What Happened

Crusoe Inc. announced on September 17, 2026 that it has closed a $3.9 billion Series F funding round at a $30.9 billion valuation. The round was led by Atreides Management, Mubadala Capital, and Valor Equity Partners, with participation from more than two dozen investors including Nvidia, Salesforce Ventures, and Robinhood Ventures Fund.

This confirms and exceeds earlier reporting from July 2026, when Crusoe was reportedly raising $3 billion at a $30 billion valuation. The final round came in $900 million larger and $900 million higher in valuation than initially signaled.

Crusoe disclosed several key operational metrics alongside the raise:

• 6+ gigawatts of contracted data center capacity in development
• $140 billion in contracted revenue from data center projects
• 20x+ year-over-year growth in Crusoe Cloud annualized revenue
• $100M+ ARR from Managed Inference within 12 months of launch

The company's flagship project is a 1.2-gigawatt AI campus in Abilene, Texas that Oracle will use to host workloads for OpenAI. A separate 900-megawatt site nearby is being built for Microsoft. According to the Wall Street Journal, some of Crusoe's contracted capacity will also serve a Google campus in Texas.

Crusoe differentiates itself by manufacturing many electrical components in-house — including industrial controls, circuit breakers, and specialized enclosures. It also offers a containerized computing module called Spark with built-in satellite connectivity for deployments lacking terrestrial networking.

Why It Matters

This round is one of the largest private funding events in AI infrastructure history, and it signals that the capital markets view AI data center buildout as a multi-year, multi-hundred-billion-dollar opportunity. Crusoe's $140 billion in contracted revenue — if the figure holds — indicates that hyperscalers and large AI labs are locking in capacity years in advance through third-party builders rather than constructing everything internally.

The competitive landscape is bifurcating. On the physical layer, Crusoe competes with traditional colocation providers and hyperscaler-built facilities. On the cloud platform layer, Crusoe Cloud's Managed Inference and Serverless Fine-Tuning services put it in direct competition with CoreWeave, Together AI, and the inference offerings from AWS, Google Cloud, and Azure.

The $100M Managed Inference ARR milestone is the most operationally significant number in this announcement. It demonstrates that inference-as-a-service — not just training — is generating real recurring revenue at scale. Crusoe's proprietary MemoryAlloy caching technology, which it claims delivers up to 5x throughput improvement for certain workloads, suggests the company is competing on performance optimization, not just raw capacity.

Nvidia's participation is strategically notable. Crusoe is both a massive GPU customer and a platform partner. Nvidia's continued investment in infrastructure operators (it has also backed CoreWeave) reflects a strategy of ensuring demand for its hardware by supporting the companies that deploy it.

Who Is Affected

GPU cloud customers now have another well-capitalized inference platform to evaluate. Crusoe Cloud's Managed Inference and Serverless Fine-Tuning services are directly comparable to offerings from CoreWeave and Together AI, and the $100M ARR figure suggests production-grade reliability.

Enterprise IT leaders at organizations considering AI infrastructure strategy should note that companies like Oracle, Microsoft, and Google are outsourcing data center construction to Crusoe. This validates the third-party builder model and suggests internal build-out may not always be the most efficient path.

AI infrastructure investors should watch the tension between Crusoe's dual identity as both a facilities builder and a cloud platform. The margins, capital requirements, and competitive dynamics of these two businesses are fundamentally different.

Strategic Implications

For AI startup founders: Crusoe Cloud's Managed Inference is now a credible inference platform with proven scale. If you're running open-source models in production, benchmark Crusoe's cost-per-token against CoreWeave and hyperscaler offerings. The MemoryAlloy caching claim of 5x throughput for some workloads is worth testing on your specific inference patterns.

For developers building with AI APIs: Crusoe's Serverless Fine-Tuning service integrates with Amazon S3 for data ingestion and uses Vast Data storage for petabyte-scale datasets. If your fine-tuning pipeline involves large datasets, this infrastructure combination may reduce data transfer bottlenecks compared to running everything within a single hyperscaler ecosystem.

For non-technical business owners: The $140 billion in contracted revenue and 6GW of locked-in capacity signal that AI compute is being reserved years ahead. If your business depends on cloud-based AI inference, consider negotiating longer-term contracts now before capacity constraints drive pricing up. The infrastructure layer is becoming a seller's market.

What to Watch Next

Monitor Crusoe Cloud's pricing and feature announcements over the next two quarters — the company will likely use this capital to aggressively expand its managed services footprint. Also watch for whether CoreWeave or Together AI respond with new funding rounds or pricing changes, as the inference-as-a-service market is now clearly a multi-billion-dollar category.

Frequently Asked Questions

Q: What is Crusoe and what do they do?

A: Crusoe is a Denver-based company that builds AI data centers for major tech firms including Oracle, Microsoft, and Google. It also operates Crusoe Cloud, an AI-optimized cloud platform offering managed inference and fine-tuning services. The company manufactures many of its own electrical components in-house and has 6+ gigawatts of contracted data center capacity in development.

Q: How much did Crusoe raise and at what valuation?

A: Crusoe closed a $3.9 billion Series F round at a $30.9 billion valuation on September 17, 2026. The round was led by Atreides Management, Mubadala Capital, and Valor Equity Partners, with participation from Nvidia, Salesforce Ventures, and Robinhood Ventures Fund among others.

Q: What is Crusoe Cloud's Managed Inference service?

A: Managed Inference is a Crusoe Cloud service that lets users run AI models without managing underlying hardware. It uses a proprietary caching technology called MemoryAlloy that Crusoe claims can deliver up to 5x throughput improvement for certain workloads. The service surpassed $100 million in annualized recurring revenue within its first year.