If you’re reading this, chances are you already know a thing or two about immersion cooling. So instead of rehashing how the technology works, let’s focus on something more important: the challenges. Despite its promise, immersion cooling has yet to see widespread adoption—and that’s not by accident. You might be wondering: Why would a company that promotes immersion cooling start by talking about its downsides? The answer is simple. That’s exactly why BluWaive was founded. We’re here to solve the real-world challenges that have held this technology back and finally make large-scale deployment a reality for our customers.
A Brief History Lesson
Before we dive in, let’s take a step back and look at the history of immersion cooling and how early adopters have implemented the technology. We’ll explore how different organizations have approached immersion technology. Then, we’ll examine the leading companies in the space, their unique value propositions, and the challenges they continue to face when deploying at scale. Finally, we’ll dig into the fundamental barriers to immersion cooling adoption—and how BluWaive is uniquely positioned to overcome them (stay tuned for part 2 of this series).
1985: Cray-2, The First Fully-Immersed Supercomputer
Immersion cooling dates back to the 1980s, when computing pioneer Seymour Cray partnered with 3M to use a novel dielectric fluid called Fluorinert to cool his next-generation supercomputer. This collaboration led to the development of the Cray-2, one of the first major systems to use liquid immersion for thermal management. Delivering 1.9 gigaflops of performance, the Cray-2 was the fastest computer in the world when it launched in 1985—a title it held until it was succeeded by the Cray X-MP in 1988. While the technology proved effective, widespread adoption remained elusive due to its complexity, high cost, and the absence of standardized infrastructure. Despite these challenges, immersion cooling has continued to evolve, with a growing number of providers now entering the space and driving innovation forward.
For you history buffs out there, here is an excellent YouTube video on the history of the Cray-2.
The Evolution Continues: Key Players and Innovations
Fast forward to the early 2010s with the push for higher compute densities and the rapid growth of hyperscale data centers, immersion cooling began to experience a renaissance. A wave of new companies entered the space, all aiming to address the mounting challenges of rising thermal densities and increasing energy costs in large-scale compute environments. Among the most notable are Submer, Sustainable Metal Cloud, and LiquidStack—each of which has played a key role in shaping the modern immersion cooling market. Their innovations and advocacy have brought immersion cooling from the confines of lab-scale experimentation to the forefront of real-world deployment. Without their contributions, the industry might still be struggling to recognize the full potential of this transformative technology.
2015: Submer Makes Single-Phase Cooling Mainstream
Founded in 2015 in Barcelona, Spain, Submer aimed to tackle both energy efficiency and infrastructure density challenges by primarily focusing on single-phase immersion cooling solutions. In recent years, they’ve gained visibility through partnerships with Intel, Supermicro and data center operators aiming to test immersion’s viability at the hyperscale level. They’ve also received multiple rounds of funding, participated in the Open Commute Project (OCP), as well as made many strides in helping standardize immersion deployments, particularly through their SmartPod and MicroPod product lines.
2017: LiquidStack and the Rise of Two-Phase Immersion
As mentioned, 2017 was a notable year with LiquidStack making headlines by debuting what was then considered the world’s largest two-phase immersion-cooled datacenter in Azerbaijan at over 30MW of IT power. This was a significant step, showcasing the potential of two-phase immersion cooling. The datacenter soon grew to support over 120MW by 2018 and still holds the record for largest two-phase focused datacenter. In the link below, LiquidStack highlights their portfolio of two-phase tanks and their advantages over conventional air-cooled deployments. The datacenter is primarily used for blockchain but the size and scale of their deployment was the first step in proving viability of two-phase cooling technology at scale.
2018: Sustainable Metal Cloud (SMC) focuses on GPU-Powered AI with Immersion
Founded in 2018, Australia-based Sustainable Metal Cloud (SMC) emerged with a strong vision to make cloud computing more sustainable, particularly for large-scale GPU AI infrastructure. Partnering with NVIDIA, SMC developed its HyperCube™ platform, an integrated single-phase immersion cooling system. They’ve reported significant energy savings—up to 50% compared to conventional air cooling and a 28% reduction in installation costs. SMC’s approach emphasizes not just efficiency but also genuine sustainability, aiming to cut CO₂ emissions significantly. Their technology, which uses a synthetic oil called polyalphaolefin, is designed to support high-density GPU hosting, crucial for the advancements in AI and large language models. They completed a 20MW research and development immersion-cooled data hall in 2021 and have since expanded, opening a global headquarters in Singapore in 2023.
The Broader Adoption and Its Hurdles
The journey of immersion cooling from niche applications to broader consideration has been accelerated by several factors. The Open Compute Project (OCP) officially embraced immersion cooling in 2018, establishing a project for Advanced Cooling Solutions (ACS) and presenting the first industry standards for immersion in 2019. The Telecommunications Industry Association (TIA) also recognized immersion cooling as a viable option in 2020. The relentless rise of high-TDP (Thermal Design Power) CPUs and GPUs, especially for AI, machine learning, and HPC, is pushing air cooling to its limits, making liquid cooling, including immersion, an increasingly necessary solution.
However, as we at BluWaive acknowledge, widespread adoption hasn’t been without its challenges. These include:
- High Upfront Costs: Investing in specialized immersion cooling infrastructure can be more expensive initially compared to traditional air cooling. Furthermore, since many equipment manufacturers (ODMs and OEMs) don’t offer standard immersion cooling products, custom designs and unique infrastructure are often necessary for each client.
- Hardware Compatibility Challenges: Not all IT hardware is designed to be submerged in immersion cooling fluids, often necessitating modifications or specific server designs. While the OCP immersion subproject committee has made significant efforts to standardize material compatibility and evaluation for various immersion fluids, compatibility testing still lacks standardization. This often requires chemical expertise to interpret results and perform specialized tests.
- Maintenance and Service Complexity: Working with liquid coolants and submerged hardware introduces new maintenance procedures that may require specialized skills or training. Immersion cooling systems can often be more challenging to service compared to air or direct-to-chip cooling methods.
- Overcoming Inertia and Standardizing a Newer Technology: As a relatively new mainstream technology, immersion cooling faces the challenge of overcoming existing practices and establishing comprehensive industry-wide standards. Many large-scale deployments are still considered early adoptions rather than the standard.
- Managing Cooling Fluids: The handling, monitoring, and eventual disposal or recycling of dielectric fluids demand careful planning. Waste management processes must be in place for when the fluid is no longer needed or during system decommissioning. For instance, fluorochemicals often need to be recycled or incinerated, though these terms are typically included in service level agreements with fluid vendors for large installations.
BluWaive: Addressing the Next Generation of Immersion Challenges
This brings us back to why BluWaive was founded. While companies like Submer, LiquidStack, Sustainable Metal Cloud have made invaluable contributions by proving the technology’s viability and pushing its boundaries, the path to truly scalable, seamless, and universally adoptable immersion cooling still has obstacles. At BluWaive, we’ve studied this history, learned from the successes and the roadblocks, and are engineering solutions to address the practical, real-world challenges that still hinder large-scale deployment. Stay tuned for part 2 of our series where we will break down each of these remaining barriers and how BluWaive is working to solve them.