Founded in 2000, Switch operates some of North America’s largest and most advanced data centers. Its visionary founder, Rob Roy, pioneered industry-standard innovations like hot-aisle containment, transforming how data centers worldwide manage energy. Today, Switch continues this legacy of sustainability and technology leadership. Its high-performance infrastructure services — including cloud computing, telecommunications and sophisticated AI workloads — run entirely on renewable energy. Delivering exceptional reliability and scalability, Switch empowers startups, global enterprises and hyperscalers alike to meet their ESG commitments while using next-generation technologies.
Switch is the only company among over 180 issuers to achieve S&P Global’s highest environmental rating (E-1), alongside top scores in its peer group for social (S-2) and governance (G-2) practices (https://www.switch.com/sustainability/). Recognizing the growing environmental concerns associated with data center water usage, Switch is committed to a net-positive water strategy. Through targeted investments, Switch aims to more than offset its water consumption, protecting the world’s most precious resources and minimizing the environmental impact within the communities it serves.
At-a-Glance
Operating some of North America’s most advanced data centers, Switch proactively adapted its infrastructure to meet AI’s exponential power and cooling demands by building a fully integrated, cloud native, private cloud. Switch partnered with Oxide Computer and SUSE to implement this innovative cloud platform, leveraging Oxide’s hyperconverged hardware and SUSE Rancher Prime, SUSE Linux Micro and SUSE Premium Support. The resulting infrastructure accelerated deployment from days to minutes, delivered real-time telemetry data every 10 seconds and scaled AI workloads sustainably — without increasing overall power consumption.
Switch Powers Sustainable Exascale Data Centers with SUSE
“We’re actually handling significantly larger workloads, but our energy consumption hasn’t increased correspondingly. SUSE Rancher Prime helps us scale sustainably.”
Challenges supporting AI workloads
The rapid rise of AI introduced unprecedented operational complexity and a critical challenge to Switch’s commitment to sustainability, reliability and technology leadership.
AI workloads typically require up to 10 times more power than traditional workloads, introducing new challenges related to power consumption, cooling efficiency, hardware reliability and sustainability (Generational Growth AI, data centers and the coming US power demand surge, Goldman Sachs, April 28, 2024). However, certain advanced hardware — such as NVIDIA’s GB200 racks used at Switch — can far exceed this average, consuming up to 23 times more power (drawing 120kW compared to traditional servers’ 5kW) (GB200 Hardware Architecture – Component Supply Chain & BOM, SemiAnalysis, July 17, 2024).
At Switch’s massive scale, with campuses consuming power comparable to entire metropolitan regions, even minor fluctuations can strain the energy grid and compromise data center reliability.
To manage these exceptionally power-intensive AI racks, Switch adopted direct-to-chip liquid cooling. But in order to optimize this solution, it needed to monitor and manage power usage and environmental conditions more precisely, and at a much greater scale.
Its existing VMware-based infrastructure and monolithic applications only allowed telemetry data collection hourly — far too infrequent for effectively managing power-intensive workloads. Additionally, Switch’s customers operating sophisticated AI environments required detailed, real-time visibility into environmental conditions, power consumption and resource availability. Without granular visibility, optimizing workload performance and energy efficiency was nearly unachievable.
The architecture also lacked the flexibility and scalability required to support modern AI and cloud native workloads. Deployments took days. Replication across multiple data centers required managing multiple sourcing contracts, separate management interfaces and distinct monitoring tools for storage, networking and compute, significantly complicating operations. Transitioning legacy applications also required extensive replatforming and redevelopment for cloud native environments, further extending project timelines and increasing costs.
Switch recognized that to overcome these operational challenges effectively, it needed to fundamentally rethink its infrastructure strategy. Rather than incremental improvements, Switch envisioned a fully integrated, private cloud platform that combined cloud-scale flexibility, operational simplicity, security and powerful real-time analytics capabilities within its own data centers.
Why SUSE and Oxide Computer?
When Switch decided to build a private, on-premises cloud, it prioritized simplicity, efficiency and operational clarity. Traditional infrastructure required assembling distinct hardware and software components, each with separate management interfaces and support contracts. When issues arose, vendors typically pointed fingers at each other, slowing resolutions and complicating operations. Switch sought a fundamentally different approach: a fully unified, scalable and secure platform with minimal fuss.
Switch discovered Oxide Computer’s fully integrated, rack-scale hardware architecture was exactly what they were looking for. Oxide’s innovative design eliminated many inefficiencies common in traditional hardware — such as redundant AC/DC conversions and fragmented hardware components — creating a unified infrastructure platform that significantly simplified management, reduced power consumption and accelerated deployment.
A long-time SUSE partner, Switch introduced Oxide to SUSE to help build a private cloud catered to Switch’s unique needs. The three organizations promptly initiated a proof of concept (POC) to evaluate SUSE’s cloud native technologies atop Oxide’s integrated hardware in Switch’s data centers.
Focusing on critical metrics such as scalability, reliability and security, the POC delivered impressive results. SUSE’s cloud native solutions (SUSE Rancher Prime and SUSE Linux Micro) managed container workloads seamlessly, reduced deployment times from days to minutes and provided the robust, real-time monitoring capabilities Switch needed. Less than a year after initiating the POC, Switch adopted SUSE Rancher Prime and SUSE Linux Micro as integral components of its private cloud platform.
Switch specifically selected SUSE Linux Micro for its lightweight, container-optimized design, security benefits and minimal operational overhead. With its reduced attack surface and rapid, disruption-free updates, SUSE Linux Micro provided an ideal operating system for Switch’s cloud native workloads, particularly as deployments extend to edge environments.
Beyond technical capabilities, Switch also chose SUSE because of its proactive approach to support. Given the scale, complexity and mission-critical nature of its operations, Switch wanted full enterprise support for reassurance and guidance.
“SUSE Premium Support isn’t just about rapid problem-solving,” says Zia Syed, Executive Vice President of Software Systems and AI at Switch. “It’s about proactively avoiding problems in the first place. Having that proactive support significantly reduces our operational risks, ensuring our business remains resilient and agile.”
Throughout the POC, Switch also discovered a deep philosophical alignment with SUSE’s open source values. “Everything was on the table for us to optimize,” says Syed. “Keep it simple, keep it simple — because simplicity’s going to scale. The alignment of that philosophy and mindset was very important for us in this journey.”
The impact of partnering with SUSE
The combined SUSE-Oxide solution provides unparalleled integration for Switch, delivering an on-premises cloud that rivals the ease and agility of public cloud environments. This unified platform enabled Switch to avoid traditional vendor complexities, streamline infrastructure management and significantly enhance operational simplicity and scalability.
“Not all open source is the same,” says Syed. “SUSE partnered with us closely to guide us through this transition.”
Accelerates deployment times from days to minutes, up to 98% faster
Previously, deploying new workloads or system updates required days of careful coordination. With SUSE Rancher Prime’s GitOps capabilities, those same deployments now happen seamlessly within minutes, up to 98% faster.
“SUSE Rancher Prime was transformative for our agility,” says Syed. “Its GitOps-based deployments and real-time orchestration delivered immediate, measurable improvements to our operational speed and efficiency. Tasks that previously took days now take less than an hour, often just minutes.”
Delivers real-time telemetry data every 10 seconds
Accurate, timely monitoring of power and environmental conditions is crucial in managing power-intensive AI workloads. Switch previously collected telemetry data only once per hour — far too infrequent for precise management. Now, SUSE Rancher Prime deployed on Oxide hardware enables telemetry every 10 seconds, representing a 360-fold increase in frequency.
“With near real-time data, we finally have the detailed visibility we need,” explains Syed. “We can now optimize power usage, efficiency and workload scheduling in ways that were previously impossible.”
This increased sampling rate allows Switch to perform multidimensional analysis and rapidly iterate on operational improvements.
Maximizes power usage efficiency
Despite substantial increases in AI-driven workloads, Switch has seen minimal increases in power consumption. SUSE Rancher Prime’s cloud native architecture, automated load balancing and efficient orchestration have enabled more intensive workloads to run across Oxide’s equally efficient stacks within the existing energy footprint.
“SUSE helped us scale our infrastructure to meet the exascale demand we face in the industry today,” says Syed. “Working with SUSE, we built a scalable, reliable and efficient software stack that aligns perfectly with our sustainability goals and innovation needs. We’re actually handling significantly larger workloads, but our energy consumption hasn’t increased correspondingly. SUSE Rancher Prime helps us scale sustainably.”
Enhances scalability and flexibility
Adopting SUSE Rancher Prime, coupled with Oxide’s hyperconverged platform, improved Switch’s infrastructure flexibility. Previously, scaling and replicating VMware-based legacy infrastructure was cumbersome. Now, SUSE Rancher Prime’s cloud native orchestration, combined with the lightweight footprint and efficient manageability of SUSE Linux Micro, facilitates seamless expansion across multiple campuses, significantly simplifying new deployments.
“Our developers can innovate and iterate much faster because of SUSE Rancher Prime’s flexibility,” says Syed. “It means quicker deployments, faster development cycles and accelerated innovation.”
Provides unmatched reliability with 100% uptime
Reliability is critical for Switch’s customers — including hyperscalers and global enterprises — that depend on uninterrupted operations. SUSE Rancher Prime’s robust orchestration capabilities combined with proactive support from SUSE Premium Support ensure rapid resolution of potential issues, helping Switch consistently achieve 100% uptime.
“SUSE Premium Support was crucial for us,” explains Syed. “We simply cannot afford downtime. Knowing SUSE experts are always available gives us confidence and peace of mind, allowing us to focus on our core mission.”
Collaborated to develop a custom Rancher node driver
The collaboration between Switch, Oxide and SUSE was foundational in building Switch’s private cloud. Oxide’s innovative rack-scale hardware with integrated DC power distribution and networking combined seamlessly with SUSE’s cloud native software stack, creating a unified platform tailored specifically for Switch’s needs.
Central to this collaboration was the custom Rancher node driver, jointly developed by Switch and Oxide, with validation and support provided by SUSE Premium Support Services. This critical innovation streamlined Kubernetes deployments, enabling rapid, efficient and secure provisioning of new infrastructure.
“This partnership delivered exactly what we needed — deep integration, technical excellence and a shared vision,” Syed says. “The joint expertise of Oxide and SUSE accelerated our timelines, simplified operational complexity and provided an innovative platform we couldn’t achieve alone.”
This successful collaboration has set a new benchmark for future deployments, providing a model for other enterprises seeking similar efficiency, scalability and sustainability.
What’s next for Switch?
Switch considers its investment in SUSE’s cloud native platform and Oxide’s hyperconverged hardware strategically critical for the coming decade. “This transformation ensures we remain ahead of industry trends and are fully prepared for an increasingly AI-driven and sustainability-conscious marketplace,” says Syed.
Switch plans to significantly expand its adoption of SUSE solutions, particularly at the network edge. As workloads and infrastructure continue moving closer to end users, Switch sees great potential in deploying SUSE Linux Micro. This extension will enable Switch to deliver enhanced performance, security and resilience directly at the edge, meeting customer demands for even faster, more responsive services.
“We’ve already experienced tremendous benefits from our partnership with SUSE, but we’ve only scratched the surface,” Syed notes. “Going forward, we’ll place significant emphasis on optimizing the intersection of hardware efficiency, software orchestration and AI-driven analytics. This will enable us to deliver the next generation of sustainable, high-performance data center infrastructure.”