Nebius Q2: $5.6B CapEx vs Revenue, Upfront Payments, and the Reality of AI Cloud Expansion

Woofun AI reports that AI cloud infrastructure company Nebius released its second quarter earnings press release and shareholder letter on August 12, covering the period ended June 30. The core financial tension emerged immediately: while the group recorded revenue of $582.3 million, cash expenditures for acquiring fixed assets and intangible assets surged to $5.6574 billion. This disparity highlights the fundamental challenge of funding massive capital investments in GPUs and data centers before corresponding service revenues are realized.

Although Nebius disclosed four significant AI cloud contracts with an average total contract value exceeding $1 billion, the narrative of seamless growth obscures the asynchronous movement of contracts, cash, equipment, and revenue. The critical question is not merely the scale of demand, but how customer commitments translate into immediate funding for capacity expansion versus what remains locked in future contractual terms or management estimates.

The financial snapshot for the quarter reveals a stark contrast between income generation and capital deployment. Revenue stood at $582.3 million, a figure dwarfed by the $5.6574 billion spent on fixed assets, intangible assets, GPUs, and data centers. This expenditure reflects the heavy upfront nature of building AI cloud infrastructure, where physical assets must be procured and installed long before they generate billable usage. The disclosure of four significant AI cloud contracts, each averaging over $1 billion in total value, underscores the scale of demand driving these investments.

However, these contract values do not equate to immediate cash inflows or recognized revenue. Instead, they represent future obligations that require substantial capital outlay today. The mismatch between the $582.3 million in recorded revenue and the $5.6574 billion in capex illustrates the intense pressure on liquidity during periods of aggressive infrastructure build-out.

Revenue growth trajectory further clarifies the dominance of the AI cloud segment within Nebius's operations. In the second quarter of 2026, group revenue climbed to $582.3 million, a significant increase from the $227.7 million reported in the fourth quarter of 2025. This surge is primarily driven by the AI cloud business, which contributed $574.9 million in the second quarter, accounting for nearly all of the group's total revenue. The rapid acceleration from $227.7 million to $582.3 million in just two quarters demonstrates the scalability of the AI cloud model once capacity is deployed.

However, this growth is concentrated in a single business line, making the company's financial health heavily dependent on the continued expansion and monetization of its AI cloud services. The near-total reliance on the AI cloud business for revenue generation amplifies the importance of understanding the specific dynamics of this segment's financial metrics.

Understanding AI cloud ARR requires distinguishing between a speedometer and locked-in revenue. The $3 billion AI cloud ARR figure represents an annualized rate calculated by multiplying the AI cloud revenue from the last month of the quarter by 12. This metric grew by 56.3% quarter-on-quarter, outpacing the group's overall revenue growth of 45.9%. While the 56.3% increase appears robust, it is crucial to recognize that ARR is not a measure of contracted revenue for the next 12 months.

Instead, it reflects the velocity of the business at the end of the quarter. The 45.9% overall revenue growth, by contrast, captures confirmed income across the entire quarter. Presenting these two figures together highlights the time gap between demand trends and recognized revenue, rather than suggesting a direct correlation between ARR and total order volume. The higher ARR growth rate indicates accelerating momentum in the final month, but it does not guarantee equivalent revenue realization in subsequent periods.

Profitability metrics reveal a divergence between operational performance and accounting losses. Non-GAAP adjusted EBITDA rose to $236.2 million in the second quarter, indicating strong underlying operational efficiency.

However, the GAAP net loss for continuing operations stood at $190.4 million, reflecting the impact of non-cash and structural costs. Adjusted EBITDA excludes depreciation, amortization, equity incentives, transaction-related costs, and interest, providing a clearer view of core business profitability. Despite the positive adjusted EBITDA, the GAAP net loss underscores the heavy burden of capital-intensive operations. Rising revenue does not automatically translate to cash profits, as significant expenses are incurred before services are delivered. The $236.2 million in adjusted EBITDA demonstrates that the business model is operationally viable, but the $190.4 million net loss highlights the financial strain of continuous investment in infrastructure and the accounting treatment of such expenditures.

Woofun AI data shows that cash flow dynamics expose the limitations of deferred revenue in covering capital expenditures. In the first quarter of 2026, changes in deferred revenue exceeded capital expenditures, but this trend reversed sharply in the second quarter. According to Nebius's unaudited consolidated financial statements, changes in deferred revenue for the first half of the year totaled $4.395 billion, while capital expenditures reached $8.1303 billion. The second-quarter change in deferred revenue, derived by subtracting the first-quarter figure from the first-half total, was significantly lower than the capex for that period.

This discrepancy means that operating cash flow, while positive, was insufficient to cover the massive capital outlays. Capital expenditures were approximately 9.7 times the quarter's revenue, illustrating the intense cash burn required for expansion. Deferred revenue changes affect operating cash flow but do not represent the full extent of upfront payments received, nor can they be directly equated to customer deposits for a specific quarter. The imbalance between $4.395 billion in deferred revenue changes and $8.1303 billion in capex highlights the funding gap that must be bridged through other means.

Contract economics provide a more nuanced view of upfront payment coverage and payback periods. The four significant AI cloud contracts signed in the quarter had an average total contract value (TCV) of over $1 billion, describing the scale of commitments rather than recognized revenue. Approximately 70% of all transactions closed in the second quarter included customer upfront payments, which covered 50% to 60% of the associated capital expenditures. This coverage ratio does not apply to every individual contract, nor does it imply that all of Nebius's capex is funded by customers.

Nebius estimates the payback period for capital expenditures and operating costs related to these new contracts at 1 year and 10 months, with a historical range of 2 to 3 years. This estimate relies on projected costs and future capacity that has already been contracted but not yet built. The company disclosed a contract value of $20 million to $25 million per megawatt-year, calculated based on revenue recognition without including upfront payments. These figures illustrate the complex interplay between upfront cash, long-term commitments, and the timeline for cost recovery.

Future capacity deployment and revenue recognition timelines further clarify the lag between investment and income. The capacity corresponding to most transactions closed in the second quarter will be gradually deployed by the end of 2026, contributing mainly to revenue in 2027. This delay means that upfront payments and capital expenditures precede revenue recognition by a significant margin. Upfront payments alleviate some funding needs for capacity expansion, but equipment installation and infrastructure build-out occur before any income is recorded on the income statement.

Revenue only appears after capacity is actually delivered and utilized by customers. This sequence creates a temporal mismatch where cash outflows for equipment and construction happen well before the corresponding inflows from service usage. The timeline underscores the importance of managing liquidity during the build-out phase, as revenue recognition lags behind both capital deployment and customer commitments.

Forward-looking estimates highlight the scale of customer commitments and power targets. Customer commitments exceeding $40 billion reflect contractual pledges that indicate future fulfillment and revenue trends. According to the company's shareholder letter, customer upfront payments are expected to exceed $9 billion in 2026, a management estimate rather than cash already reflected in the second-quarter balance sheet. Another key metric is the 5GW contracted power target expected to be achieved by the end of 2026, supported by existing land and power commitments.

This 5GW figure is not the current GPU capacity online, nor can it be directly converted into hash rate available to customers in the current quarter. It represents the physical infrastructure capacity that Nebius aims to bring online, which will support future GPU capacity and hash rate offerings. The distinction between current operational metrics and forward-looking targets is critical for assessing the company's growth trajectory and resource allocation.

The misalignment of growth metrics in capacity expansion requires careful interpretation to avoid inflated perceptions of financial health. Contracts, upfront payments, equipment, and revenue enter the capacity expansion process at different speeds, creating a complex web of financial indicators. Adding these figures together to derive a single growth metric would obscure the underlying realities of cash flow timing and revenue recognition. The $582.3 million in quarterly revenue, the $3 billion in AI cloud ARR, the $40 billion in customer commitments, and the $9 billion in expected upfront payments each tell a different part of the story. Recognizing the distinct roles of these metrics is essential for understanding Nebius's financial position.