AI buildout fuels DCPI growth
AI infrastructure keeps fueling another quarter of rapid growth across the data center supply chain. Dell’Oro Group says the global market for data center physical infrastructure (DCPI) expanded 28% year over year during the first quarter, marking a fifth consecutive quarter of growth above 20% as hyperscalers, neoclouds, and colocators continue investing aggressively to support AI deployments. Thermal management led all infrastructure segments, growing nearly 50%, while direct liquid cooling continues to reshape data center cooling architectures and power distribution technologies such as busbars also outpaced the broader market.
The report argues that AI demand remains constrained less by compute availability than by the industry’s ability to deliver power. Speaking with RCRTech, Dell’Oro research director Alex Cordovil said the primary bottleneck sits upstream of the white space, where transformers, medium-voltage switchgear and on-site generation face multi-year lead times that were never designed for today’s pace of AI infrastructure expansion. “Liquid cooling gets the attention, but you can stand up a cold plate loop far faster than you can energise a substation,” he said.
Cordovil also noted that while liquid cooling is transforming thermal architectures, another area deserves greater attention: heat rejection. Once heat is removed from AI chips, operators must still dissipate it efficiently through chillers, dry coolers, and related systems, making local factors such as water availability and climate increasingly important in facility design. Meanwhile, permitting delays and community opposition are emerging as additional risks in some markets, potentially pushing future AI capacity toward regions where power can be brought online more quickly.
Juan Pedro Tomas
Editor
RCRTech
AI Infrastructure Top Stories
Tranxform edge chips: Founded by a veteran of Apple and Amazon, Taiwanese startup Tranxform is developing low-power edge processors to run large generative AI models directly on devices rather than in data centers.
Floating data center push: Korean company HD KSOE and Schneider Electric inked a deal to jointly develop infrastructure technologies for floating data centers, targeting offshore engineering solutions as interest grows in alternatives to land-constrained, energy-intensive facilities.
AI Today: What You Need to Know
Meta’s Canada AI hub: Meta will invest more than $9.16 billion in its first Canadian AI data center, a 1GW Alberta campus designed for AI workloads, expanding hyperscale capacity while strengthening power and community infrastructure.
Cerebras targets Europe: Cerebras plans 200MW of AI data center capacity across Europe by 2027, targeting sovereign AI demand with inference-focused infrastructure as it challenges Nvidia’s dominance in regional AI compute.
Malaysia adds 210MW: Malaysia’s Global Telecommunications Group acquired land for a planned 210MW data center in Selangor, expanding the country’s AI infrastructure footprint near one of Southeast Asia’s largest digital infrastructure hubs.
Korea’s AI push: NFD Korea is developing a 300MW AI-ready data center campus with secured power, permits, and phased expansion, underscoring South Korea’s accelerating investment in AI infrastructure and hyperscale capacity.
Gujarat courts hyperscalers: Gujarat has launched its first data center policy, targeting 7.5GW of new capacity through incentives aimed at accelerating hyperscale, cloud, and AI infrastructure while strengthening India’s digital infrastructure ambitions.
AI accelerators surge: The global data center accelerator market could exceed $270 billion by 2035 as AI infrastructure investment fuels demand for GPUs, ASICs, and custom silicon across hyperscale and enterprise deployments.
RCR Events
Industry Resources
Webinar, July 7th: Noise-Figure Measurements with RFmx and PXI VSTs
Webinar, July 16th: NTN in motion — evolving standards, expanding services
Whitepaper: Powering sovereign AI at scale
Whitepaper: Scalable database design for 5G and beyond
Report: Scaling AIOPs from insight to action
Summit Access: GSMA Device Enablement Summit: How operators can fix device-network fragmentation
Whitepaper: Telco AI Enabler: Mediation’s defining role
Report: Securing telecom infrastructure for the quantum era
Report: Scaling optical networks for the AI and hyperscale era