Nebius Lands Clarifai’s Zeiler as Research Chief Amid 684% Revenue Explosion and New Hyperscaler Challenge

Nebius licenses Clarifai AI, hires founder Zeiler. Q1 revenue up 684% to $399M; stock falls after Google-Blackstone cloud JV.

Nebius is doubling down on software just as Wall Street recalibrates its view of the competitive landscape. The AI infrastructure provider has struck a strategic licensing deal with Clarifai, bringing on board the company’s founder and chief executive Matthew Zeiler as its new head of research – a move designed to sharpen the margins on inference workloads that will define profitability in the coming years.

The appointment comes on the heels of a quarterly revenue leap that saw Nebius generate $399 million in the first quarter of 2026, a 684% jump from the $50.9 million posted a year earlier. Adjusted net loss came in at roughly $100 million, well below the $174 million analysts had penciled in, while the AI cloud unit’s adjusted EBITDA margin more than doubled to 45% from 24%. The company closed the quarter with more than $9 billion in cash, boosted by a $2 billion equity injection from Nvidia and proceeds from convertible bond issuance.

Yet the stock has struggled to hold those gains. On May 19, the day Google and Blackstone unveiled a joint venture to build a massive cloud platform around Google’s Tensor Processing Units, Nebius shares traded in a wide band between $183.34 and $202.77 before settling near $195 – a sharp drop from recent highs around $233. Analysts see the Big Tech–private equity alliance as a direct threat to the pricing power that specialized GPU-cluster operators have enjoyed.

The software push is Nebius’s answer to that pressure. Under the Clarifai deal, Nebius will license key technology, absorb the core research and development team, and acquire a portfolio of patents covering AI inference and compute orchestration. The license is perpetual but non-exclusive, with older computer-vision models and business tied to U.S. government and defense left out. Zeiler’s mandate is to strengthen the so-called Token Factory, the service that takes trained models and runs them efficiently in production – an area where costs explode as model size and request volumes rise.

The strategy mirrors an earlier acquisition of Eigen AI, which optimizes model performance. The two pieces are meant to fit together: Eigen works on the model side, Clarifai on the infrastructure layer, giving Nebius more control over hardware utilization, data routing, and ultimately margins.

The financial foundation for this expansion is already laid. Two landmark customer deals underscore demand from the largest technology companies. Microsoft has a multiyear commitment for dedicated GPU capacity worth up to $19.4 billion. Meta Platforms signed a five-year contract valued at $12 billion, with an option for an additional $15 billion that could push the total to $27 billion. Both are anchored by Nebius’s data center in Vineland, New Jersey.

Capacity is being added at breakneck speed. Contracted power now exceeds 3.5 gigawatts, and Nebius has lifted its year-end 2026 target to more than 4 gigawatts, up from an earlier goal of 3 gigawatts. Capital expenditure for the full year is expected to run between $20 billion and $25 billion. Construction is under way on a gigawatt-scale AI factory in Independence, Missouri, and the company has secured a 1.2-gigawatt power agreement for another site in Pennsylvania, with phased delivery starting in 2027.

Wall Street remains split on what all this means for the valuation. Citi’s Tyler Radke raised his target to $287, a 70% jump, and kept a buy rating, citing robust demand for AI cloud services and rising GPU pricing. Citizens analyst Greg Miller increased his target to $270 from $175 and highlighted hypergrowth characteristics alongside the company’s control over technology and power. Morgan Stanley was more cautious, edging its target up to $144 from $126 with an equal-weight rating, acknowledging better execution but less upside.

Nebius has guided for full-year 2026 revenue of $3 billion to $3.4 billion and an annualized recurring revenue run rate of $7 billion to $9 billion. The next test is whether the Clarifai and Eigen AI integrations can indeed lift inference margins even as the hyperscalers – armed with new capital from the Google-Blackstone venture – crowd into the same space. For now, the company is betting that owning more of the software stack will keep its massive capacity build from becoming a commodity race.