CoreWeave's Q2 Revenue Surges 100% Amid AI Demand

CoreWeave, a cloud-based AI infrastructure service provider, reported a significant surge in inference workload usage during its second-quarter earnings call. The company's CEO highlighted that inference, which involves processing data through AI models to generate insights, is becoming a critical revenue driver for AI applications. This trend is evident in the substantial growth of CoreWeave's VFX cloud service product, Conductor, which saw a usage increase of over fourfold in the first half of 2025.

CoreWeave's diverse customer base is expanding rapidly, with increasing demand for AI applications across various industries, including media and entertainment, healthcare, and financial services. The company's Conductor service, which is used for visual effects, has seen a remarkable increase in usage, driven by partnerships with AI video generation startups like Moonvalley. Additionally, CoreWeave has expanded its collaborations with major financial institutions such as Jane Street, Morgan Stanley, and Goldman Sachs, further solidifying its position in the financial services sector.

To meet the growing demand and address structural supply shortages, CoreWeave is aggressively expanding its capacity. The company aims to achieve over 900 megawatts of power delivery capability by the fourth quarter, representing a significant increase from its current capacity. This expansion includes the addition of over 400 megawatts of new power capacity. The CEO emphasized that the company is focusing on "expanding" rather than simply "renewing" contracts, as clients often upgrade to the latest hardware to maintain their competitive edge.

CoreWeave's remaining performance obligations (RPO) reached 301 billion, marking an 86% year-over-year increase and a 40 billion increase from the first quarter. This growth is attributed to the company's strategic partnerships and the signing of two major contracts with large-scale cloud customers within the past eight weeks. One of these contracts has already been reflected in the second-quarter results, while the other is expected to impact the third-quarter data.

The company's financial performance for the second quarter was robust, with revenue exceeding 10 billion for the first time in a single quarter. CoreWeave also achieved an adjusted operating profit of 2 billion, marking a significant milestone. Despite the strong performance, the company's earnings call revealed that while revenue growth remained in the triple digits, it did not exceed market expectations as much as in the previous quarter. The company has revised its full-year revenue guidance upward due to strong AI demand but maintained its operating profit guidance.

CoreWeave's capital expenditure for the second quarter was 29 billion, reflecting the company's efforts to rapidly expand its infrastructure to meet client needs. The CEO noted that increasing capacity and services remains a key factor in the company's success in a supply-constrained market. The company has made significant progress in deploying next-generation AI infrastructure, including the integration of NVIDIA's GB200, NVL72, and HGXB200 into its CoreWeave Mission Control system.

In addition to its core services, CoreWeave has introduced several new products and services, including an innovative archival object storage product, full-stack observability features, and a reasoning service in collaboration with Weights and Biases. These offerings support advanced AI models from OpenAI, Meta's Llama4, and DeepSeek, among others.

CoreWeave's vertical integration strategy is also progressing, with the acquisition of Weights & Biases to enhance its technology stack and the planned acquisition of CoreScientific to strengthen its infrastructure control. The integration of CoreScientific is expected to add approximately 1.3 gigawatts of total power capacity and provide over 1 gigawatt of future expansion capability. This move is anticipated to eliminate over 100 billion in future lease debt and achieve 5 billion in annual cost savings by the end of 2027.

The company's strong financial performance and strategic initiatives have been supported by robust capital market activities. CoreWeave has successfully reduced its financing costs and expanded its funding sources, completing its first and second high-yield bond issuances within the past three months. The company also completed a landmark secure GPU financing transaction, collaborating with leading global banks to create a new financing structure. The recent 26 billion delayed draw term loan has a capital cost of SOFR+400, representing a 900 basis point reduction from the previous loan.

Since the beginning of 2024, CoreWeave has secured over 250 billion in debt and equity financing. As of June 30, the company holds 21 billion in cash, cash equivalents, and restricted cash, with no debt maturing before 2028, excluding supplier financing. These financial achievements underscore CoreWeave's ability to sustain its growth and innovation in the AI infrastructure market.