3 AI Infrastructure Stocks With 100% or Greater Growth Rates

CoreWeave, Nebius, and Applied Digital are all posting explosive AI infrastructure growth, with Nebius revenue up 547% year over year to $228 million, CoreWeave revenue up 110% to $1.6 billion, and Applied Digital revenue up 139% to $127 million. Management and market estimates point to continued rapid expansion, including Nebius's expected annual run rate rising to $7 billion-$9 billion by end-2026 and CoreWeave expecting at least $28 billion of revenue recognition over the next 24 months. The article is broadly bullish on AI data center demand and these names' growth prospects.

Analysis

The first-order read is that capacity remains the bottleneck, but the second-order winner is whoever can lock in power, land, and financing before pricing power normalizes. In this setup, NBIS looks best positioned on product breadth: a fuller stack makes it stickier for developers and less exposed to pure GPU commoditization, which matters if customers begin arbitraging between providers on cost per training run rather than headline capacity. CRWV remains the cleaner demand proxy, but its concentration in a narrower offer means it is more sensitive to any future change in GPU availability, utilization mix, or hyperscaler procurement timing.

APLD is the more interesting market-structure trade: it benefits from being the picks-and-shovels landlord, but that also makes it a spread business between capex intensity and contracted rents. The key question is whether current contracts were signed before peak optimism; if so, incremental upside could come from lease re-pricing and faster pre-leasing of new sites, but if not, the stock may already embed much of the growth runway. The next 2-4 quarters matter more than the next few days because these names are being rerated on forward capacity additions rather than trailing revenue alone.

The contrarian risk is that growth remains real while returns get diluted by a wave of supply. Any acceleration in competitors adding power, shells, or GPUs could compress margins even as top-line growth stays elevated, which is a classic trap in infrastructure booms. Meta-style hyperscaler demand is supportive, but it also increases customer concentration risk and bargaining power over time; if one large customer pauses deployment, the market will likely punish the higher-beta names before fundamentals fully roll over.

Key Decisions for Investors

• Long NBIS vs. short CRWV on a 3-6 month horizon: favor the fuller-stack platform as the higher-quality growth compounder; target relative outperformance if customers prioritize integrated workflows over raw GPU access.
• Trade APLD as a structured call on data-center scarcity: buy on pullbacks after financing/lease milestones rather than momentum spikes, with a 6-12 month horizon and a tight stop if pre-leasing slows.
• Reduce outright CRWV exposure into strength unless utilization and backlog conversion remain accelerating; the risk/reward worsens if GPU supply loosens and pricing power shifts to buyers.
• For option risk, consider NBIS call spreads 6-9 months out to capture upside from revised 2026 capacity guidance while limiting multiple-compression risk if the market starts discounting future supply.
• Pair long AI infrastructure against short high-duration software that depends on the same capex cycle; if hyperscaler budgets stay elevated, infrastructure should outperform software with weaker near-term monetization.