CRWV CoreWeave, Inc. 8-K 2026-09-17 slides

Safe Harbor Disclaimer

This presentation is strictly confidential and is being furnished solely in reliance on applicable exemptions from the registration requirements under the U.S. Securities Act of 1933 (the “Act”). This presentation contains selected information derived from the preliminary offering memorandum (the “offering memorandum”) of CoreWeave, Inc. (the "Company") and the portions of the documents that the Company has filed with the U.S. Securities and Exchange Commission (the “SEC”) that are incorporated by reference in the offering memorandum. The information contained in this presentation is provided as of the date of this presentation and is subject to change without notice.

The notes have not been and will not be registered under the Act, or any state securities laws or the laws of any foreign jurisdiction. The notes will be offered in the U.S. only to persons reasonably believed to be “qualified institutional buyers” pursuant to Rule 144A under the Act and outside the U.S. in reliance on Regulation S under the Act. Accordingly, this document is being provided only to persons that are reasonably believed to be “qualified institutional buyers,” as defined in Rule 144A under the Act and to persons that are non-U.S. persons (as defined in Regulation S under the Securities Act) in transactions outside the United States. By accepting this presentation, you will be deemed to represent that you are a qualified institutional buyer or a non-U.S. person outside the United States.

The notes have not been approved or disapproved by the SEC, or any other securities regulating body or agency, nor has any such authority, commission, or body passed on the accuracy or adequacy of this presentation. Any representation to the contrary is a criminal offense. The notes will be subject to restrictions on transferability and resale and may not be transferred or resold except as permitted under the Act and applicable state securities laws pursuant to registration or exemption therefrom. Investors should be aware that they may be required to bear the financial risks of an investment in the notes for an indefinite period of time. There is no obligation on the part of any person to make a market for the notes. No action has been made or will be taken that would permit a public offering of any securities in any jurisdiction in which action for that purpose is required. No offers, sales, resales or delivery of any securities or any distribution of any offering material relating to any securities may be made in or from any jurisdiction except in circumstances which will result in compliance with any applicable laws and regulations and which will not impose any obligation on the Company.

Before you make a decision to invest in the notes, you should read the offering memorandum and the portions of the documents that the Company has filed with the SEC that are incorporated by reference in the offering memorandum for more complete information about the Company and the offering to which this presentation relates. Any statements, whether oral or written, that are not consistent with the offering memorandum are not authorized and should not be relied upon by prospective investors. This presentation does not constitute an offer to sell or a solicitation of an offer to buy the notes. The offer to sell, or the solicitation of an offer to buy, the notes is made only pursuant to the offering memorandum.

This presentation and the accompanying oral commentary contain “forward-looking” statements based on the Company’s beliefs and assumptions and on information currently available to the Company. Forward-looking statements include all statements other than statements of historical fact contained in this presentation, including information or predictions concerning the future of the Company’s business or financial performance. The words “believe,” “may,” “will,” “potentially,” “estimate,” “continue,” “anticipate,” “intend,” “could,” “would,” “project,” “target,” “plan,” “expect,” and similar expressions are intended to identify forward-looking statements, although not all forward-looking statements contain these identifying words. These forward-looking statements are subject to a number of risks, uncertainties, and assumptions, including those described in the section titled “Risk Factors” in the Company’s Quarterly Report on Form 10-Q for the three months ended June 30, 2026 filed with the Securities and Exchange Commission (the “2026 Q2 10-Q”). Moreover, the Company operates in a very competitive and rapidly changing environment, and new risks emerge from time to time. The Company may not actually achieve the plans, intentions, or expectations disclosed in these forward-looking statements, and you should not place undue reliance on these forward-looking statements. Actual results or events could differ materially from the plans, intentions, and expectations disclosed in these forward-looking statements. In addition, the forward-looking statements included in this presentation represent the Company’s views as of the date of this presentation. The Company anticipates that subsequent events and developments will cause its views to change. However, while the Company may elect to update these forward-looking statements at some point in the future, it specifically disclaims any obligation to do so except as required by law. These forward-looking statements should not be relied upon as representing the Company’s views as of any date subsequent to the date of this presentation.

In addition, this presentation contains certain projected financial information with respect to the Company's future results. Such projected financial information was not prepared with a view to public disclosure or compliance with the guidelines established by the Public Company Accounting Oversight Board, published guidance or rules of the SEC or U.S. generally accepted accounting principles ("GAAP"). Such projected financial information constitutes forward-looking information, is presented for illustrative purposes only and should not be relied upon as being indicative of future results. The projected financial information is based on estimates and assumptions that the management of the Company believed to be reasonable at the time they were made and are inherently subject to significant uncertainties and contingencies, many of which are beyond the Company’s control, that could cause actual results to differ materially from the projected financial information. In particular, the value of terms of actual contracts, interest rates and financial results and measures may vary, and the figures included in this presentation are presented for illustrative purposes only and should not be considered to be an indication of the actual contract terms or values, or the Company’s financial position or results of operations for any historical or future periods. The projected financial information also does not reflect future changes in general business or economic conditions, or any other transaction or event that may occur and that was not anticipated at the time this information was prepared and is subject to risks, uncertainties, and other factors, including those described in the section titled “Risk Factors” in the 2026 Q2 10-Q. The inclusion of such information in this presentation should not be regarded as a representation by any person that the results reflected in such projections will be achieved. Neither the Company's independent auditors, nor any other independent accountants, have compiled, examined, or performed any procedures with respect to the prospective financial information contained herein, nor have they expressed any opinion or any other form of assurance on such information or its achievability, and assume no responsibility for, and disclaim any association with, the prospective financial information.

This presentation includes references to Revenue Backlog. Revenue Backlog includes remaining performance obligations, plus other amounts the Company estimates will be recognized as revenue in future periods under committed customer contracts, in each case, subject to the satisfaction of delivery and availability of service requirements. Revenue Backlog is expected to fluctuate from period to period, given the nature of the Company's committed contract business and the size of committed contracts.

This presentation includes certain financial measures not presented in accordance with GAAP, including Adjusted Operating Income (Loss), Adjusted Operating Income (Loss) margin, Adjusted EBITDA, Adjusted EBITDA margin, Adjusted Net Loss and Adjusted Net Loss margin. Such non-GAAP measures are used by management in making operating decisions, allocating financial resources, and for internal planning and forecasting and business strategy purposes. The non-GAAP measures included in this presentation have certain limitations, and should not be construed as alternatives to financial measures determined in accordance with GAAP. The non-GAAP measures as defined by us may not be comparable to similar non-GAAP measures presented by other companies. The Company's presentation of such measures, which may include adjustments to exclude unusual or non-recurring items, should not be construed as an inference that the Company's future results will be unaffected by other unusual or non-recurring items. Please refer to the documents filed and furnished by the Company with the SEC for more information with respect to the Company's use of non-GAAP financial measures, including reconciliations thereof. Please refer to the Appendix for a reconciliation of each non-GAAP financial measure presented herein to the most directly comparable financial measure stated in accordance with GAAP.

This presentation also contains estimates and other statistical data made by independent parties and by the Company relating to market size and growth and other data about the Company’s industry and estimated total addressable market. This data involves a number of assumptions and limitations, and you are cautioned not to give undue weight to such estimates. Although we are responsible for all of the disclosures contained in this presentation and we believe the third-party market position, market opportunity, and market size data included in this presentation are reliable, we have not independently verified the accuracy or completeness of this third-party data. In addition, projections, assumptions, and estimates of the Company’s future performance and the future performance of the markets in which the Company operates are necessarily subject to a high degree of uncertainty and risk.

This presentation shall not constitute an offer to sell or the solicitation of an offer to buy securities, nor shall there be any sale of securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction. Any decision to purchase securities of the Company should be made solely on the basis of the information contained in a prospectus to be issued by the Company in relation to a specific offering.

All third-party trademarks, including names, logos, and brands, referenced by the Company in this presentation are property of their respective owners. All references to third-party trademarks are for identification purposes only and shall be considered nominative fair use under trademark law.

Offering Summary

Executing at Scale: Recent Momentum Across the Business

1. Exceptional Q2 execution: accelerating revenue and margin growth, record revenue backlog, marquee customer wins, and a strengthened financial position
2. Pricing and unit economics continue to expand, with both price and margins increasing across SKUs and contract tenors
3. The essential cloud enabling both proprietary AI labs and the open-weight ecosystem across training, inference, and managed inference
4. Enterprise adoption is intensifying across verticals, from industrial and physical AI to financial services and cybersecurity
5. Significant demand for prior generations of cloud infrastructure, driving strong pricing and contract terms as AI diffuses globally
6. A cost-optimized, systematic financing approach that funds the majority of capex at the asset level

Continued Momentum with Exceptional Execution in Q2

Revenue Backlog note: Does not include more than $25 billion of net new customer commitments added in early Q3.

Scaling Our Capacity

Strengthened Financial Position

An AI Cloud Built to Scale

1. First AI cloud provider to bring up and validate NVIDIA's Vera Rubin NVL72
2. Delivered seven new AI platform capabilities to accelerate AI development and deployment
3. Posted record MLPerf® benchmarks across training and inference
4. Launched unified agentic AI capabilities, including Sandboxes and ARIA to fuel the AI loop
5. Unlocked faster AI deployments for enterprises through cross-cloud services

Select New Customers

Select Customer Expansions

Note:

1. Non-GAAP measure; see Appendix for GAAP reconciliation.
2. As of Q2 2026 earnings call on August 11, 2026.

Unit Economics Are Inflecting with Highly Attractive Contract Terms

• +5–10 pts — higher expected contribution margin on Q2-signed contracts vs. recent quarters¹
• ~25% — price increase across SKUs in July 2026
• ~$40M/MW — recent pricing range for signed short-dated contracts (3–6 months) in Q3²
• ~70% — of deals signed in Q2 included customer prepayment

Note:

1. Reflects contribution margin across all Q2-signed contracts.
2. Calculated as annualized revenue divided by power required to service clusters for related short-term contracts.

CoreWeave Powers Both the Proprietary and Open AI Ecosystems

The essential cloud enabling the world's leading proprietary AI labs and the open-weight ecosystem across training, inference, and managed inference

Proprietary / Closed-Source AI

World's leading frontier labs and AI-native companies: Anthropic, Cursor, Google, Hippocratic AI, Isomorphic Labs, Microsoft, Midjourney, OpenAI, Periodic Labs, Runway, World Labs

How leading labs and AI-native companies use CoreWeave

• Frontier-scale training clusters for the world's leading closed-model labs, deployed first-to-market on each new NVIDIA generation
• Low-latency inference powering consumer and enterprise AI applications at massive scale
• Deep, longstanding partnerships spanning multiple products and long-term take-or-pay commitments

Open / Open-Weight Ecosystem

Customers / Strategic Relationships: Cognition, Cohere, Databricks, IBM, Meta, Mistral, NVIDIA

Leading Open Models Served: DeepSeek, Gemma, Granite, Kimi, MiniMax, Mistral AI, NVIDIA Nemotron, Poolside, Qwen, GLM-5.3

How CoreWeave serves the open ecosystem

• Training and fine-tuning infrastructure for both open-model developers and enterprises
• CoreWeave Managed Inference serves the leading open models at scale, with booked ARR ramping from $1 million to more than $100 million in only a few months
• Consistent leadership in Artificial Analysis benchmarks

Enterprises Are Developing New Vertical Use Cases on CoreWeave

Established enterprises across industries are moving into production on CoreWeave.

Industrial / Physical AI

Select Use Case: physical AI, robotics, and simulation
Training and simulating the world models, foundation models, and generative-design systems behind physical AI, from autonomous construction equipment and vehicles to humanoid robots and engineering simulation.

Recent customer win
Caterpillar deploys NVIDIA Vera Rubin with CoreWeave as its data factory to train specialized physical-AI models at industrial scale.

Financial Services

Select Use Case: quantitative trading and research
Migrating high-intensity foundation-model training and AI research onto CoreWeave, deployed for quantitative trading strategies, market making, and systematic research at scale.

Recent customer wins
~10% of revenue backlog driven by financial-services customers¹

Cybersecurity

Select Use Case: AI-native security
Running large-scale AI training and low-latency inference to power real-time threat detection, autonomous response, and network protection across billions of events each day.

Recent customer win
CoreWeave selected to power training and inference for CrowdStrike’s new agentic AI cybersecurity system SafeMind.

Note:
¹ As of September 13, 2026.

Significant Demand for Prior Generations of Infrastructure on CoreWeave Cloud

• Greater level of experience and sophistication with prior SKUs
• Workloads already mapped to existing architectures allow for continuity of research, product development, and deployment
• Switching costs keep mission-critical workloads on proven infrastructure
• ROI proving to be highly attractive, driving pricing stability and expansion

CoreWeave’s Cost Optimized Financing Approach

Asset-Level Financing
(Funds Majority of Capital Expenditures)

• Secured by long-term take-or-pay customer contracts, plus the related GPUs, PP&E and data center leases
• Attractive pricing given seniority of instrument and look through to credit-worthy customers
• Utilized only when costs are incurred and infrastructure is deployed via delayed draw structure
• Typically amortizes as customers make contractual payments, such that the debt is fully repaid by the time the customer contract expires
• Excess cash flow is distributed as dividends to CoreWeave, Inc. during the life of the customer contract
• Investment-grade ratings received when the underlying customers are investment grade (e.g., DDTL 4.0, which achieved an “A-” equivalent rating from three rating agencies)

Parent-Level Financing
(Funds Remainder of Capital Needs)

• Funds capital needs beyond the asset-level facilities
• Sources:
  • Customer prepayments
  • Free cash flow from stabilized contracts
  • CoreWeave, Inc. financing (debt, OEM financing, convertibles, equity)

CoreWeave – The Essential Cloud for AI

• Cloud Platform Purpose-Built for Artificial Intelligence Delivering Unmatched Performance¹
• Rapidly Delivering New Generations of Infrastructure at Scale with Unparalleled Track Record of Being Among the First to Market
• AI-Native Platform Offers Unmatched Product Market Depth across Managed Inference, Development Tools, Orchestration, and Observability
• Serving Most of the World’s Leading AI Labs, Hyperscalers, and Enterprises
• Systematic Approach to Financing at Scale
• Unique Combination of Growth at Scale with Attractive Unit Economics

Note:

1. Based on MLPerf benchmark results, NVIDIA Exemplar Cloud status, and SemiAnalysis ClusterMAX rating

Workloads Required for AI Are Fundamentally Different

Legacy Cloud Compute Platforms Were Built for a Different Era

Serialized Workloads

• Website Hosting
• Databases
• Email Systems
• SaaS Applications
• If… Single Chip Failure: Simple Tasks Offloaded to Other Processors
• Then

Parallelized Workloads

• Generative and Agentic AI
• VFX and Rendering
• Drug Discovery
• High-Frequency Trading
• If… Single Chip Failure: Entire Job Fails, Massive Waste of Compute Resources
• Then

CoreWeave’s AI Cloud Platform Is Singular

Only Player to Achieve Top-Ranked Platinum Status in Each of SemiAnalysis’ GPU Cloud ClusterMAX™ Ratings

Below Silver Status: ~200 Other Cloud Platforms

Recent Expansion of Collaboration with NVIDIA Follows Track Record of Delivering Bleeding-Edge Compute at Scale

Established Track Record of Being First-to-Market for Latest Generations of NVIDIA Accelerated Computing Technology

Continued Expansion of Strategic Relationship With NVIDIA to Deliver AI

Distribution of CoreWeave’s Cloud Stack

• Test and validate CoreWeave’s AI-native cloud stack to work towards including within NVIDIA’s reference architectures for NVIDIA cloud, enterprise, and sovereign customers

Accelerated Buildout of 5GW of Cloud Infrastructure

• Build and deliver more than 5GW of cloud infrastructure to meet customer demand
• Selectively leverage NVIDIA’s financial strength to accelerate CoreWeave’s procurement of land, power and data center shells

Further Platform Alignment

• $2BN investment in support of CoreWeave’s platform, team, and shared vision for AI infrastructure at scale

Large and Growing Footprint of AI Data Centers

• Technologies to Maximize Rack Density
• Systematized Processes and Modular Deployments
• Embedded Security
• Broad Geographical Footprint Minimizing End User Latency
• Cutting-Edge Liquid Cooling Technology
• Massive Scale
• High-Speed Interconnects

Map legend: Data Center; Region and/or Point of Presence¹; Backbone²

Callout: Our Extensive Data Center Network Enables Artificial Intelligence Across Use Cases and Geographies

Note: Figures as of June 30, 2026 (other than Contracted Power which is as of August 11, 2026). This graphic provides an illustrative representation of our infrastructure footprint and may not precisely reflect all current locations or capacities.

1. Region represents a local grouping of data centers where customers can deploy services. Point of Presence represents a network entry and exit point.
2. Represents the core, high-capacity network infrastructure that interconnects data centers and carries primary inter-data-center traffic.

CoreWeave Has Built Expert Know-How from Delivering AI Cloud Services Across 51 High Performance Data Center Sites

Diversified Set of Data Center Partners

Illustrative Timeline for AI Cloud Services Delivery¹

Callout: Exceptional time to market drives competitive advantage for customers

Note: 1. Timelines presented are illustrative and actual delivery dates and timelines are subject to a number of different factors, many of which may be outside of our control, as such, such illustrative timelines may not be indicative of actual timelines.

Our Platform Is Purpose-Built for AI

Every layer is purpose-built for AI workloads. CoreWeave Cloud is optimized for low latency, high throughput, and operational efficiency to support the complexity of large-scale AI training and inference.

Platform layers

• Application Software Services
• Managed Software Services
• Infrastructure Software Services
• Data Centers

Model and Agent Development
Tools for teams to build, evaluate, deploy, and monitor models and agents—speeding time to production.
Key Product: W&B Models

Runtime Acceleration
AI-native software that accelerates training and inference by reducing startup latency, improving throughput, and increasing utilization at runtime.
Key Product: Slurm on Kubernetes (SUNK)

Infrastructure Control
Integrated, AI-native orchestration and bare-metal control that deliver reliability, flexibility, and efficiency for complex workloads.
Key Product: CoreWeave Kubernetes Service (CKS)

Data and Storage
Purpose-built storage services combining exascale, AI-optimized object and file storage with GPU-local caching to deliver high-throughput data access, cross-cloud reach, and predictable economics for training and inference workloads.
Key Product: CoreWeave AI Object Storage

Foundational Infrastructure
Purpose-built data centers and infrastructure that maximize performance with first-to-market GPU clusters, ultra-high density, and high-speed interconnects that enable AI breakthroughs and lower TCO.

Diagram labels: Specialized Workloads; Training; Inference

CoreWeave Mission Control™
Security, Talent Services, Observability

CoreWeave Mission Control™ integrates security, observability, and talent services—including node, rack, and fleet lifecycle management—to enable intelligent, unified orchestration from foundational infrastructure to agent development.

CoreWeave’s Cloud Infrastructure Flywheel: Continuous Optimization at Scale

Chart: Massive Troves of Data Unlock Deep Insights into Infrastructure Management

Horizontal-axis labels: 10:00, 12:00, 14:00, 16:00, 18:00
Vertical-axis labels: 330.0 TFLOPS to 364.0 TFLOPS, in 2.0 TFLOPS increments.

Data-Driven Advantages Driving Differentiation

1. Customer-led approach to software innovation
2. Develop new products (e.g. AI Object Storage) to solve key customer pain points
3. IaaS and cloud software platform optimized for AI
4. Software sold as standalone license or bundled with infrastructure as unified cloud platform

CoreWeave Serves the World’s Leading AI Companies

• Anthropic: Anthropic partnership reinforces CoreWeave’s role as trusted infrastructure partner to leading AI labs
• Cohere: CoreWeave’s new Canadian data center allows Cohere to train some of its LLMs in Canada, previously not an option due to the lack of necessary compute
• Google Cloud: High performance compute partner to Google Cloud as of Q2’25, has already signed multiple contracts
• IBM: CoreWeave delivers supercomputers to IBM, and partners to deliver IBM Storage Scale system to CoreWeave customers
• Meta: Expanding partnership with Meta to power next generation workloads
• Microsoft: Longstanding customer relationship with several contracts signed, demonstrating commitment to CoreWeave
• Mistral AI: Partnership with CoreWeave offloads burden of managing infrastructure and other operational tasks
• NVIDIA: CoreWeave serves as scaled AI Hyperscaler used to pressure-test next-gen NVIDIA chips; NVIDIA is also a CoreWeave customer
• OpenAI: CoreWeave’s relationship with OpenAI represents ability to deliver reliable and performant infrastructure for some of the most demanding, bleeding-edge AI Labs

Track Record of Expansion with Our Customers

Land + Expand: Customer A

• Land with test clusters
• Expand with new contracts for more compute capacity to serve increasing workload demand

Incremental TCV

Callout: 392× increase in cumulative TCV less than 36 months after first contract

Advance with New Compute: Customer B

• Consistent purchase behavior of upgrading to later generations of chips while expanding capacity with older generations

Hardware Generation Expansion

Callouts: + New Generations (three callouts); 4 successive generations of compute purchases

Note: Based on contract close dates

Holistic Distribution Approach to the World’s Leading Enterprises

CoreWeave is a central enabler for Hyperscalers, LLMs and Direct Cloud Consumption

• Multiple ways for organizations to consume compute: CoreWeave distributes via each of them
• The CoreWeave Cloud Platform simplifies the complexity of engineering, assembling, running, and monitoring state-of-the-art infrastructure at a massive scale to deliver cutting-edge performance and efficiency for AI workloads

Distribution channels:

• Through Hyperscalers
• Through Direct Cloud Consumption
• Through LLMs

Team with Track Record of Execution

Operating Team and Board of Directors

Co-Founders

• Michael Intrator — CEO, Co-Founder, and Chairman of the Board
• Brian Venturo — CSO, Co-Founder, and Board Member
• Brannin McBee — CDO and Co-Founder
• Peter Salanki — CTO and Co-Founder

Senior Leadership

• Nitin Agrawal — CFO
• Jon Jones — CRO
• Jim Higgins — CISO
• Chen Goldberg — EVP Product & Engineering
• Jean English — CMO
• Jeff Baker — CAO
• Sachin Jain — COO
• Ernie Rogers — Chief Architect of Strategic Financing
• Sandy Venugopal — CIO

Independent Board of Directors

• Karen Boone — Board Member
• Meg Whitman — Board Member
• Glenn Hutchins — Board Member

Q2’26 Highlights

• $2.6B — Revenue; up 112% YoY
• $104.2B — Revenue Backlog¹˒²; up 246% YoY
• $9.4B — Capital Expenditures³
• $1.5B — Adjusted EBITDA⁴; 59% Margin
• $128M — Adjusted Operating Income⁴; 5% Margin
• $(567)M — Adjusted Net Loss⁴; (22)% Margin

Note:

1. Does not include more than $25 billion of net new customer commitments added in early Q3.
2. See Slide 35 for definition of Revenue Backlog.
3. Capital expenditures are additions to property and equipment plus assets acquired under finance leases, less changes to construction in progress.
4. Adjusted EBITDA, Adjusted EBITDA Margin, Adjusted Operating Income, Adjusted Operating Income Margin, Adjusted Net Loss and Adjusted Net Loss Margin are non-GAAP financial measures. Please see the reconciliation from GAAP to non-GAAP measures contained in the Appendix.

Continued Progress in Driving Down Cost of Debt

Weighted Average Interest Rate (%)¹,²

Right arrows appear between the chart bars.

Callout: Improvement from Q2-25 to Q2-26 represents nearly $1.1B of annualized interest savings based on Q2-26 debt balance.

Notes:

1. Weighted average interest rates are calculated using effective interest rates and principal amounts of debt for each period as shown in our SEC filings, weighted by outstanding debt at the end of each period.
2. We have estimated 10% effective interest rate for OEM and software license financing arrangements for each period shown.
3. FY-23 weighted average interest rate is calculated using effective interest rates as of FY-24 and principal amounts of debt as of FY-23.
4. Q2-26 pro forma debt illustratively shown as Q2-26 debt + $3.5B Convertible Note issuance at 2.625% interest, does not include DDTL 5.5.
5. Revenue backlog includes remaining performance obligations, plus other amounts we estimate will be recognized as revenue in future periods under committed customer contracts, in each case, subject to the satisfaction of delivery and availability of service requirements. Revenue Backlog as of December 31, 2023, 2024, and 2025, and June 30, 2026 were $9.9B, $15.1B, $66.8B and $104.2B, respectively.

Revenue Backlog¹

Revenue Backlog: $104.2 billion, up 246% YoY, driven by continued diversification from enterprise and AI-native customers.

Chart: Revenue Backlog¹
Note: $ in billions

up arrow: +246% YoY Growth

Callout: Does not include more than $25 billion of net new customer commitments added in early Q3.

Notes:

1. Revenue backlog includes remaining performance obligations, plus other amounts we estimate will be recognized as revenue in future periods under committed customer contracts, in each case, subject to the satisfaction of delivery and availability of service requirements. Revenue backlog is expected to fluctuate from period to period, given the nature of our committed contract business and the size of committed contracts. The percentages shown in the bar charts represent the percentage of revenue backlog we expect to recognize in the months following the end of the quarterly reporting period.
2. Q2'25 figures represent unsatisfied revenue backlog as of June 30, 2025. Q2'26 figures represent revenue backlog as of June 30, 2026.

Illustrative Impact from Ramping Multiple New Deployments Across Periods

Ramping Multiple Deployments Simultaneously Compresses Contribution Margins in the Near-term, Prior to Stabilizing in the Mid 20%s Once Fully Delivered

Chart — Aggregate Contribution Margin %

Assumption: One new deployment of equal size begins to ramp in each period.

Callouts

• Initial stabilized margin before new deployments
• Deployment 1 begins with active power increasing ~30% period-over-period
• DC operationalization and GPU installation commences
• Margin troughs as Deployment 2 commences while Deployment 1 not yet stabilized
• Margins recover meaningfully with active power base 3x+ the size of the initial power base
• However, significant ongoing deployments remain
• Aggregate contribution margin profile at stabilization

Legend

• Gray: Existing Stabilized Revenue
• Blue shades: New Deployment 1–8 Revenue
• Gray: Run-Rate Stabilized Revenue
• Gray line: Aggregate Contribution Margin %

Note:

1. Margin impact in the illustrative periods may vary based on size and timing of deployments ramping

Quarterly Financial Trends

Revenue

Driven by continued strong execution and customer demand for CoreWeave's AI cloud platform

Q2'26: +112% YoY Growth

Adjusted EBITDA¹

Q2'26: 59 % Margin

Adjusted Operating Income¹

Q2'26: 5% Margin

Capital Expenditures ($bn)²

Driven by scaling our platform to meet customer demand

Note: $ in millions unless otherwise noted

1. Adjusted EBITDA, Adjusted EBITDA Margin, Adjusted Operating Income, and Adjusted Operating Income Margin are non - GAAP financial measures. Please see the reconciliation from GAAP to non - GAAP measures contained in the Appendix
2. Capital expenditures are additions to property and equipment including assets acquired under finance leases, less changes in con struction in progress

Reconciliation from GAAP to Non-GAAP Measures

(In millions, except percentages)

Adjusted Operating Income (Loss) and Adjusted Operating Income (Loss) Margin

The following table presents a reconciliation of operating income (loss) and operating income (loss) margin, the most directly comparable financial measures stated in accordance with GAAP, to adjusted operating income and adjusted operating income margin, respectively, for each of the periods presented:

Note: Figures may not sum due to rounding.

1. Acquisition related costs include direct transaction costs, such as due diligence, advisory, and professional services fees, and certain compensation and integration related expenses. We exclude acquisition related costs, as we believe these transaction-specific expenses are inconsistent in amount and frequency, and do not correlate to the operation of our business.
2. In the second quarter of 2025, we began including an adjustment for the amortization of acquired intangibles in our calculation of adjusted operating income (loss). Prior period non-GAAP calculations for acquired intangible amortization are not being adjusted as these amounts were insignificant.

Reconciliation from GAAP to Non-GAAP Measures

(in millions, except percentages)

Adjusted Net Loss and Adjusted Net Loss Margin

The following table presents a reconciliation of net loss and net loss margin, the most directly comparable financial measures stated in accordance with GAAP, to adjusted net loss and adjusted net loss margin, respectively, for each of the periods presented:

Note: Figures may not sum due to rounding.

1. Primarily relates to losses recognized upon the early extinguishment of certain OEM financing arrangements, as well as accelerated amortization of debt discount and debt issuance costs related to our term loan facility governed by a previously terminated credit agreement, which was repaid in connection with the initial public offering.
2. Acquisition related costs include direct transaction costs, such as due diligence, advisory, and professional services fees, and certain compensation and integration related expenses. We exclude acquisition related costs, as we believe these transaction-specific expenses are inconsistent in amount and frequency, and do not correlate to the operation of our business.
3. In the second quarter of 2025, we began including an adjustment for the amortization of acquired intangibles in our calculation of adjusted net loss. Prior period non-GAAP calculations for acquired intangible amortization are not being adjusted as these amounts were insignificant.
4. Primarily relates to a net unrealized gain on our strategic investments.
5. In the second quarter of 2025, we began including an adjustment for the income tax effect related to our non-GAAP adjustments. Prior period non-GAAP calculations for the income tax effects on our non-GAAP adjustments are not being adjusted as these amounts were not material. Additionally, the third quarter of 2025 includes an adjustment for amounts related to the impact of the passage of the One Big Beautiful Bill Act on the first and second quarters of 2025, that were recorded in third quarter of 2025.

Reconciliation from GAAP to Non-GAAP Measures

Adjusted EBITDA and Adjusted EBITDA Margin

The following table presents a reconciliation of net loss and net loss margin, the most directly comparable financial measures stated in accordance with GAAP, to adjusted EBITDA and adjusted EBITDA margin, respectively, for each of the periods presented:

(in millions, except percentages)

Note: Figures may not sum due to rounding.

¹ Acquisition related costs include direct transaction costs, such as due diligence, advisory, and professional services fees, and certain compensation and integration related expenses. We exclude acquisition related costs, as we believe these transaction-specific expenses are inconsistent in amount and frequency, and do not correlate to the operation of our business.

Revenue Backlog

(in billions)

As of June 30,

Note:

1. Does not include more than $25 billion of net new customer commitments added in early Q3.
2. Revenue backlog includes remaining performance obligations, plus other amounts we estimate will be recognized as revenue in future periods under committed customer contracts, in each case, subject to the satisfaction of delivery and availability of service requirements.

Capital Expenditures Calculation

(in millions)

Note:

1. Capital expenditures are additions to property and equipment including assets acquired under finance leases, less changes in construction in progress.