Mike Mattacola, CBO of CoreWeave: Building the Next Generation of Cloud in Europe

Acquisition of Core Scientific:

• Announced a planned all-stock acquisition of Core Scientific, targeting completion within six months.
• The deal adds over 1.3 gigawatts of available power in the U.S., including 850 megawatts already contracted and partially online.
• The transaction aims to eliminate intermediary margins, reduce costs, and accelerate capacity deployment for customers.

Financial Performance and Strategy:

• Revenue grew from $15.8 million in 2022 to nearly $2 billion in 2024.
• Growth was driven by selective customer partnerships rather than broad market adoption, focusing on large training clusters and rapid execution.
• The company pioneered the launch of NVIDIA H100, H200, GB200, and recently announced the GB300 to the market.
• Capital was raised through financial innovation, collateralizing credit-worthy contracts and infrastructure assets before a recent successful IPO.

European Expansion and Infrastructure:

• Originally planned to open 3–4 data centers in Europe; expanded to 6 within 12 months, exceeding initial commitments.
• European investment now totals €1 billion in the UK and $2.2 billion in the rest of Europe.
• A seventh data center is scheduled for announcement within 90 days.
• Regional team grew from zero to approximately 120 employees in the last year.
• Current infrastructure includes Blackwell chips; the roadmap includes preparing for NVIDIA Rubin next-generation hardware.

Strategic Outlook and Market Dynamics:

• Sovereignty View: Michael argues against 100% sovereign infrastructure in Europe, citing a lack of local capital and skilled talent required to build new clouds; advocates for localized data management (GDPR compliance) while utilizing global supply chains.
• Competitive Landscape: The AI cloud sector is expected to consolidate over the next two years; however, CoreWeave does not plan to pursue M&A as a primary strategy, believing talent will naturally migrate to high-performing environments.
• Key Challenges: Securing access to power and high-caliber talent remains the primary bottleneck, alongside the physical complexity of managing supercomputers with over 100,000 GPUs.
• Customer Relationships: Maintains a deep, long-standing partnership with French LLM developer Mistral, which utilized CoreWeave infrastructure from its inception in the US and Europe.

Advice for AI Startups:

• Success relies on building unique products, differentiating from competitors, executing flawlessly, and hiring the right talent rather than focusing solely on fundraising mechanics.
• The most difficult aspect of building a company is assembling a trusted team capable of navigating discomfort and providing mutual support.