AI rack density is reaching its practical limits, with electrical engineering and resilience planning, rather than chip counts, becoming the primary constraints for systems exceeding 100 kW.
Thermal Management Shift
emerging tech
deals
sentiment shift
Driven by soaring chip power densities, evidenced by Microsoft's microfluidics and major vendor pivots, facility cooling is transitioning from macro to integrated solutions. This requires professionals across hyperscale and enterprise sites to urgently re-evaluate power distribution, plumbing, and the entire thermal loop supporting dense compute.
Data centers are increasingly employing artificial intelligence for optimized operations. Digital Realty, AWS, DPR Construction, and Schneider Electric are leading this trend, with analysts predicting significant industry evolution.
As artificial intelligence workloads drive data center rack densities beyond air cooling capabilities, direct-to-chip, immersion, and rear-door heat exchanger liquid cooling solutions present distinct advantages and trade-offs.
opinion
AI data centers can improve efficiency and performance by adopting strategies similar to those used by Formula One racing teams, focusing on operational excellence and rapid optimization.
ZutaCore is implementing strategies to scale its global artificial intelligence infrastructure, focusing on enabling efficient and robust AI operations worldwide.
The article examines the practicality and implications of high-temperature superconducting feeders for data center operators, exploring their potential benefits and technical aspects.
Thermoplastics are playing a crucial role in supporting the demanding requirements of modern data center workloads, contributing to essential infrastructure solutions.
Industry veteran Ron Vokoun discusses navigating market changes driven by artificial intelligence, focusing on prefabrication, liquid cooling, and current market bottlenecks.
Becky Wacker of Trane discussed how artificial intelligence is transforming the balance between power, cooling, and compute performance in data centers during a DCD Talks session.
A direct-on-die two-phase cooling method is proposed to reduce thermal resistance and enable warmer facility water for next-generation artificial intelligence graphics processing units.
DCD Talks featured Schneider Electric's Tuan Hoang discussing efficient water usage in data centers through closed-loop cooling for sustainable AI infrastructure.
Server efficiency and cooling benefits can be achieved through reduced density, according to the Uptime Institute's 2025 Global Data Center Survey results, highlighting operational optimization strategies.
This article details the critical first 30 minutes following a liquid cooling alarm in a data center, emphasizing immediate response protocols for preventing failures.
Hydrogen fuel cells offer a path to zero-emission backup power for data centers, enhancing capacity and resilience while supporting sustainability goals.
High-performance AI accelerators are overheating data centers, necessitating advanced two-phase cooling solutions to manage thermal challenges.
This explainer examines data center cooling methods—air, evaporative, liquid, and hybrid—detailing tradeoffs in energy, water, density, cost, and complexity.
→
Rethinking data center engineering for the artificial intelligence era involves adapting to higher power density, adopting liquid cooling, evolving electrical systems, and implementing hybrid energy strategies.
Addressing rising chip temperatures is transforming data center cooling strategies, requiring heat rejection technologies to be matched with chip performance, climate, and long-term efficiency goals for artificial intelligence workloads.
The global market for data center chillers is expected to surpass 6.4 billion dollars by 2035, reflecting increasing cooling demands.
Two-phase liquid cooling is presented as the emerging standard for data centers, supporting next-generation artificial intelligence infrastructure.
Liquid-cooled load banks can simplify data center readiness by validating cooling infrastructure before the deployment of artificial intelligence IT equipment.
This article offers insights for data center owners on navigating power quality, cooling, and backup energy challenges to effectively support the increasing demands of artificial intelligence workloads.
Two-phase cooling is attracting investment as a successor to water cold plates, indicating a shift in data center thermal management strategies.
This article advises on developing a cooling strategy for next-generation artificial intelligence by aligning thermal architecture with future demands in graphics processing unit density and performance.
The market for data center refrigerants is anticipated to exceed USD 1.62 billion by 2035, driven by the growing need for efficient cooling solutions in global data centers.
Sufficient power and cooling are essential requirements for the optimal functioning of artificial intelligence applications within data centers.
HPE's Andrew DesRochers explained that as artificial intelligence infrastructure scales, power availability, cooling efficiency, utility project timelines, and water consumption are becoming critical considerations in IT planning and decision-making.
Increasing thermal density in AI hardware is rendering traditional air-cooled data center assets obsolete, driving a shift towards liquid-cooling solutions and creating a bifurcation in the tenant market, with emerging markets facing a significant cooling penalty.
A minor error in coolant temperature, as little as half a degree Celsius, can inadvertently cost data centers millions of dollars.
As escalating AI workloads increase rack densities, data center operators are compelled to reassess their cooling strategies due to colliding concerns over power consumption, water usage, and existing infrastructure limitations.
Lambda is showcasing Nvidia's CPO switch with the Microsoft-backed neocloud's liquid-cooled Quantum-X Q3450-LD, highlighting significant power savings that enable 'more tokens' and eliminate the need for traditional transceivers.
Continuous thermal monitoring is crucial for building resilient, AI-ready power infrastructure that enhances capacity, improves efficiency, and supports the future growth of data centers.
The global market for data center chillers is expected to exceed USD 6.4 billion by 2035.
This article provides a practical guide to implementing rack-level cooling solutions for high-performance GPU environments, focusing on retrofitting for increased density.
Accelsius is securing funding to support new products and a 300 MW customer project, advancing its focus on two-phase cooling solutions and coolant fluid research.
Advanced cooling technologies are significantly influencing and reshaping the design decisions being made for data centers.
During a DCD>Talks session, Vertiv's Maurizio Frizziero discussed how increasing chip temperatures affect global cooling infrastructure and identified three key pillars for advancing the AI transformation.
Barocal has raised $10 million to advance its solid-state cooling system for data centers, which could provide an efficient alternative to traditional air cooling methods.
Data Centre World London 2026 will explore critical questions for artificial intelligence-driven operators regarding their partnerships with infrastructure providers.
The Data Center World 2026 Expo highlighted advancements in liquid cooling technologies, showcasing the latest equipment designed for cooling artificial intelligence infrastructure.
The article discusses the adoption of liquid cooling technologies, specifically direct-to-chip architecture, for AI data centers to enhance cooling efficiency, reduce resource dependency, and support next-generation infrastructure with higher rack densities.
As artificial intelligence workloads intensify, cooling systems are transitioning from a design element to a critical deployment challenge, requiring operators to adapt to significantly increased power densities.
To address the extreme power densities driven by artificial intelligence, data center operators are increasingly adopting hybrid cooling solutions, modular construction methods, and tightly integrated systems to maintain optimal performance, efficiency, and reliability.
Experts at Data Center World are discussing how the increasing demands of generative and agentic artificial intelligence are compelling a reassessment of power, cooling, and overall infrastructure strategies for data centers.
Michel Chartier of WSP discusses the evolution of engineering teams in the data center industry, emphasizing the need for specialized training to handle advanced cooling systems and emerging 800V DC infrastructure driven by artificial intelligence workloads.
The article discusses the adoption of liquid cooling for high-density computing needs, particularly in the context of AI workloads, while noting a slower uptake by traditional enterprise IT departments.
Precision liquid cooling and industrial resiliency are crucial for scaling artificial intelligence infrastructure, driving data center performance through enhanced resilience and speed.
Brian Lillie of ZutaCore discussed the challenges of rising heat and power demands driven by artificial intelligence and explained how their waterless two-phase liquid cooling technology supports next-generation, high-density data center deployments.
While microfluidic cooling shows promise for future data centers, its widespread adoption faces significant hurdles related to manufacturing complexity and ongoing maintenance requirements.
The increasing demand for artificial intelligence data centers is leading to significant backlash over water consumption, prompting exploration into technologies like atmospheric water harvesting to generate sustainable cooling water from the air.
Akash Systems, in collaboration with AMD and Nvidia, is pioneering diamond-based cooling solutions to address the thermal challenges hindering the scalability of artificial intelligence in data centers.
Vertiv is enhancing its thermal management offerings for AI infrastructure through the acquisition of ThermoKey, aiming to address critical cooling bottlenecks and strengthen its position in the AI hardware market.
Artificial intelligence workloads have significant implications for the design of high-density data centers.
Waterless cooling technology is emerging as an essential component for building scalable infrastructure to support artificial intelligence factories.
The data center HVAC market is projected to exceed USD 36 billion by 2035, indicating significant growth in cooling infrastructure.
Jason Hoffman of Switch DataCenters discusses the impact of artificial intelligence on data center operations, including exascale facilities, high-density systems, and integrating power and cooling infrastructure for optimal performance with Kat Sullivan during DCD>Talks.
The strategy of 'icebanking' is being explored as a method to mitigate increasing energy expenses within digital infrastructure operations.
Executives from Flex detailed their new integrated, factory-built platform—incorporating liquid cooling, higher-voltage DC power, and global manufacturing—designed to significantly reduce deployment timelines for AI-scale infrastructure.
Eaton's $9.5 billion strategic pivot toward thermal solutions is analyzed as a key indicator potentially signaling the next trillion-dollar phase of power, cooling, and deployment strategy across the US artificial intelligence data center market.
Vertiv's recent $1 billion strategic acquisition is an aggressive attempt to establish dominance over the critical reliability layer that now ultimately governs the performance outcomes of artificial intelligence cooling solutions.
Microsoft introduced an in-chip microfluidic cooling system that uses etched pathways within the silicon to remove heat up to three times more effectively than standard cold plates, promising denser and more cost-efficient AI hardware.