CoreWeave contracted power capacity in fiscal 2026

CoreWeave plans substantial 2026 capital expenditure to boost power capacity. The company projects a significant capital expenditure between $31 billion and $35 billion for full-year 2026, primarily to scale infrastructure and fulfill existing customer contracts ^{[1][2][3][4][5]}. This investment aims to increase CoreWeave's active power capacity from approximately 1GW in Q1 2026 to over 1.7GW by the close of 2026, with deployments accelerating in the second half of the year ^[1][5][6].

This capacity growth is supported by an $8.5 billion loan secured in March 2026. This financing mechanism incorporates an ABS-style draw feature, allowing CoreWeave to access additional capital once underlying customer contracts have stabilized ^[1][7][8]. This funding strategy is specifically designed to support the planned infrastructure expansion and the increase in power capacity throughout 2026 ^{[1][2][3][4][5]}.

No major new client partnerships are indicated for H2 2026 impacting growth. Research findings do not show any new client partnerships emerging specifically in the second half of 2026 that would significantly alter CoreWeave's contracted power capacity growth. While CoreWeave's total contracted power capacity was over 3.86GW as of early August 2026, including a 360MW expansion in Indonesia and ongoing collaborations with partners like NVIDIA, these reflect existing capacity and continuous relationships rather than new partnerships established in H2 2026 ^[4][8][9].

CoreWeave projects significant data center expansion and capital investment in 2026. The company plans an aggressive buildout, aiming to double its active power capacity from 850MW at the end of 2025 to over 1.7GW by the close of 2026 ^[1][2][3][4]. This significant infrastructure buildout is supported by anticipated capital expenditures of $30–$35 billion for the full year 2026, marking a substantial increase from approximately $10.3 billion in 2025 ^[1][5][6]. CoreWeave also reported over 3.5GW of total contracted power capacity as of Q1 2026 and has a long-term objective of reaching 8GW of active power by 2030 ^[1][3][4].

Hyperscalers dominate the cloud market but direct comparison data for 2026 is absent. While CoreWeave demonstrates rapid growth in its specialized GPU-intensive AI cloud services, major hyperscalers such as AWS, Microsoft Azure, and Google Cloud collectively hold a dominant position, commanding over 60% of the global cloud infrastructure market as of Q1 2026 ^[7]. These established providers leverage extensive and diversified ecosystems, contrasting with CoreWeave's specialized focus ^[7]. However, the available research does not provide specific metrics on the rate of data center expansion, including power capacity growth or dedicated capital expenditures for data centers, for AWS, Microsoft Azure, or Google Cloud for the year 2026. This absence of information makes a direct quantitative comparison of their expansion rates with CoreWeave's challenging based on the available data.

CoreWeave's revenue backlog demonstrates a direct correlation with its contracted power capacity expansion. The company's new customer commitments are a primary driver for acquiring and deploying additional contracted power, which is then converted into active, revenue-generating capacity ^[2][5][3][4]. This relationship was clearly observed in 2025, as CoreWeave's revenue backlog grew to $66.8 billion, aligning with the contracting of approximately 2 GW of additional power. By the end of 2025, the company's total contracted power reached 3.1 GW, alongside 850 MW of active capacity ^[6][7][8].

Growth in backlog and capacity continued into early 2026. This trend extended into the first quarter of 2026, with the revenue backlog further increasing to $99.4 billion, reflecting over $40 billion in new commitments ^[1][2][3][4]. During this quarter, contracted power capacity expanded to 3.5 GW, and active power surpassed 1 GW ^[1][2][3][4]. CoreWeave is projected to achieve over 1.7 GW of active power by the end of 2026, supported by its revenue guidance of $12 billion to $13 billion for the year ^[5][3][4].

CoreWeave defines contracted power capacity for take-or-pay agreements. This metric represents the total power capacity secured under committed contracts, which customers utilize on a take-or-pay basis throughout the specified contract term ^[1]. As of December 31, 2025, CoreWeave reported a total of approximately 3.1 gigawatts (GW) in contracted power capacity ^[1][2].

CoreWeave anticipates deploying all contracted capacity by late 2027. This 3.1 GW of contracted capacity is scheduled for deployment across future periods, with virtually all of it expected to become operational by the end of 2027 ^[3][2]. Specifically for fiscal year 2026, CoreWeave's guidance projects a doubling of its active power capacity to surpass 1.7 gigawatts by the conclusion of that year ^[3][2].

CoreWeave's infrastructure build-out hinges on strategic supply agreements. CoreWeave's infrastructure build-out schedule is significantly influenced by its 2026 supply chain agreements with NVIDIA and Solidigm. These collaborations are crucial for securing access to essential hardware, such as GPUs and enterprise SSDs, and are designed to mitigate potential supply bottlenecks, directly synchronizing procurement for timely deployment ^{[1][2][7][8][9][6]}.

NVIDIA partnership ensures early access to advanced infrastructure. The deepened partnership with NVIDIA grants CoreWeave early access to various generations of NVIDIA's infrastructure, including Rubin, Vera CPUs, and BlueField storage systems. This access is vital in supporting CoreWeave's long-term ambition to develop over 5 gigawatts of AI factories by 2030 ^[1][2].

Solidigm agreement and power synchronization optimize deployment efficiency. Concurrently, a multi-year agreement with Solidigm, effective August 5, 2026, guarantees priority access to enterprise SSD capacity, specifically addressing storage as a critical constraint in scaling CoreWeave's integrated AI cloud platform ^[7][8][9]. CoreWeave's deployment timeline is primarily determined by its ability to synchronize the procurement and delivery of power, GPUs, and storage, with the company aiming to minimize the interval between powered shell readiness and full operational capacity. As of Q1 2026, CoreWeave has exceeded 1 gigawatt of active power and holds over 3.5 gigawatts of contracted power, with the majority expected online by the end of 2027. These strategic supply agreements are vital for achieving this synchronization and ensuring the timely deployment of its contracted capacity ^{[6][9][3][4][5]}.

Investment catalysts for CoreWeave center on the conversion of the $99.4 billion revenue backlog, successful execution of sequential margin expansion, and the upcoming Q2 2026 earnings release scheduled for August 11, 2026 ^[1][2]. The bull case rests on strong AI demand and backlog visibility ^[1][2]. Conversely, the bear case focuses on high leverage ($50.8 billion in total liabilities), GPU depreciation risks, and customer concentration ^[1][2][3].

Operationally, CoreWeave reported over 3.5 gigawatts of contracted power capacity as of Q1 2026, with more than 400 megawatts added during that quarter ^[4][5]. For fiscal 2026, the company reaffirmed revenue guidance of $12 billion to $13 billion and guided for full-year capital expenditures of $31 billion to $35 billion ^[6][7][8]. Year-end 2026 active power is targeted to exceed 1.7 gigawatts ^[6][7][8].