Microsoft has entered into a landmark multi-year cloud agreement with AI infrastructure provider Nebius Group, valued at up to $19.4 billion. The deal, which extends through 2031, will see Nebius supply Microsoft with dedicated graphics processing unit (GPU) infrastructure capacity from a new data center in New Jersey, a move designed to bolster Microsoft's artificial intelligence capabilities amid surging demand. This partnership underscores a significant trend among major cloud providers who are increasingly turning to specialized infrastructure companies to navigate the high costs and supply chain complexities of the AI boom. The agreement is a substantial validation for Nebius, an Amsterdam-based company that spun off from Russian tech giant Yandex, and signals a strategic shift in how hyperscalers are managing their AI infrastructure expansion.

The partnership with Nebius is a direct response to the capacity constraints Microsoft has been facing within its Azure cloud platform.[1][2][3][4][5] The rapid adoption of AI technologies and the intensive computational power required for large language models have created a demand for GPU resources that has outstripped the available supply, a challenge acknowledged by Microsoft executives.[1][2][3] To address this, Microsoft has been making record investments in building out its own data centers, but has also increasingly relied on third-party providers to supplement its capacity.[1][4] This strategy allows Microsoft to expand its AI service offerings more rapidly without shouldering the entirety of the capital expenditure for new data centers.[6] The deal with Nebius, similar to arrangements with other providers like CoreWeave, enables Microsoft to secure a long-term supply of crucial computing power, thereby mitigating supply concentration risks and ensuring it can meet the growing needs of its Azure AI customers.[1][7]

Nebius Group has emerged as a key player in the AI infrastructure space, offering a full-stack platform designed for intensive AI and machine learning workloads.[8][9][10] Headquartered in Amsterdam, the company has a global presence with research and development hubs across Europe, North America, and Israel.[9][10] Nebius provides not only large-scale GPU clusters but also a suite of cloud services and developer tools, supported by its own proprietary software and in-house hardware design.[11][9][10] The company's origins trace back to Yandex, once considered the "Google of Russia."[12][1] Following Russia's invasion of Ukraine, Yandex's international businesses were spun off and rebranded as Nebius Group, which is now an independent entity focused on the global AI market.[12][13][6] This strategic pivot has been backed by significant investment, including a $700 million funding round with participation from Nvidia, which has bolstered its expansion efforts, including the development of data centers in Finland, France, and the United States.[8][13] The deal with Microsoft is seen as a monumental step for Nebius, providing a significant revenue stream and solidifying its credibility as a leading "neocloud" provider.[1][14]

The Microsoft-Nebius agreement carries significant implications for the broader AI and cloud computing industries. It highlights the immense and growing demand for specialized AI infrastructure, a market that is projected to experience explosive growth.[15] For Microsoft, it represents a flexible and capital-efficient approach to scaling its AI capabilities, allowing it to remain at the forefront of the AI race against competitors like Amazon and Google.[7] For Nebius, the deal is transformative, providing a long-term, stable revenue source and a powerful endorsement from a leading hyperscaler that is likely to attract other high-profile customers.[6][14] The stock market's reaction was immediate and positive, with Nebius's shares surging dramatically following the announcement.[1][7] This partnership signals a maturing of the AI infrastructure market, where specialized providers with deep technical expertise can play a crucial role in supporting the ambitious goals of the world's largest technology companies. The move also suggests a future where hybrid cloud strategies, involving a mix of in-house and third-party infrastructure, become the norm for managing the intensive demands of artificial intelligence.