**[Analysis: Understanding Five Major Trends in the AI Cloud Computing Power Sector, Bottlenecks Shift from Customers to Power Delivery, Price Hikes Driven by Short-Term Emergency Capacity Rather Than Traditional Bare Computing Power]** BlockBeats News, August 13 — Analyst qinbafrank stated that five major trends in the AI Cloud Computing Power (CSP) sector can be distilled from the latest financial reports of CoreWeave and Nebius. Demand continues to significantly exceed near-term deliverable supply—CoreWeave had approximately 1.5GW of operational power and 3.7GW of contracted power by the end of Q2, with a backlog of orders worth around $104 billion, excluding over $25 billion in new commitments added at the beginning of Q3. Nebius plans to deploy over 1GW annually starting in 2027, with a mid-term goal of 5GW in contracted capacity. The industry's bottleneck has shifted from "whether there are customers" to "when power can be delivered, GPUs deployed, and billing initiated." Pricing power is indeed strengthening, but the increases are not uniform: CoreWeave raised prices for various SKUs by approximately 25% in July; Nebius saw prices for older-generation GPUs rise by over 30% compared to Q1, while newly signed contracts in Q2 averaged annualized revenue of over $20 million per MW. At the beginning of Q3, short-term emergency capacity even reached $40 million to $50 million per MW. The most significant price hikes are seen in short-term capacity, next-generation GPUs, large-scale clusters, and production-grade inference, rather than traditional low-priority, long-term locked bare computing power. Project-level ROI is now calculable, but company-level ROIC has yet to be proven. Nebius disclosed relatively clear project payback periods for the first time, while CoreWeave uses five-year contracts and asset-level financing to cover GPU investments, though depreciation and interest continue to consume most of its operating profits. Both companies are upgrading toward an "AI infrastructure operating system," with the competitive focus shifting from hourly GPU rentals to providing integrated platforms for training, inference, storage, networking, model deployment, monitoring, governance, agent runtime environments, and cross-cloud operations. Paths for introducing third-party capital are diverging: Nebius's asset-light model is closer to "capital partners funding the construction of AI factories, with Nebius providing the operating system and operational capabilities," while CoreWeave Omni leans toward "deploying a complete cloud infrastructure on customers' own data centers and GPUs." While the direction is the same, Nebius currently emphasizes reducing capital investment, whereas CoreWeave focuses more on hybrid cloud and sovereign AI delivery. [Original Link]