CoreWeave and Nebius Touted as Top AI Data‑Center Stocks Amid Explosive Backlog Growth
Investor research flags CoreWeave and Nebius as the premier AI‑focused cloud infrastructure plays, pointing to a $104 billion contracted backlog for CoreWeave and a $40 billion backlog for Nebius. Both firms are projected to post double‑digit revenue growth through 2028 despite mounting debt and rising political resistance to data‑center construction.
Why It Matters
CoreWeave and Nebius exemplify the emergence of AI‑native infrastructure as a distinct SaaS vertical, where compute is packaged, sold, and monetized through subscription‑style contracts. Their massive backlogs signal sticky, high‑margin revenue streams that can fund further data‑center rollouts, creating defensible moats against hyperscalers. At the same time, the surge in community opposition highlights a new regulatory risk layer that could reshape cost structures and geographic expansion strategies for AI‑infrastructure providers.
For SaaS operators, the story underscores the importance of building product‑led growth engines around mission‑critical infrastructure, while also preparing for non‑technical headwinds—political, environmental, and community‑based—that can impact capital deployment and unit economics. The valuation gaps suggest opportunities for investors willing to bet on execution, but also warn of the volatility inherent in capital‑intensive, rapidly scaling SaaS businesses.
Key Points
- CoreWeave FY 2025 revenue $5.1 B (+168% YoY); projected 2026 revenue $12.9 B
- Nebius projected revenue CAGR 253% (2025‑2028) with $40 B backlog
- CoreWeave backlog $104 B (246% YoY); Nebius backlog $40 B (4× YoY)
- CoreWeave debt‑to‑equity 14.3×; Nebius maintains low leverage (0.1×)
- Data‑center opposition blocked $68 B (Q2 2026) and $130 B (Q1 2026) in projects
Analysis
The rapid ascent of CoreWeave and Nebius reflects a broader shift from generic cloud services to specialized AI‑compute platforms that operate on a SaaS model—selling capacity, managed services, and software integration through recurring contracts. This verticalization creates higher net‑retention rates and pricing power, especially as large enterprises lock in multi‑year commitments to meet the compute demands of LLM training and inference. However, the capital intensity of building and maintaining GPU‑dense data centers introduces a classic SaaS paradox: the need for massive upfront capex to secure future subscription revenue. Companies that can efficiently convert backlog into billable ARR while managing debt will set a new benchmark for AI‑infrastructure economics.
Regulatory resistance adds a layer of strategic complexity. Unlike traditional SaaS firms that scale primarily through software development, AI‑infrastructure providers must secure physical sites, negotiate local utility contracts, and address community concerns about energy consumption and environmental impact. The $68 B to $130 B of delayed projects signals that future growth may be increasingly contingent on political capital as much as technical prowess. Firms that develop modular, portable data‑center solutions or partner with jurisdictions offering favorable incentives could mitigate these risks.
From an investor perspective, the valuation disparity—CoreWeave trading at ~7× adjusted EBITDA versus Nebius at ~15×—suggests differing market expectations about execution risk and geographic diversification. CoreWeave’s larger footprint and marquee customer list provide visibility but also concentration risk, especially with Microsoft accounting for 67% of FY 2025 revenue. Nebius’s more diversified vertical focus may offer a smoother path to profitability, albeit with a smaller scale. As the AI compute market is projected to expand at a 58% CAGR through 2031, the winners will be those that blend aggressive infrastructure rollout with disciplined financial management and proactive stakeholder engagement.