Crusoe Series D
Despite the excitement, organizations seeking to develop and deploy AI solutions face a series of challenges ranging from resource instability to platform complexity to a lack of customer support. In addition, data center vacancy rates are at historic lows, and most existing sites lack the power and cooling capacity needed to support the high density of next-generation GPUs. These challenges span from the physical to the digital, from steel to fiber, from electrons to bits, and from breakers to software. They continue to impede the large-scale adoption of AI solutions, preventing organizations from reaching their full AI potential.
Solving these problems is one of the most important challenges of our generation. Success will unlock a whole new chapter of human prosperity leading us from the information age to the age of intelligence. Abundant intelligence will lead to abundant innovation, abundant clean energy, and abundant progress. We are embarking on a new era of AI industrialization and Crusoe is leading the way.
At Crusoe, we believe the best way to solve complex problems is to “make the problem bigger.” This means addressing interconnected challenges as a unified whole. This is the approach we are taking to develop AI cloud infrastructure. Crusoe is bringing critical cross-functional expertise that spans the great divide between the physical and digital worlds. We bring systematic solutions to the challenges brought by the convergence of energy and computing.
Crusoe is the industry’s first vertically integrated, purpose-built AI infrastructure provider. We provide a reliable, cost-effective, and environmentally aligned AI platform by harnessing large-scale clean energy, building AI-optimized data centers, and simplifying the development of AI solutions. Crusoe Cloud, our reliable, feature-rich platform caters to organizations seeking to develop and deploy AI solutions. Our data centers, which underpin Crusoe Cloud, also provide hyperscalers and enterprises with state-of-the-art AI-optimized infrastructure.
Since launching Crusoe Cloud, we have continuously honed our efforts, learned from customer feedback, and optimized our approach. This resulted in a 7x growth in our customer base over the last year alone. Building on this momentum, we were pleased to announce earlier today Crusoe Cloud’s general availability. Starting today, developers, machine learning engineers, and data scientists can access our cutting-edge resources, managed services, and enterprise-grade support for AI exploration, model training, fine-tuning, and scalable inference workloads.
As part of today’s announcement, we are excited to share that we raised an additional $600 million in Series D funding. This capital will accelerate our investment in data centers and further expand Crusoe Cloud’s capabilities to simplify and accelerate AI development. To explain some of the reasons behind our growth and enthusiasm among the investment community, let’s unpack the innovations we’re delivering at each layer of the stack.
Crusoe Cloud
Our aim is to make Crusoe Cloud the world’s favorite cloud for AI. It is designed to reduce complexity and accelerate time to value. We are experts in AI workloads and have invested heavily in the usability of the platform to streamline the experience and help customers get their workloads up and running quickly. Crusoe Cloud’s managed services reduce complexity by freeing customers from spending time orchestrating underlying infrastructure, enabling them to focus on what matters most - building their products.
“We rely on the Crusoe Cloud to train our models on massive amounts of data. Crusoe’s usability is a key differentiator. Their managed services free up precious resources to enable us to focus on our core expertise: building models.” Marvin Gabler, Co-Founder & CTO Jua
Crusoe Cloud offers industry-leading start times for H100s and delivers consistent performance at scale. According to a recent SemiAnalysis report, our VMs can boot up in less than 90 seconds and we set the “gold standard” for GPU scaling. This performance is in part due to our intimate knowledge of building out data center clusters. We are experts on InfiniBand and cabling, enabling us to deliver resilient performance that scales smoothly even for the most demanding workloads, like real-time video generative experiences.
"With Crusoe, we scaled our capacity 5x within hours to serve all of our Oasis users across Europe. This enabled Oasis to seamlessly scale to over 2 million users in just 4 days." Dean Leitersdorf, Co-Founder & CEO Decart
H200s will be available soon, with GB200s anticipated in the Spring 2025.
Details matter, especially when it comes to reliability. We have carefully architected Crusoe Cloud and developed capabilities that allow us to quickly swap out nodes when needed. The result is an enterprise-ready platform, where customers can “set it and forget it.”
"Codeium's NVIDIA H100 Tensor Core GPUs on Crusoe have been incredibly reliable with a cluster uptime of 99.98%.” Varun Mohan, CEO & co-founder of Codeium
We succeed when our customers succeed. This is why we built an enterprise-grade support team that offers a mean response time of as little as four minutes. Our product roadmap is heavily influenced by customer feedback. If customers need a new feature, chances are we can make it happen very quickly.
“Crusoe provides a level of quality of service, responsiveness, and support for early access programs that we couldn’t find with any other cloud provider. Crusoe’s quality of service in particular is exceptional and we look forward to partnering with them in 2025 and beyond." Prasanth Veerina, Co-Founder Pixelcut
Data Centers
Another challenge we are tackling is the lack of AI-optimized data centers. GPU power and cooling requirements continue to increase dramatically, which means that the traditional hyperscaler data center design is no longer optimal. Our data centers are built from the ground up for GPUs, feature high-density racks and direct liquid-to-chip cooling, and enable us to support the most demanding AI workloads that traditional data centers can’t handle.
In addition, Crusoe is able to bring new data center capacity online at unmatched speed. We are on pace to set a record for the greenfield development of a hyperscale data center at our site in Abilene, with 206 megawatts of liquid-cooled GPU capacity coming online in the first quarter of next year. We broke ground on this site in June 2024 and we have already turned over access to the first data hall for tenant improvements - a delivery timeline of less than six months (unheard of in the data center industry at this scale for greenfield construction). Crusoe’s speed of delivery is positioning our customer to activate one of the largest GB200 clusters in the world next year.
Speed matters in this business. There are several reasons why we are able to bring new capacity online faster than anyone. The insights we gain from Crusoe Cloud enable us to understand the evolving needs of next-generation GPUs and plan accordingly. We also in-source the manufacturing of long-lead data center equipment, reducing our dependence on third parties and enabling us to source critical items in a fraction of the time.
Another factor is our team. We are experts in data center design, engineering, construction, and operations. Our team has decades of experience and a track record for building state-of-the-art infrastructure. Collectively, we employ several hundred manufacturing and construction workers, all with the goal of powering the AI revolution.
Energy
The lack of power needed to support the global demand for AI workloads is a key challenge we are taking on. Following the launch of ChatGPT and other large language models, the power demand from data centers has skyrocketed. In fact, data center energy consumption is on track to top 10% of the global energy supply by 2030, placing a significant burden on existing power resources. This is already hindering AI adoption and some hyperscalers have admitted that the lack of power is an obstacle to their growth.
Crusoe’s ability to quickly access energy at scale enables another critical speed advantage. We are energy experts. We have decades of experience in the engineering, deployment, and operation of power generation and electrical infrastructure. Blending this skillset with the urgency and creativity of a startup has positioned Crusoe to de-bottleneck the energy constraint across our portfolio of data center development opportunities.
With this expertise, Crusoe is developing on-site power plants, utility connection points, and energy optimization strategies for our data centers. Our focus is to source clean energy and to do so at prices well below market rates. By sourcing (and in some cases generating) our own energy, we derisk our infrastructure timelines and bottlenecks, and can operate with the agility of a startup rather than depending on entrenched energy utilities.
Get started today
Crusoe is paving the way to help organizations realize their full AI potential. While some may be daunted by today’s challenges, our philosophy is the obstacle is the way. We are Crusoe. And we are powering the AI industrial revolution.
Join us for the journey and see why so many customers rely on Crusoe Cloud and our data centers for their AI needs.
