Powering the AI Revolution: What Goldman’s Forecast Means for Data Infrastruc
- DPS
- Jul 17
- 2 min read
“From kVA to MW: How AI is Redefining Data Center Power Planning”
Introduction
As AI workloads surge, we’ve reached an inflection point in data center design—shifting from thinking in kilovolt-amperes (kVA) to megawatts (MW). Goldman Sachs’ latest server forecast signals that facilities must be architected for over‑1 MW per installation—a seismic change from traditional scales. At Data Power Supply, we recognize this as a pivotal moment: the conversations that once revolved around kVA are now firmly in the megawatt era.
1️⃣ Why the Megawatt-Scale Matters
Explosive compute demands: AI, ML, and large language models now drive power requirements into multi-megawatt ranges.
New architectural constraints: Design strategies must consider power density, thermal management, and infrastructure capable of sustained high-load operations.
Beyond traditional planning: It’s not just about generators or UPS—it’s about full-system coordination across power, cooling, energy sourcing, and reliability.
2️⃣ Engineering Challenges in Mega-Scale Deployments
Precision electrical infrastructure: From MV switchgear and redundant feeders to advanced PDUs—upgrading capacity and ensuring resilience is non-negotiable.
Advanced cooling systems: High-density servers push heat loads beyond what legacy air-cooling can handle. Liquid cooling, immersion, and hot‑aisle containment are rapidly becoming table stakes.
Integrated microgrid support: On-site generation (diesel, gas turbines), renewable integration, and energy storage systems like UPS or BESS are essential for resiliency and efficiency.
3️⃣ Data Power Supply’s Strategic Approach
Megawatt-grade PDUs & MV solutions – designed for ultra‑high density environments.
Modular power architectures – phased MV/LV deployment supports evolving compute needs.
Cooling & energy synergy – pairing compute with chilled-water systems, heat reuse, and integrated energy storage.
Edge-ready pre-fab power modules – rapid deployment with containerized MV/LV substations and UPS.
Sustainable electrification – embracing hydrogen-ready generators, renewables, and future‑proof microgrid interfaces.
4️⃣ Why Holistic Design Wins
Speed to market: Designing with 1 MW+ awareness avoids redesign and deployment delays.
Future scalability: Modular and UPGR‑architected sites can scale with growth and evolving workload demands.
Operational reliability: Resilience against power disturbances and hot swap upgrades ensure near-zero downtime.
Efficiency gains: Proper cooling and load balancing reduces OPEX and carbon footprint.
💡 Final Takeaway
Goldman Sachs’ pivot—from kVA to MW—is more than a metric shift. It signals an era where compute, power, and cooling are engineered in unison. At Data Power Supply, we see this as the blueprint for tomorrow’s AI-ready architecture.
Your hyperscale design doesn’t have to stall. Let’s rethink power as a system—with MV backbone, resilient cooling, and smart energy integration—all built from the ground up for megawatt workloads.
Want to architect the next generation of AI data centers?
Connect with our team to explore how modular power systems, high-density PDUs, and integrated energy storage can give you a scalable, reliable edge.
Comments