Startups > CoreWave

CoreWave

CoreWeave provides GPU cloud infrastructure optimized for AI workloads at 40-50% below AWS pricing.

Debt Financing $12.487B total Founded 2017 Roseland, New Jersey 60 employees

CoreWeave is an AI-native GPU cloud platform providing purpose-built infrastructure for AI workloads, including GPU compute, intelligent storage, high-performance networking, and managed services. The company serves AI developers and enterprises needing scalable compute resources, positioning itself as 35x faster and 80% cheaper than legacy cloud providers like AWS. Founded by former commodities traders, CoreWeave went public in March 2025 and has raised over $12.4B across equity and debt financing, making it one of the most heavily funded infrastructure startups.

Problem solved

Enterprises lack access to affordable, flexible GPU compute infrastructure that scales with AI workload demands without vendor lock-in or excessive costs.

Target customer

AI developers, machine learning teams, enterprises running large language models and complex AI workloads, CGI/rendering companies requiring GPU acceleration

Website LinkedIn Crunchbase Twitter / X

Founders

Michael Intrator

CEO & Chairman

Commodities and climate finance executive; founded Hudson Ridge Asset Management (natural gas hedge fund) and held senior portfolio management roles at Natsource Asset Management. MBA from Columbia University SIPA.

Brian Venturo

Chief Strategy Officer

Co-founder; former commodities trader.

Brannin McBee

Chief Development Officer

Co-founder; former commodities trader.

Peter Salanki

Chief Technology Officer

Co-founder; leads technical architecture and product development.

Funding history

Seed $1.2M February 2019 Led by Unknown · Unknown

Series A $1.1B May 2023 Led by Unknown · Unknown

Series B $642M December 2023 Led by Unknown · Unknown

Equity $420M April 2023 Led by Unknown · NVIDIA ($100M)

Debt Financing $2.3B August 2023 Led by Magnetar Capital, Blackstone · Unknown

Series C $1.1B May 2024 Led by Coatue Management · Unknown

Debt Facility $7.5B May 2024 Led by Blackstone · Magnetar, Coatue, Carlyle, CDPQ, DigitalBridge Credit, BlackRock, Eldridge Industries, Great Elm Capital

Secondary Share Sale $650M November 2024 Led by Unknown · Jane Street, Magnetar Capital, Fidelity Management, Macquarie Capital

IPO $1.5B March 28, 2025 Led by Public Markets · Unknown

Equity Investment $2B January 2026 Led by NVIDIA · Unknown

Convertible & Bond Offering $5.25B April 2026 Led by Unknown · Unknown

Delayed Draw Term Loan $8.5B April 2026 Led by Unknown · Unknown

Total raised: $12.487B

Industries

Cloud Infrastructure Cloud Computing Information Technology Machine Learning

Pricing

Usage-based, per-hour model starting at $10/instance/hour. H100 PCIe priced at $4.25/hour (GPU only); full 8-GPU nodes around $49.24/hour. Up to 60% discounts available for committed/reserved capacity. Enterprise custom pricing available. No data ingress/egress fees.

Notable customers

Not publicly disclosed; enterprise AI developers and ML teams

Integrations

Kubernetes-native platform; NVIDIA H100/H200 GPU partnerships; integrations with major cloud data platforms and AI frameworks

Tech stack

GSAP (JavaScript frameworks) jQuery (JavaScript libraries) core-js (JavaScript libraries) Open Graph HTTP/3 DocuSign HubSpot Analytics (Analytics) Matomo Analytics (Analytics) Linkedin Insight Tag (Analytics) Google Analytics (Analytics) Facebook Pixel (Analytics) Google Font API (Font scripts) Google Hosted Libraries (CDN) cdnjs (CDN) Cloudflare (CDN) HubSpot (Marketing automation) Google Tag Manager (Tag managers) Webflow (Page builders) Sendgrid (Email) Cloudflare Rocket Loader (Performance) Workable (Recruitment & staffing)

Website

coreweave.com/

Competitors

AWS

CoreWeave is 40-50% cheaper for GPU instances and purpose-built for AI workloads vs AWS's generalist approach.

Google Cloud Platform

CoreWeave offers specialized AI-native infrastructure with transparent pricing vs GCP's broader enterprise cloud services.

Microsoft Azure

CoreWeave focuses exclusively on GPU compute and AI infrastructure with no lock-in, while Azure bundles multiple services.

Lambda Labs

CoreWeave offers larger scale, Kubernetes-native environment, and institutional-grade debt financing vs Lambda's smaller-scale GPU rental service.

Crusoe Energy

CoreWeave is a generalist GPU cloud provider while Crusoe focuses on AI inference acceleration with custom hardware.

Why this matters: CoreWeave is the most heavily funded infrastructure startup globally at $12.4B+ raised, with an institutional-grade capital structure including $7.5B in debt from Blackstone and $2B from NVIDIA. The company's IPO and subsequent mega-rounds signal massive market demand for AI-specific GPU infrastructure, positioning it as a critical infrastructure layer for the AI economy.

Best for: AI teams and enterprises needing scalable, cost-effective GPU infrastructure for training, inference, and complex compute workloads without vendor lock-in.

Use cases

Large Language Model Training

Data science teams training large language models can leverage CoreWeave's Kubernetes-native H100 infrastructure at 40-50% below AWS pricing. The platform's high-performance networking enables efficient multi-node scaling across distributed training jobs, reducing training time from weeks to days.

AI Model Inference & Serving

Companies deploying AI models to production can use CoreWeave's flexible spot and reserved pricing to optimize costs. Customers pay only for compute used and gain the ability to scale inference workloads on-demand without upfront capital investment in hardware.

CGI & Rendering Workloads

Visual effects studios and CGI render farms can burst compute capacity using CoreWeave's GPU infrastructure during peak production periods. The flexible hourly pricing allows studios to avoid expensive on-premise GPU clusters while maintaining cost control.

AI Research & Experimentation

Researchers prototyping new ML architectures can rapidly iterate using CoreWeave's on-demand GPU access without procurement delays. The transparent pricing and no lock-in model encourages experimentation across multiple hardware configurations.

Alternatives

AWS EC2 AWS offers broader cloud services but CoreWeave is 40-50% cheaper for GPU workloads and purpose-built for AI.

Google Cloud AI Platform GCP integrates AI tools but charges more per GPU hour and locks customers into Google's ecosystem vs CoreWeave's flexibility.

Lambda Labs Lambda offers GPU rental but lacks CoreWeave's scale, institutional backing, Kubernetes integration, and enterprise-grade financing.

Crusoe Energy Crusoe specializes in inference acceleration hardware while CoreWeave is a general-purpose GPU cloud for both training and inference.

FAQ

What does CoreWeave do? +

CoreWeave provides AI-native GPU cloud infrastructure through a Kubernetes-native platform. It offers GPU compute, managed storage, high-performance networking, and software services designed specifically for AI workloads. The company serves AI developers and enterprises running large language models, machine learning training, and inference at 40-50% lower cost than AWS.

How much does CoreWeave cost? +

CoreWeave uses per-hour usage-based pricing starting at $10/instance/hour. H100 GPUs are priced at $4.25/hour (GPU only), with full 8-GPU nodes around $49.24/hour. Customers can receive up to 60% discounts with committed/reserved contracts. Enterprise pricing is available on request, and there are no data transfer fees.

What are alternatives to CoreWeave? +

Major alternatives include AWS EC2 (broader but more expensive), Google Cloud AI Platform (integrated but pricier), Lambda Labs (smaller-scale GPU rental), and Crusoe Energy (specialized inference hardware). CoreWeave differentiates through AI-specific design, lower pricing, and no vendor lock-in.

Who uses CoreWeave? +

CoreWeave serves AI developers and enterprises training/deploying large language models, research institutions, visual effects studios, and ML teams. Notable investors include NVIDIA, Blackstone, and Coatue. Specific end-customer names are not publicly disclosed.

How does CoreWeave compare to AWS for GPU workloads? +

CoreWeave is 35x faster and 80% cheaper than legacy providers like AWS for comparable GPU instances. CoreWeave is purpose-built for AI with Kubernetes-native architecture and intelligent storage, while AWS is a generalist cloud platform. CoreWeave also offers transparent pricing with no hidden data transfer fees, making it easier for teams to predict costs.

Is CoreWeave a public company? +

Yes, CoreWeave went public on Nasdaq under ticker CRWV on March 28, 2025, at $40/share. The IPO raised $1.5B and valued the company at approximately $23B. It has since received major follow-on investments including $2B from NVIDIA in January 2026.