The McKinsey Podcast Data centers: The race to power AI
11 snips
Jun 5, 2025 Leaders are racing to scale AI, but this requires expanding data centers, which consume immense energy. The podcast dives into the challenges of building these facilities rapidly, the shift in their locations, and innovative cooling solutions needed to meet sustainability goals. It also discusses navigating the energy demands of data centers in the new AI economy, highlighting the urgency for investment and the exploration of renewable energy initiatives. Collaboration and resilient business models emerge as critical to addressing these challenges.
AI Snips
Chapters
Transcript
Episode notes
AI Economy's Massive Data Needs
- The AI economy will create $4 trillion value by 2030, driving massive data center expansion.
- AI's rapid growth relies on vast computing power and data storage availability.
Data Center Power Demand Surge
- Global data center power use will triple from 70 to 220 gigawatts in five years.
- AI drives 75% of growth and changes data center design with high power needs and liquid cooling.
Energy Challenges in Data Centers
- AI demands huge energy and construction investments for new data centers.
- Many sites lack grid capacity to support multi-hundred megawatt power needs, requiring infrastructure upgrades.