To The Moon

On November 10, $CoreWeave, Inc.(CRWV)$ released its third-quarter earnings report, reporting revenue of approximately $1.36 billion, marking over 100% year-over-year growth and slightly exceeding market expectations (around $1.29 billion). However, its adjusted gross margin/ adjusted operating margin faced pressure, declining from about 21% in the same period last year to approximately 16%. Management reaffirmed robust demand for large-scale AI infrastructure while noting accelerated capital expenditures and expansion pace. Overall, the company maintains strong growth momentum, but volume significantly outpaces pricing, with lingering concerns over margins and cost structures. Specifically, let's examine the core information in the financial report. Revenue reached $1.365 billion this quarter, surging 134% year-over-year (compared to $584 million in the same period last year) and growing approximately 25% quarter-over-quarter, significantly exceeding the market consensus of $1.29 billion. The primary drivers stemmed from explosive demand for AI training and inference, notably Meta's $14.2 billion mega-order and OpenAI's $6.5 billion expansion contract (totaling $22.4 billion), alongside the sixth agreement with a leading hyperscale cloud provider; In terms of business structure, significant diversification among clients is evident, with the share of a single customer dropping from 85% at the beginning of the year to below 35%. Investment-grade clients now account for over 60% of the customer base, signaling a shift from reliance on a few AI giants toward broader adoption across enterprises, governments, and public sectors. Adjusted EBITDA reached $838 million, surging 121% year-over-year with a margin of 61% (down from 65% in the same period last year). This exceeded market expectations, driven by economies of scale and improved operational efficiency, though the marginal decline reflects pressure from interest expenses (projected at $1.21-$1.25 billion for the full year). The key shift lies in the rising contribution from software and services revenue, driving the evolution from pure GPU leasing to a full-stack AI platform and strengthening profit resilience. Revenue backlog surged to $55.6 billion, nearly doubling from the previous quarter and nearly quadrupling year-over-year; this far exceeded expectations, driven by $25 billion in new backlog added during Q3, primarily from large renewals and new customer onboarding. The logic lies in AI supply falling short of demand, with the company holding an advantage as a pioneer specializing in GPU cloud services. However, this also exposes sensitivity to AI cycles, where demand fluctuations could amplify risks. In terms of profitability, the company posted a net loss of $110 million, or a loss of $0.22 per share, representing a significant improvement year-over-year (compared to last year's $360 million loss) and far exceeding the expected loss of $0.57 per share. Capital expenditures totaled $3.3 billion, below expectations due to delayed procurement caused by project setbacks. Improvements stemmed from revenue scaling, though high Capex underscores cash burn pressures inherent in the capital-intensive model. Structurally, debt financing included $1.75 billion in new senior notes and $2.6 billion in extended drawdown loans, optimizing capital costs while eliminating near-term maturity risks. Active power capacity reached 590 megawatts, with a sequential increase of 120 megawatts; exceeding expectations, driven by the launch of eight new data centers in Europe and the United States, as well as the early deployment of NVIDIA GB300 NVL72 systems. A notable shift is the increased proportion of self-built data centers, aimed at reducing reliance on third parties, with the annual target raised to over 850 megawatts. Key Investment Considerations Strong revenue growth but high expansion costs The company achieved revenue of approximately $1.36 billion in the third quarter, significantly exceeding analysts' prior expectations of around $1.29 billion. This performance reflects its rapid expansion capabilities as an AI infrastructure provider—with contracts such as the $14 billion agreement with Meta Platforms and the $6.5 billion deal with OpenAI further driving demand. The robust revenue growth confirms demand momentum. However, given the rapid expansion and capital-intensive nature of the business, the pace of "investment versus returns" remains a key focus area. Signs of pressure on profit margins are evident. Despite substantial revenue growth, the company reported an adjusted operating margin of approximately 16%, a significant decline from 21% in the same period last year. Contributing factors include rising costs for GPU/AI infrastructure, increased expenditures on data center construction and power facility expansion, and higher unit costs due to equipment supply constraints. Profitability structure has yet to improve in tandem, with margins facing near-term pressure under the "high growth, high investment" model. In other words, while volume growth is assured, improvements in pricing and structural efficiency will take time. Capital expenditure and expansion scale have entered a critical phase. Company management previously disclosed during the Q2 earnings call that full-year capital expenditures for 2025 are projected to reach $12 billion to $14 billion. Concurrently, the company reported its current revenue contract backlog has reached approximately $30.1 billion, representing an 86% year-over-year increase. The high capital expenditure reflects the company's extreme optimism regarding AI infrastructure demand, but it also implies increased financial leverage, depreciation of fixed assets, and rising operational costs such as power and cooling. An overly rapid expansion pace could introduce execution and cost risks. Top clients are concentrated, but diversification is underway. The company disclosed significant agreements with Meta and OpenAI, noting that demand from mid-to-large-sized AI clients downstream is growing rapidly beyond its top-tier customers. Contract backlog figures have increased. On the other hand, the company has yet to provide detailed breakdowns of client share and renewal rates. Securing top AI clients demonstrates a competitive advantage, but high client concentration also poses risks. Should top clients gain greater bargaining power or competition intensify, it could undermine the company's pricing flexibility. CoreWeave's GPU-intensive AI cloud infrastructure business operates within a sustainable, long-term growth trajectory, akin to AWS's early positioning during the cloud computing wave. Its focus on high-performance computing offers enduring advantages amid the sustained expansion of AI large-model training and inference demands, particularly through penetration in the public sector (such as the newly launched CoreWeave Federal) and enterprise applications. However, some growth remains tied to short-term AI hype and topic-driven factors, such as large contracts with Meta and OpenAI. Should the AI investment cycle slow or shift toward edge computing, volatility could ensue. Compared to previous quarters, this period saw positive signals in customer diversification (eight out of nine major clients renewed contracts) and software stack expansion (e.g., RTX PRO 6000 instances), signaling a transition from infrastructure provider to ecosystem platform. Conference Call and Key Points Capital expenditures in Q4 2025 are projected to exceed the combined total of the first three quarters, indicating that "Q3 was not the peak of expansion" and that Q4 remains the peak investment period. The company continues to emphasize the "global shortage of AI infrastructure supply," positioning itself as an "accelerator for AI cloud and HPC infrastructure" by leveraging its GPU cluster advantages. We express cautious optimism regarding profit margin improvement, believing that future economies of scale and more efficient/lower-cost equipment will drive profitability enhancement. Overall, our view is that CoreWeave demonstrates strong growth potential. As a rapidly expanding player in the AI infrastructure sector, it has established a clear advantage in terms of scale. However, from an investment perspective, the following key considerations exist: 1. While scale is growing rapidly, pricing and cost control have not yet improved in tandem. If equipment/GPU/power costs rise further, profit recovery could be delayed. 2. High capital expenditures, substantial fixed asset investments, and elevated operating expenses have become the norm, exerting significant short-term pressure on cash flow and profitability. 3. While AI infrastructure demand remains robust, uncertainties persist regarding the global economy, interest rates, equipment supply, and geopolitical factors (e.g., GPU supply chains). 4. The company has garnered high market expectations with widespread anticipation for growth. Should it fail to deliver improvements in its profit structure over the next few quarters, the risk of valuation pullbacks cannot be overlooked.