Projecting the IT and Cooling Landscape in 2030

From Homogeneity to Specialization: Reflecting on Inflection

(Article header: A Sycamore chip mounted in the printed circuit board during the packaging process. Source: Google AI Quantum)

In the first article of this series, we explored how necessity-driven adoption reshaped the datacenter cooling landscape with the rapid rise of liquid cooling and in particular cold plate technology. This inflection point demonstrated a key industry truth: transformative changes and innovations happen when technological requirements leave no alternative.

According to the Cushman & Wakefield Global Data Center Market Comparison, 2024, the global data center operational capacity stands at approximately 35GW today, with projections of an additional 45GW driven by new developments. This explosive growth reflects the increasing demands of digitalization, AI, HPC workloads, big data, analytics and edge computing. These interconnected trends are not only expanding data center capacity but also intensifying the complexity and diversity of thermal challenges.

Building on these insights, this article skips forward to examine the data center ecosystem beyond the next major inflection points. By 2030, the IT landscape will look very different, with specialized platforms driving increasingly diverse thermal challenges. This article provides a framework for understanding how the diversification of workloads and the corresponding thermal demands will shape cooling technology adoption. We’ll explore the diverse IT Equipment ecosystem of 2030, highlighting the roles of air cooling, cold plates, immersion and hybrid solutions, and why their coexistence is essential for optimizing performance and sustainability across the industry.

From One-Size-Fits-All to Purpose-Built IT Systems

In the early days of modern data centers, IT equipment was largely standardized. General-purpose servers, housed in air-cooled environments, were designed to meet a broad range of applications, and cooling was a secondary concern. Air cooling systems sufficed for the relatively modest thermal demands of these setups, allowing for a one-size-fits-all approach to thermal management. However, as computing needs grew, so did the complexity of the IT ecosystem.

Today, the data center landscape is no longer defined by homogeneity. The rapid increase of compute-intensive workloads, including artificial intelligence (AI), high-performance computing (HPC) and hyperscale cloud platforms has driven the development of highly specialized IT systems. These systems are increasingly optimized for specific workloads, each with unique performance requirements and unique power and thermal challenges.

Just as IT systems have transitioned from general-purpose to application-optimized, power and cooling technologies will evolve to address these diverse requirements. This shift lays the foundation for a more diverse ecosystem of cooling solutions, where air cooling, cold plates, immersion and hybrid technologies coexist to meet the specific demands of modern IT applications.

5 Major Trends Shaping The Future

The IT landscape is evolving rapidly. Emerging compute-intensive workloads, such as generative artificial intelligence (AI), machine learning (ML) and quantum computing are reshaping the types of systems in demand. By examining these changes, we can begin to envision what the ecosystem may look like in 2030.

1. Emergence of New Workloads and Shifting Relevance: As workloads become increasingly specialized, the prominence of general-purpose systems is expected to decline. AI, ML and quantum computing workloads will demand entirely new system architectures tailored to their specific needs, such as high-density compute and low-latency interconnects. For instance, quantum-optimized systems and ASIC-dense platforms could emerge as key players in this ecosystem. Until true quantum computing becomes widely available, accessible, or usable in mainstream computing environments, quantum simulation on GPUs or hybrid processors will continue to play a critical role, offering significant performance improvements over GPU-only setups but demanding exceptionally high power densities to achieve optimal results.
2. Integration of Functionality and Specialization: The increasing complexity of IT workloads is already driving the adoption of application-optimized processors like Tensor Processing Units (TPUs) and Intelligence Processing Units (IPUs) for AI workloads, Data Processing Units (DPUs) for advanced networking, Reduced Instruction Set Computer (RISC) architectures for energy-efficient computing and Application-Optimized Integrated Circuits (ASICs) for highly specialized tasks like cryptocurrency mining or video processing. Their integration into new systems will further diversify the IT landscape. Simultaneously, systems optimized for AI, edge-specific deployments and other specialized workloads will gain traction, reflecting the growing need for highly tailored solutions in data centers. Traditional standalone systems, such as those for storage or networking, are expected to become embedded within these more specialized platforms, streamlining functionality and enhancing their overall performance.
3. Technological Innovation and Obsolescence: Legacy technologies, such as low-power networking devices or older storage controllers, may be phased out in favor of more efficient, scalable alternatives like SmartNICs and heterogeneous computing platforms. Innovations in semiconductor technologies, including domain-specific accelerators and integrated CPU-GPU designs, are likely to drive entirely new equipment categories.
4. Thermal Challenges and Sustainability Pressures: Thermal management will play an increasingly pivotal role as power densities rise. Immersion-optimized systems and liquid-cooled platforms will be crucial for managing these thermal challenges. Additionally, regulatory and sustainability goals may push for designs focused on heat reuse and low power consumption, reshaping the cooling and power infrastructure of data centers, from the grid to the chip.
5. Proliferation of Edge Computing: Edge computing is poised to grow as industries adopt localized solutions for real-time processing in applications like retail, healthcare, smart cities and telecommunications. These systems will require compact, energy-efficient designs with highly variable power densities. This shift underscores the need for modular and flexible cooling solutions capable of handling the thermal diversity of edge environments.

Projecting the 2030 Ecosystem

The IT equipment landscape in 2030 will likely be vastly different, shaped by these evolving compute demands, technological innovation and increasingly complex thermal management requirements. While this projection should be treated as speculative, it offers some valuable insights into the potential trajectory for the industry.

• Decline of General-Purpose Systems: General-purpose systems, once a cornerstone of the ecosystem, will likely shrink as specialized systems replace them. These systems will persist and cover a larger variation in load per system (0.8-3 kW) in use cases requiring flexibility but may no longer dominate the market.
• Evolution of Cloud-Optimized Platforms: Cloud-optimized systems will likely remain significant. Although their design should adapt to handle more diverse workloads efficiently, the architecture and power load per system (0.8-1.8 kW) will likely remain relatively stable to ensure flexibility.
• Transition of Modular HPC Compute: By 2030, modular HPC systems are expected to be absorbed into other categories, including Heterogeneous Compute, AI/GPU-Optimized Systems and Edge Computing. This reflects a trend toward greater specialization and the adaptation of modular platforms to meet workload-specific demands in high-performance and localized computing environments.
• The Rise of AI and GPU-Optimized Systems: AI and GPU-optimized platforms may account for 35% of shipped units in 2030, driven by their dominance in AI training and inference workloads. These systems, with diverse loads ranging from 6-60 kW per system, will represent a critical shift toward high-power-density platforms.
• Evolving Role of Storage: Standalone storage platforms are projected to decline as a distinct category. Storage functionalities are increasingly integrated into multi-functional platforms like AI/GPU-optimized, cloud-optimized and heterogeneous compute systems, reflecting the industry’s shift toward unified, workload-specific solutions. The increasing power requirements will likely shift the demands towards 0.8-6 kW.
• Shift from Traditional Networking to HPC Networking: Traditional networking platforms are expected to decline as their functionalities become increasingly integrated into modular and heterogeneous systems. Meanwhile, HPC networking will expand to meet the rising data transfer demands of high-density and AI platforms. This transition reflects the growing need for high-performance interconnects capable of managing the data intensity and bandwidth requirements of next-generation workloads, driving HPC networking power demands to 1.5-4 kW, while traditional networking systems remain at lower power levels (<1 kW).
• Specialized Systems for Emerging Technologies: Quantum, RISC-based architectures and other domain-specific systems will emerge as niche but critical segments by 2030. These platforms will exhibit a wide variety of power densities, form factors and designs. For example, edge-deployed RISC systems optimized for IoT could operate as efficiently as <100 W, while high-performance platforms for quantum simulations using GPUs or similar processors may demand over 10 kW per node.
• Growth in Edge Computing: Edge systems will stand apart due to their localized deployments, latency-sensitive workloads and operation in constrained environments. They are designed for real-time processing in industries like retail, healthcare and autonomous vehicles. Power densities will range from 0.5 kW for IoT gateways to 5 kW for high-performance AI inference platforms.
• Emergence of Heterogeneous Compute: Heterogeneous systems integrating CPUs, GPUs and accelerators are currently embedded within other categories, such as AI-optimized and HPC systems. These are expected to comprise an increasing part of the ecosystem, addressing the need for flexible, high-performance computing across a range of workloads with power densities ranging from 1.5 - 6 kW.

This projection represents one potential outcome based on historical and current trend trajectories. It excludes certain transformative technologies, such as breakthroughs in quantum computing or novel paradigms in chip design, which could drastically alter the ecosystem. Additionally, factors such as regional market and supply chain dynamics linked to regulatory changes will influence the final distribution.

What remains clear is the growing complexity of the IT ecosystem and the corresponding thermal challenges. High-power-density systems, particularly GPU-optimized platforms, will require advanced cooling solutions to ensure operational efficiency and sustainability in the data centers of tomorrow.

Implications for the cooling technology ecosystem in 2030

The story of cooling technologies in data centers is, in many ways, a journey of circular innovation. In the 1960s, engineers faced the thermal challenges of early computing with liquid cooling solutions, exemplified by IBM’s System 360 computers. As silicon technology advanced, reducing power densities and thermal output, air cooling became sufficient for decades. However, as we look toward 2030, the narrative is shifting once again. Driven by the relentless demand for more powerful computing, application-optimized hardware is growing at an unprecedented scale, evidenced by the surge in AI silicon companies and specialized processors, and the industry is returning to liquid cooling.

This back to the future moment highlights how historical lessons can inform the evolving cooling landscape. There is no longer a one-size-fits-all approach to cooling. The increasing diversity of workloads and hardware platforms necessitates equally specialized cooling solutions tailored to meet the unique thermal demands of each application. This evolution calls for the development of application-optimized cooling systems to ensure performance, efficiency, operations and sustainability in next-generation data centers.

• Traditional air cooling will retain its relevance, especially for systems with lower power densities, high interface requirements, or frequent maintenance needs. Networking equipment and legacy general-purpose servers, for instance, remain well-suited to air cooling due to its cost-effectiveness and operational simplicity. While its penetration in new deployments may decline, it will continue to serve as a cornerstone for specific segments of the ecosystem. Although its share will decrease in terms of new equipment which relies on air cooling with 45% in 2030, eventually it will stabilize around 30-35%.
• Cold plate cooling will play a pivotal role in addressing the thermal demands of high-power xPU components like CPUs and GPUs. With precision-targeted cooling capabilities and the ability to efficiently interface with neat chip surfaces, cold plates are indispensable for AI and HPC workloads, where chip power densities surpass the limits of air cooling. Frequently paired with door heat exchangers to manage residual heat, cold plates are expected to dominate high-performance compute environments with approximately 45% of new IT equipment in 2030 being equipped with various types of this technology. Additionally, they will complement immersion cooling solutions to handle the most extreme chip power requirements effectively.
• Immersion cooling is emerging as the ideal solution for environments characterized by distributed heat loads and high-density deployments. Its ability to manage system-wide thermal challenges makes it particularly well-suited for compact platforms, such as edge computing and for ultra-high-density AI and ML systems where the overall thermal load exceeds the air cooling capabilities. Innovations like immersion-precision cooling, which incorporates directed flow technology, further enhance its capability to handle escalating thermal loads efficiently. We can expect approximately 30% of all new IT equipment in 2030 to rely on some form of immersion cooling technology.
• Hybrid cooling solutions, combining the strengths of various cooling solutions (e.g., cold plates within immersion systems), represent a next-generation approach to managing complex thermal demands while improving performance. These systems merge the pinpoint accuracy of cold plates Door Heat Exchangers and with the broad thermal management coverage of immersion cooling, enabling the highest-density workloads, especially for AI and other high density workloads. Hybrid systems are expected to become indispensable for handling the most extreme power densities and diverse thermal profiles.

The Takeaway: Coexistence, Not Competition

This analysis provides a high-level view of cooling technology adoption in 2030, emphasizing functional outcomes rather than distinctions between specific variants like one-phase versus two-phase immersion cooling or open-bath versus enclosed chassis systems. It serves as a framework to understand how the datacenter industry will evolve to address increasing thermal challenges.

As the IT ecosystem continues to diversify, the interplay between these cooling technologies will define the operational efficiency, scalability and sustainability of future data centers. While air cooling will remain the cornerstone for low-power workloads, liquid cooling technologies, particularly cold plates and immersion systems, will dominate the high-density segments. Each technology, whether air cooling with door heat exchangers, cold plates, or immersion, with all its diverse variations, brings unique strengths to the table. By embracing this diversity, the data center ecosystem will remain agile and equipped to meet the challenges of next-generation workloads.

Questions: The article references the return to liquid cooling as a Back to the Future moment. What other historical lessons do you think the industry should revisit to address the challenges of 2030 and beyond? How can data center operators balance the coexistence of diverse cooling technologies while ensuring performance, efficiency and sustainability?

Share your insights below and join me next month as we explore the road to 2030 and delve into the ways in which the industry can adapt and prepare for the next generation of compute, power and cooling.

About Promersion

Promersion is a trusted partner for organizations navigating the dynamic landscape of liquid cooling technologies. Committed to collaboration, Promersion works with a wide network of industry stakeholders to accelerate adoption, drive innovation and establish best practices across the liquid cooling ecosystem. By bridging the gap between technological advancements and real-world business strategies, Promersion empowers the industry to unlock the full potential of liquid cooling technologies.

Acknowledgments

Special thanks to Dev Tyagi and Moises Levy, Ph.D. for proofreading and providing valuable insights which helped shape this article.

For further reading about the drivers and challenges shaping the data center industry, including AI workloads, liquid cooling and sustainability, also see Moises Levy, Ph.D. ’s article, Four Key Trends Disrupting Data Centers in 2025 on DatacenterDynamics .

Image Use Policy

Images by Promersion may be freely used with appropriate credit. Licensed images from Cartoonstock are not available for reuse.

Recreating Graphs

Readers are welcome to recreate graphs based on the data presented in this article for non-commercial purposes, provided proper credit is given to Promersion. Please ensure that any recreated visuals accurately represent the data and context provided here.

Source Data Access

The data presented in this article is intended to provide high-level insights. Access to more detailed source data and context is reserved for Promersion clients as part of Promersion's consulting services. For inquiries, please contact me directly.

#empoweringopen #ocpvolunteers #edge #OCP #datacenter #innovation #opensource #sustainable #sustainability #datacentres #datacentre #DontFearLiquidCooling #immersion #immersioncooling #liquidcooling #netzero #efficiency #heatreuse #supplychain #PlanforIT