsoft and Meta. 

In November 2025, Nebius announced a deal worth approximately $3 billion with Meta, which will provide AI infrastructure for Meta within five years. Previously, in September, it reached a deal worth $17.4 billion with Microsoft, and the company's revenue in the third quarter of that year increased more than fourfold. 

Arkady Volozh said when talking about this financing: "Nebius was born for AI from the very beginning. It was not transformed from a general cloud platform but was specifically designed to meet the actual needs of developers. Now, with the help of NVIDIA, we are extending this concept to the entire technology stack, from gigawatt-scale AI factories to inference and software, to build one of the first and largest cloud platforms for all AI developers globally." 

▲A real - shot photo of Nebius' data center (Source: Nebius)

02 Deploy over 5GW clusters by the end of 2030 and carry out four major collaborations

Before receiving this round of financing from NVIDIA, Nebius had continuously deployed NVIDIA chips on its global platform, including building multiple gigawatt-scale AI factories in the United States. After receiving the financing, Nebius will more promptly adopt NVIDIA's latest products to support the deployment of computing power centers with a capacity of more than 5 gigawatts (GW) by the end of 2030. 

According to the terms of the cooperation agreement announced by NVIDIA, the two companies will cooperate in the following aspects: 

1. AI factory design and support: Including access to partner design materials, design review processes and acceptance, early samples and system software support, startup support, and regular system partner business and technical reviews. 

2. Inference: Leveraging NVIDIA's latest software technologies, optimized models, and libraries to build a first-class inference and intelligent AI stack for developers and enterprises. 

3. AI infrastructure deployment: Deploying multiple generations of NVIDIA infrastructure on the Nebius platform by early adopting NVIDIA's computing architecture, including the NVIDIA Rubin platform, NVIDIA Vera CPU, and NVIDIA BlueField storage system. 

4. Device cluster management: Optimizing the overall health of Nebius' device clusters by deploying NVIDIA's latest GPU health monitoring and software recommendations. 

03 For the Agent era, NVIDIA deeply binds with AI cloud providers

"AI is at another turning point - the emergence of AI Agents has driven astonishing computing demand and accelerated infrastructure construction," said Jensen Huang. "Nebius is building an AI cloud specifically designed for the Agent era, fully integrated from chips to software and supported by NVIDIA's next-generation accelerated computing technology." 

▲Jensen Huang, founder and CEO of NVIDIA

In addition to investing in Nebius, NVIDIA is also an early investor in well-known AI cloud provider Lambda. The specific investment amount and round have not been publicly disclosed. In addition to capital support, the cooperation between the two parties is very in - depth - NVIDIA is not only a shareholder of Lambda but also an important customer. According to public reports, NVIDIA once signed a server leasing agreement worth billions of dollars with Lambda, which is equivalent to "both investing money and giving orders." 

At the same time, NVIDIA has gradually increased its investment in the new U.S. AI cloud star CoreWeave. Initially, NVIDIA joined in an earlier financing round of CoreWeave and became one of its shareholders. The milestone in the cooperation between the two parties occurred in January 2026 when NVIDIA announced an additional investment of $2 billion in CoreWeave, aiming to help CoreWeave build more than 5 gigawatts of AI computing power by 2030. 

It can be seen that NVIDIA's cooperation with AI cloud providers belongs to the strategic synergy model of 'investment + deep binding': Through financial support and priority computing power supply, these AI cloud providers are incorporated into its ecological territory to jointly expand the computing power infrastructure for the AI market. 

04 Conclusion: Supporting new cloud forces, NVIDIA secures a position in the AI Agent competition

NVIDIA has invested a total of $4 billion in Nebius and CoreWeave in just three months, both aiming at a computing power deployment of more than 5 gigawatts by 2030, which shows the logic of consolidating its leading position in the AI field: By using capital ties to deeply bind emerging AI cloud providers to its ecosystem, it not only locks in huge orders for next - generation chips but also can counter the potential threat from self - developed chips of cloud giants. 

The wave of AI Agents is pushing the computing power competition to a new height, and NVIDIA's "investment + technology transaction" strategy is essentially securing a position in the future AI factory standards in advance. When general cloud providers are seeking computing power autonomy, the "new cloud" forces supported by NVIDIA are becoming a key move for it to maintain architectural dominance and penetrate enterprise - level AI applications. 

This article is from the WeChat official account "TechNode" (ID: zhidxcom). The author is Li Shuiqing. It is reprinted by 36Kr with permission.