$NBIS --- In early June, $NBIS teamed up with NVIDIA to launch a new program powering robotics startups across the UK and Europe with NVIDIA’s latest Blackwell architecture (RTX PRO 6000 Blackwell GPUs). As a deep NVIDIA partner, $NBIS gets top-priority access to NVIDIA’s cutting-edge high-performance AI chips (H200, Blackwell, etc.)—a moat no second-tier cloud provider can match.
Also in early June, $NBIS announced plans to invest ~£1.7 billion (~$2.16 billion) in UK AI infrastructure. It locked in a 10-year, 22MW partnership with Kao Data to deploy a large-scale AI cloud platform in Harlow, with existing customers including fintech unicorn Revolut. Nebius also secured 1.2 GW of power and land in Pennsylvania to build a massive new AI factory.
Back in mid-May, $NBIS announced a long-term partnership with Bloom Energy ($BE) to deploy 328 MW of solid oxide fuel cells, delivering behind-the-rack power to solve data centers’ biggest pain points: crippling power shortages and grid interconnection queues. The deployment goes live this year.
1. The "Arms Dealer" for AI Unicorns
Hyperscalers like Microsoft and Google build cloud services for their own ecosystems and Fortune 500 enterprises. But Meta, xAI, and hundreds of mid-to-large AI startups are desperate for pure, efficient, cost-effective GPU compute clusters—exactly what Nebius’s Neocloud delivers.
Nebius sells full-stack AI cloud services, full stop. It never competes with customers on model development, it just sells compute. Its main rival is CoreWeave (currently prepping for IPO), making $NBIS one of the only pure-play AI compute assets publicly traded in the U.S.
2. Blue-Chip Order Backing = Self-Reinforcing Growth Flywheel
Nebius already has publicly disclosed contracts with mega-customers including Meta and Microsoft. With these internet giants as anchor assets, Nebius can raise capital at rock-bottom rates via asset-backed financing and convertible debt. That creates a seamless high-growth flywheel: raise capital → buy GPUs → rent to giants for profit → raise more capital to scale even faster.
3. Elite Engineering DNA From the ex-Yandex Dream Team
This isn’t some shell compute company thrown together overnight. The restructured Nebius inherited over 1,000 elite European engineers from Yandex. They have a decade+ of experience running hyperscale search architecture, building in-house distributed systems, and developing proprietary hardware liquid cooling. That’s why Nebius’s data centers deliver far better orchestration efficiency and Power Usage Effectiveness (PUE) than generic industry peers.