In today’s world where computing power is the new gold, CoreWeave has emerged as a dominant force in high-performance AI infrastructure. Backed by strategic support from NVIDIA, the company has rapidly transformed from a small-time Ethereum mining operation into a multi-billion-dollar cloud computing provider serving giants like Microsoft. With a valuation surpassing **$2 billion** and a groundbreaking $2.3 billion debt financing secured using NVIDIA H100 GPUs as collateral, CoreWeave’s journey is nothing short of extraordinary.
This is the story of how a group of former commodities traders leveraged market downturns, GPU scarcity, and the AI explosion to build one of the most sought-after AI infrastructure companies in the world.
The Humble Beginnings: Mining on a Pool Table
In 2016, what began as a hobby would soon evolve into a tech powerhouse. Michael Intrator, Brannin McBee, and Brian Venturo—three former commodities traders—purchased their first GPU out of curiosity and set it up on a pool table overlooking the East River in Lower Manhattan. They used it to mine on the Ethereum network, successfully mining their first block.
What started as a weekend experiment quickly turned into a full-blown business. With early backing from friends and personal networks, the team moved operations from that pool table to a garage in New Jersey—essentially their first makeshift data center. Their company, initially named Atlantic Crypto, was born.
As cryptocurrency miners, their most valuable asset was clear: GPUs. And when the crypto winter hit in 2018–2019, many competitors folded. CoreWeave saw opportunity. While others sold off hardware at rock-bottom prices, they bought aggressively—scaling from hundreds to tens of thousands of GPUs across seven data centers.
👉 Discover how GPU-powered infrastructure is reshaping the future of AI and cloud computing.
Pivoting to Cloud: Solving Real-World Compute Demand
During this expansion, CoreWeave began offering GPU cloud services to other miners. But soon, they noticed something unexpected: developers outside the crypto space were reaching out, seeking powerful GPU acceleration for tasks like CG rendering, scientific simulations, and machine learning.
These clients shared a common frustration: traditional cloud providers like AWS and Google Cloud offered limited GPU options, often at inflated prices—making large-scale projects prohibitively expensive.
Recognizing this gap, CoreWeave officially pivoted in 2019. They repositioned themselves as a GPU-accelerated cloud service provider, catering to enterprises that needed raw, scalable computing power without the constraints of mainstream clouds.
According to CoreWeave’s official claims, their infrastructure delivers 35 times faster performance and costs up to 80% less than general-purpose public clouds—thanks to optimized hardware, networking, and storage solutions tailored for GPU-intensive workloads.
By 2021, the company had raised $50 million in funding, solidifying its shift toward enterprise cloud services. Interestingly, they still used idle customer compute capacity to mine Ethereum—a clever way to monetize underutilized resources during the transition phase.
Riding the AI Wave: The 2023 Breakthrough
Everything changed in 2023.
The global launch of ChatGPT triggered an unprecedented surge in demand for AI training infrastructure. Suddenly, every tech company wanted access to thousands of high-end GPUs to train large language models (LLMs). And CoreWeave was perfectly positioned to capitalize.
In April 2023, the company secured $221 million in Series B funding**, followed by an additional **$200 million extension in May—bringing total funding to $576.5 million. More importantly, NVIDIA joined the round, signaling a major strategic endorsement.
Why NVIDIA Bet on CoreWeave
NVIDIA’s involvement wasn’t just financial. Despite extreme shortages of H100 chips, NVIDIA prioritized CoreWeave over even Amazon and Google Cloud for hardware allocations—a rare and telling move.
Why? Because CoreWeave doesn’t compete with NVIDIA; it amplifies its ecosystem. Unlike tech giants investing in custom silicon (like Google’s TPUs), CoreWeave runs entirely on NVIDIA hardware and software stacks. They’re a pure-play partner—making them an ideal ally for expanding NVIDIA’s reach in AI cloud infrastructure.
In September 2023, CoreWeave became NVIDIA’s first Elite-tier Cloud Service Provider (CSP), gaining access to advanced technologies like:
- NVIDIA InfiniBand: Delivers ultra-low latency, high-bandwidth networking critical for AI training clusters.
- BlueField DPUs: Offload infrastructure tasks from CPUs, improving efficiency and scalability in data centers.
Setting Records in AI Training Performance
The partnership bore immediate fruit.
In July 2023, CoreWeave demonstrated a stunning feat: training a GPT-3-scale model in just 11 minutes using H100-based cloud infrastructure. For context, OpenAI reportedly took 14.8 days using 10,000 A100 GPUs.
This leap in efficiency wasn’t magic—it was architecture. CoreWeave optimized cluster design, interconnectivity, and software stack integration to maximize throughput and minimize bottlenecks.
Such performance attracted major clients. In June 2024, Microsoft signed a multi-billion-dollar agreement with CoreWeave to ensure sufficient AI compute capacity for future versions of ChatGPT and Azure AI services—a testament to CoreWeave’s reliability and scalability.
Financial Trajectory and Future Vision
The numbers speak volumes:
- 2023 revenue: $30 million
- 2024 projected revenue: $500 million
- 2025 contracted revenue: Nearly $2 billion
These figures reflect not just current demand but long-term enterprise commitments. CoreWeave isn’t just riding a trend—it’s becoming foundational infrastructure for the AI era.
Beyond generative AI, the company is expanding into life sciences, including:
- Drug discovery
- Protein folding simulations
- Genomic analysis
- Molecular modeling
During the 2020 pandemic, CoreWeave donated computing power from 6,000 GPU rigs to Stanford University’s Folding@Home project, helping researchers analyze coronavirus proteins and potential treatments—an early sign of their commitment to impactful science.
Brannin McBee, Co-Founder and Chief Strategy Officer, reflected:
“Walking into our data centers is one of the most awe-inspiring experiences. I’ve been building computers since I was a kid. Now, we’re scaling this up to power some of the coolest companies on the planet with raw computing muscle. It’s surreal.”
Frequently Asked Questions (FAQ)
Q: What made CoreWeave different from traditional cloud providers?
A: Unlike general-purpose clouds, CoreWeave specializes exclusively in GPU-accelerated computing. Their infrastructure is optimized end-to-end for AI, rendering, and scientific workloads—offering significantly higher performance and lower costs for GPU-intensive tasks.
Q: How did CoreWeave secure so many NVIDIA H100s amid global shortages?
A: Their status as a non-competitive partner to NVIDIA—relying entirely on NVIDIA hardware without developing alternatives—earned them priority access. Being named an Elite CSP further solidified trust and supply chain advantages.
Q: Is CoreWeave still involved in cryptocurrency mining?
A: No. While they used idle capacity for Ethereum mining in earlier years, the company has fully transitioned to enterprise cloud services since 2023, focusing on AI and high-performance computing.
Q: Who are CoreWeave’s main customers?
A: Key clients include Microsoft, life science research institutions, AI startups, and enterprises running large-scale machine learning models. Their client base spans tech, healthcare, entertainment (e.g., CGI studios), and academia.
Q: Can other companies replicate CoreWeave’s model?
A: It’s extremely difficult today. Access to NVIDIA’s top-tier GPUs is limited and tightly controlled. CoreWeave’s early mover advantage, deep partnerships, and proven scalability create significant barriers to entry.
👉 Learn how next-generation cloud providers are redefining what’s possible in artificial intelligence.
Final Thoughts
CoreWeave’s journey—from mining Ethereum on a pool table to powering billion-dollar AI systems—is a modern tech fairy tale rooted in timing, vision, and strategic execution. By identifying market gaps during crypto winters and pivoting toward scalable GPU cloud services, they positioned themselves at the epicenter of the AI revolution.
Backed by NVIDIA, trusted by Microsoft, and delivering unmatched performance, CoreWeave is no longer just another cloud provider—it’s a critical enabler of the next wave of technological innovation.
As demand for AI compute continues to explode, companies like CoreWeave won’t just thrive—they’ll shape the future.