Amsterdam-based AI cloud provider Nebius Group has inked a landmark five-year agreement valued at approximately $3 billion with Meta Platforms, the parent company of Facebook, to supply high-performance AI infrastructure. This deal, announced alongside Nebius’s third-quarter financial results, underscores the surging demand for specialized computing power in the artificial intelligence sector as tech giants race to fuel their next-generation models.
The partnership marks Nebius’s second major hyperscaler contract in 2025, following a multi-billion-dollar agreement with Microsoft earlier this year, and positions the company as a key player in the “neocloud” space focused on GPU-intensive AI workloads.
Deal Details and Timeline
Under the terms of the agreement, Nebius will deliver advanced AI infrastructure, including GPU-based compute resources and data center capacity, to support Meta’s ambitious AI initiatives over the next five years. The deal’s value reflects committed capacity rather than upfront payments, with Nebius capping the amount due to overwhelming demand that exceeded its current available supply—highlighting the tight market for high-end AI hardware like Nvidia GPUs.
Deployment is set to begin within three months, allowing Meta to integrate Nebius’s resources into its ecosystem for training and running large language models, such as those powering its Llama series. This move helps Meta diversify beyond traditional hyperscalers like AWS and Google Cloud, amid a global shortage of AI-capable infrastructure.
Nebius, which specializes in “neocloud” services tailored for AI rather than general-purpose computing, will leverage its expanding network of data centers in the U.S., Europe, and Israel to fulfill the contract. The company’s recent launch of one of Israel’s first public GPU clusters using Nvidia’s latest chips further bolsters its capability to handle such large-scale demands.
Nebius’s Financial Surge and Strategic Positioning
Nebius reported third-quarter 2025 revenue of $146.1 million, a staggering 355% increase year-over-year, driven by sold-out capacity in its AI cloud offerings and the broader boom in AI adoption. Despite this growth, the company posted a wider net loss, attributed to heavy investments in expansion, with capital expenditures jumping to $955.5 million—primarily on Nvidia GPUs, real estate, and power infrastructure.
CEO Arkady Volozh, the founder who previously led Yandex before spinning off Nebius in 2024, described the year as foundational for future acceleration. “We are at the forefront of one of the most significant technological revolutions in history—and Nebius has quickly become a core enabler of the AI-driven economy,” Volozh stated, emphasizing the Meta deal as a validation of their model.
To support ongoing growth, Nebius launched an at-the-market equity program for up to 25 million Class A shares, aiming to raise funds efficiently while minimizing dilution. Ending the quarter with $2.43 billion in cash and equivalents, the company projects an annualized revenue run rate of $7 billion to $9 billion by the end of 2026, up from about $551 million currently.
Nebius’s origins as a Yandex offshoot have not hindered its global ambitions; re-established in Amsterdam, it has rapidly scaled by focusing on AI-specific infrastructure, differentiating itself from broader cloud giants. This deal follows a $17.4 billion pact with Microsoft in September, potentially expanding to $19.4 billion by 2031, and signals more hyperscaler partnerships on the horizon.
Implications for AI Infrastructure Market
The Meta-Nebius partnership arrives at a pivotal moment when AI demand is outstripping supply, with companies like Meta investing billions to build sovereign AI capabilities outside dominant providers. For Meta, this deal reduces reliance on a few suppliers and accelerates its push into advanced AI applications, from content moderation to generative tools across its platforms.
In the broader market, it highlights the rise of specialized “neocloud” providers like Nebius and rivals such as CoreWeave, which are capitalizing on the GPU crunch to secure premium contracts. Analysts see this as a boon for AI stocks, though execution risks remain—Nebius must now rapidly deploy new sites to reach 2.5 gigawatts of contracted power by 2026.
For investors, the announcement drove Nebius shares higher, reflecting optimism about its trajectory despite short-term losses. As AI infrastructure becomes the backbone of the tech economy, deals like this could reshape how Big Tech sources its computing muscle.
This collaboration not only cements Nebius’s role in the AI ecosystem but also illustrates the intensifying global competition for the hardware powering tomorrow’s innovations.
