CoreWeave (CRWV.US) 2026年第二季度业绩电话会
文章语言：
简
繁
EN
Share
Minutes
原文
会议摘要
Corbe Cloud and Core Weave report significant growth in AI-related services, with Core Weave achieving record revenue of $2.6 billion, a 500 MW increase in active power capacity, and surpassing 1 billion model training runs. Both companies emphasize expanding margins, strategic opportunities, and robust demand for AI cloud infrastructure, with Core Weave on track for 8 GW of active power by 2030. They highlight innovative partnerships, strong customer engagement, and a managed inference business, reflecting their commitment to continuous AI development and deployment at scale.
会议速览
Core We's Q2 2026 Earnings Call: Forward-Looking Statements and Non-GAAP MeasuresThe dialogue introduces Core We's second quarter 2026 earnings call, emphasizing the inclusion of forward-looking statements and non-GAAP financial measures, while inviting questions and providing guidance on accessing the call's replay.
Core Weave's Q2 Success: AI Market Expansion and Strategic GrowthCore Weave reports record revenue, capacity growth, and customer demand in Q2, highlighting strategic execution, AI market development, and broadening customer adoption across sectors, positioning for long-term market share gains.
Revolutionizing AI Operations: Continuous Model Improvement and Platform InnovationsAI operations have evolved from one-time model deployments to continuous improvement loops, necessitating ongoing compute resources. Our AI-native platform facilitates this shift, offering integrated services from model training to deployment and monitoring, accelerating innovation and customer success. With over 1 billion model training runs tracked, we're at the forefront of AI advancement, fostering a community of developers and researchers. Our managed inference platform, bolstered by recent industry-first innovations, is experiencing rapid growth, constrained only by capacity, highlighting the expanding demand for AI infrastructure services.
Core Weave's AI Platform: Balancing Cost, Speed, and Scale for Enhanced RevenueCore Weave offers a specialized AI platform, optimizing cost and speed for customers deploying models at scale, resulting in significant ARR growth. Its comprehensive approach from power to managed services supports the continuous AI lifecycle, distinguishing it as a purpose-built solution.
Core Weave's Leadership in AI Cloud: Scaling for Future Demand with Enhanced Security and EconomicsCore Weave has achieved significant milestones in AI cloud infrastructure, including first-to-market innovations, record ML performance, and enhanced security. With a focus on vertical integration, global expansion, and securing critical resources, the company is poised for continued growth, aiming for 8 GW by 2030 and reinforcing its position as the essential cloud for AI.
Strong Demand and Strategic Capacity Allocation Drive Growth and Margins in Q2Discussed exceptional Q2 results, highlighting robust demand for GPU and AI services, disciplined capacity allocation, margin improvements, and strategic pricing adjustments. Emphasized growth in non-GPU businesses and long-term customer relationships, with cash flow implications noted.
Long-Term Returns and Recontracting Opportunities in AI Cloud Infrastructure InvestmentsDiscusses the financial benefits and opportunities of AI cloud infrastructure investments, highlighting strong returns post-initial contract terms, the value of recontracting, and the potential for further upside in a supply-constrained market. Emphasizes the proven ROI of existing assets and the growing demand for AI cloud services.
Strong Q2 Revenue Growth and Strategic Capital DeploymentThe dialogue highlights significant Q2 revenue growth of 112% year-over-year, driven by strong customer demand for AI solutions. It details strategic investments in infrastructure, a robust balance sheet with $6.9 billion in cash and equivalents, and a diversified capital structure including $18 billion raised through various financial instruments. The company anticipates further growth, with a focus on scaling operations and managing supply chain complexities.
NVIDIA's Strategic Financial Milestones and AI ExpansionAnnouncement of a significant term loan enhancing NVIDIA's ability to scale enterprise services and AI infrastructure, leading to reduced debt costs, increased revenue guidance, and a strong outlook for 2026.
Renewal Opportunities and Supply Chain Strategies for AI InfrastructureDiscusses renewal of older AI contracts, potential customer intentions, and the value of legacy infrastructure for AI use cases. Also covers strategies for securing long-term supply chain agreements to support capacity expansion.
Managing Supply Chain Complexity for AI-Driven Product DeliveryDiscusses the importance of maintaining strong supplier relationships and managing a complex supply chain to deliver high-quality AI products, highlighting increased output value and margin expansion due to efficient input management.
Expanding Managed Inference Offerings: Lessons Learned and Future Allocation StrategiesDiscusses the successful scaling of managed inference products, emphasizing the importance of diversifying offerings for higher margins and software solutions. Highlights the strategic use of off-contract GPUs to maximize value, leveraging control over silicon for market advantage.
Factors Driving Higher Margins in Short-Term Contracts and Market Pricing TrendsDiscussed drivers behind 5%-10% better margins on recent deals, attributing it to the superior value of delivered infrastructure, enhanced by quality, reliability, security, and TCO. Also highlighted clients' willingness to pay higher margins for access to premium compute resources, reflecting broader market pricing trends.
Navigating Regulatory Challenges and Growth Targets in Data Center ExpansionDiscusses upward pricing pressures and reaffirms ARR targets for 2026, while addressing regulatory hurdles in deploying GW-scale data centers, emphasizing community engagement and infrastructure development.
Adapting Data Centers for AI Inference and Training NeedsThe dialogue explores the necessity of evolving data center infrastructure to accommodate growing AI inference demands, emphasizing the importance of building versatile AI infrastructure that supports both training and inference processes, ensuring future capacity and efficiency.
Navigating AI Workloads: Edge vs. Cloud in a Competitive MarketDiscusses the evolution of AI workloads between edge and cloud, emphasizing the flexibility of their platform to cater to both low-latency and scale-sensitive demands, highlighting a competitive advantage through informed client feedback and growing market trends.
Analysis of Power Addition Linearity and Vera Rubin's Monetization PotentialDiscussed the linearity of power additions, noting a significant 500 MW increase, with 300 MW in June alone, impacting Q3 and Q4. Early market signals indicate strong demand and pricing power for Vera Rubin, expected to drive a 5% to 10% margin expansion, showcasing its potential as a successful SKU.
Strategies for Recontracting CPU Fleets and Leveraging Inference OpportunitiesThe dialogue discusses strategies for managing older generation CPU fleets as contracts roll off, emphasizing the dynamic nature of the inference market. It highlights the decision-making process for recontracting or placing assets on the spot market, considering both short-term opportunities and long-term company stability.
Expanding Market Reach Through Flexible Contract Durations in Compute LeasingThe dialogue highlights the strategy of offering both long and short-term contracts to maximize profit and accessibility in the compute leasing market, addressing enterprise clients' preference for shorter contract cycles.
Closing Remarks and Gratitude for Team, Customers, and PartnersThe speaker concludes the call by expressing gratitude to the team, customers, and partners for their contributions to Core Weave's achievements. Anticipation for future progress is expressed, thanking all attendees for their support and looking forward to future updates.
要点回答
Q：What were the financial highlights of Core Weave's second quarter?
A：Core Weave reported record revenue of $2.6 billion, up 112% year over year, and increased revenue backlog to $100 billion. The company also experienced rapid capacity growth and expanded its enterprise adoption.
Q：What progress has Core Weave made in terms of power strategy and AI platform capabilities?
A：Core Weave has reached 1.5 GW of active power, adding nearly 500 MW in the quarter, more than any quarter in its history, with plans to reach at least 10 GW by 2026. The company grew adjusted operating income to $128 million, with margins expanding as technology stack broadens and AI platform capabilities are delivered, including multiple industry first milestones.
Q：What is Core Weave's vision for the future of AI and how is it guiding the company's strategy?
A：Core Weave believes the AI era is here and will touch every part of the global economy, transforming every organization. The company's vision guides its strategy, product road map, capital allocation, partnerships, and customer service, focusing on broadening demand and customer adoption, continuous AI development tools, performance and economics, and a Power and Supply Chain Foundation for future growth.
Q：What are some examples of industries that are expanding their use of AI with Core Weave?
A：AI demand is intensifying across various sectors including life sciences, with new customers like Isomorphic Labs; financial services, with customers such as Flow Traders and IMC; and the public sector, with a collaboration with Leos to deliver secure AI capabilities for defense and national security.
Q：How is AI moving from experimentation into core operations within organizations?
A：AI is transitioning from experimentation into core operations, and the companies that act decisively are creating an advantage. With AI moving into production, the way applications are built is changing, and the leaders will be those who can learn and iterate the fastest.
Q：What is the new approach to managing AI models in production described in the speech?
A：The new approach involves a continuous loop of training, inference, evaluation, and improvement where real-world data informs model evaluation, driving new experiments that improve the model before redeployment into production. This creates a feedback loop that compels ongoing improvement and capability compounding.
Q：What are the core components of the AI native platform mentioned?
A：The AI native platform comprises cutting-edge cloud infrastructure, a managed inference business, leading developer tooling and agent solutions, and an orchestration and observability layer powered by mission control.
Q：What were the achievements of the AI platform in Q2?
A：In Q2, the AI platform introduced 7 new capabilities, achieved multiple industry firsts, and saw strong adoption, including over 1 billion model training runs tracked. It also experienced growth in its managed inference platform, with a significant increase in booked revenue.
Q：How does Core Weave facilitate monetization for companies and what are the benefits for customers?
A：Core Weave allows companies like Grammarly and u dot com to move from experimentation to real production traffic, run AI coding agents, serve their fine-tuned models, and deploy open weights models at scale. This flexibility in consumption and monetization benefits customers by providing a total cost of ownership advantage and enables faster deployment and better application performance.
Q：Why is Core Weave purpose-built for AI according to the speech?
A：Core Weave is purpose-built for AI because it addresses the unique needs of the AI lifecycle with a focus on power cooling and rack design, networking, orchestration, observability, developer tools, and managed services. Its depth, breadth, and technical capability are tailored to support the continuous AI lifecycle.
Q：How does Core Weave ensure enterprise-grade security and reliability?
A：Core Weave delivers enterprise-grade security and reliability by integrating capabilities with a broader portfolio of services across data centers, and has been recognized as a visionary in the Gartner 2026 Magic Quadrant for Cloud AI Infrastructure.
Q：What is the strategy for achieving scale and operating leverage?
A：The strategy involves securing power sites, cooling hardware, storage, networking, and supply chain inputs well ahead of need and operating them as an integrated system. This approach is working to transform scale into operating leverage, resulting in expanding margins.
Q：What is the significance of the new power contracts and potential capacity?
A：The significance lies in the ability to sustain growth over a multi-year horizon. The company ended up with 3.7 GW of contracted power and plans to add more, excluding 1.5 GW of potential power from powered land. This, along with global expansion and long-term agreements, supports enhanced long-term margins and meets customers where they operate.
Q：What momentum is Core Weave experiencing in the second half of the year?
A：Core Weave enters the second half of the year with more momentum than at any point in its history, supported by AI reshaping every industry and a diverse set of customers with improving operating leverage and visibility.
Q：What are the financial implications of the pricing changes and the increase in customer spending?
A：The pricing changes, which include a 25% increase across SKUs, are in response to the strong demand environment and the increasing return on investment observed by customers in the Core V platform. Customer spending on Core V has increased, contributing to higher operating margins before the July pricing changes.
Q：How is the economics of a typical 500-year contract structured for Core Weave?
A：A typical 500-year contract has strong and still expanding unit economics across its term. However, the cost, primarily in the form of CapEx, is front loaded requiring financing through debt, customer prepayments, and other capital sources for the build-out. Once the cluster is delivered, contracted revenue ramps up, becoming predictable and cash flow generative.
Q：What is the projected revenue impact of renewing or recontracting cloud infrastructure?
A：Recontracting offers the potential to deliver strong returns, as the company is largely sold out of prior generations of Nvidia GPUs and the current ones. Every resale or renewal adds to the returns already earned within the initial term. The economics do not rely on recontracting after the initial customer term and there is potential for significant further upside with tailwinds such as strong demand and constrained supply of new AI cloud infrastructure.
Q：What were the financial results for Q2, and how do they compare to the previous year and the expectations?
A：Revenue for Q2 was $2.6 billion, up 112% year over year and 24% sequentially. The adjusted EBITDA was $1.5 million, doubling from Q2 of the previous year, with a margin of 59%. Adjusted operating income was $128 million, up from $21 million in the last quarter and above the high end of guidance. The net loss for Q2 was $626 million, compared to a net loss of $290 million in the prior year's Q2. Despite the net loss, the income tax provision was made due to the valuation allowance on net deferred tax assets.
Q：What is the outlook for future CapEx and debt levels for Core Weave?
A：Core Weave's CapEx in Q2 was $9.4 billion, slightly above the high end of the guided range, reflecting customer deliveries and construction in progress. Cit increased to $11.9 billion from $9.6 billion, signaling a significant amount of PPE expected to be deployed early in Q3. Core Weave has made significant progress in strengthening its balance sheet, expanding access to capital, and raising approximately $18 billion through various debt, convertibles, and equity transactions, including several firsts like an inaugural Eurobond and a HPC infrastructure delayed term loan.
Q：How is the recent financing expected to impact Nvidia's ability to serve the enterprise market and its managed inference platform?
A：The financing is expected to unlock Nvidia's ability to serve a critical part of the enterprise market scale, accelerate the ramp of its managed inference platform, and allow growth of exposure to shorter dated contracts which typically have higher ASPs and margins.
Q：What are the expected Q3 financial figures for Nvidia?
A：Q3 revenue is expected to be in the range of 3.45 to 3.6 billion, Q3 adjusted operating income of 200 to 260 million, and Q3 interest expense is expected to be in the range of 860 to 940 million.
Q：What is the anticipated end-of-year annualized run rate revenue for Nvidia?
A：The anticipated end-of-year annualized run rate revenue for Nvidia is raised to 18.5 to 19.5 billion.
Q：How is the demand environment for Nvidia's technology stack and what progress has been made in the capital structure?
A：The demand environment for Nvidia's technology stack is strong, evidenced by strategic expansion of the customer base to support the next wave of enterprise AI adoption. There has been progress in the capital structure, including reduction of the weighted average cost of capital and securing financing for long-term growth.
Q：What details were provided regarding the renewal opportunities and the installed base of equipment for Nvidia?
A：While specific figures were not provided, the speaker mentioned that the older generations of infrastructure continue to have significant value for various AI use cases. The company has seen a mix of customer intent regarding contract periods for renewal, with some customers opting for shorter terms. They also mentioned that the managed inference platform is expected to grow rapidly, with an aim to reach $250 million in Adjusted Recurring Revenue (ARR) by the end of the year.
Q：How is Nvidia managing the supply chain and what is the strategic approach regarding long-term agreements?
A：Nvidia is managing the supply chain by building long-standing deep relationships with original design manufacturers (ODMs), original equipment manufacturers (OEMS), and companies that supply memory, among others. The strategic approach includes assessing what is necessary to ensure the company has the required infrastructure to deliver products at an acceptable price and quality. The company aims to secure long-term agreements to assure supply and support the execution of power-only capacity needs.
Q：How does the speaker's company plan to meet the evolving needs of AI infrastructure?
A：The company plans to meet the needs of AI infrastructure by building AI-specific infrastructure that includes all necessary components for serving the full range of AI use cases, optimizing for the cycles and movements of data to clients and consumers.
Q：What are the company's views on AI workloads and the role of edge computing?
A：The company believes that some workloads will be served from the edge and others do not require the same level of latency protection, while the cloud is built to accommodate both types of workloads. They plan to continue to build in this fashion, adapting to clients' needs for more edge or less latency-sensitive computing.
Q：How is the company responding to the competitive market and what is the impact on their business?
A：The company is responding to the competitive market by seeing increasing demand, pricing, and margins across their products, signaling growth in their core business and in the massive trend they are a part of.
Q：What insights does the company have regarding the performance and contribution of the Vera Rubin generation?
A：The company has insights that the Vera Rubin generation has shown margin expansion from the start and that the demand for this platform is substantial, with strong pricing power being realized by the company.
Q：How does the company manage the transition of older generations of CPUs off contract and onto new ones?
A：The company manages the transition by scaling their inference product and exploring the managed inference opportunity. They decide whether to renew existing contracts or place the infrastructure back into term contracts based on economic feasibility and the long-term stability and obligations of the company.
Q：What strategy is the company employing with short-term and long-term contracts in response to market demand?
A：The company is employing a strategy to sell compute on both long-term and short-term contracts, aiming to extract additional margin with the latter. They have been aggressive in introducing products like the 5.5 year DDT L to support this strategy and align with the evolving needs of the market.
Q：How does the new contract duration of 5.5 years benefit the company and its clients?
A：The new contract duration of 5.5 years benefits the company and its clients by allowing the company to diversify its contract terms and enter into new markets with clients seeking shorter-term commitments, thus providing a solution that aligns with their consumption cycles.
