What it is
Session recording from Sept 2024 Data Center Frontier Trends Summit exploring how AI workloads increase rack power density and create physical infrastructure challenges. Covers Schneider Electric/NVIDIA collaboration, direct-to-chip liquid cooling solutions, and the role of DCIM, EPMS, and BMS software tools in managing high-density environments.
Why it matters
AI workload proliferation directly correlates with escalating rack power density, creating operational and design challenges for data center operators. Facilities teams must evaluate next-gen cooling strategies (direct-to-chip liquid cooling) and leverage software tools (DCIM, EPMS, BMS, digital electrical design tools) to manage thermal and power constraints in high-density deployments.
Evidence from source:
- Session explores ‘direct impact of AI on rack power density and the ensuing physical infrastructure challenges faced by data center operators’
- Covers ‘direct-to-chip liquid cooling solutions’ as approach to ‘manage and dissipate heat in high-density AI environments’
- Discusses role of ‘DCIM, Energy and Power Management Systems (EPMS), Building Management Systems (BMS), and digital electrical design tools’
Links
- Canonical source: https://www.datacenterfrontier.com/videos/video/55248395/dcfts-2024-session-2-absorbing-the-impact-of-ai-on-rack-power-density-actionable-strategies
- Player: /players/other/
- Topic: /topics/ai-infrastructure/
- Topic: /topics/power-quality-surge/
Open questions
- What specific rack power density thresholds trigger the need for liquid cooling vs. traditional air cooling in AI deployments?
- How do DCIM and EPMS tools integrate with electrical design software to model power distribution constraints in high-density environments?