Jon Lin, Equinix | Nvidia GTC 2024
summary
Equinix describes a concrete Nvidia partnership that already powers private DGX deployments to keep enterprise data isolated, and engineers for 100 data centers supporting liquid cooling with closed‑loop chillers. The footprint spans 71 metros in 30 countries with over 90% of the world within 10 ms, and more than 95% renewable power coverage. The claim: colocating GPUs next to data and clouds cuts latency and enables ROI horizons customers report as about 12 months.
AI-written summary · about
chapters (10)
- 0:00Introduction to Equinix and Nvidia
- 1:04Key takeaways from Nvidia keynote
- 2:51Equinix's private AI infrastructure model
- 4:16AI factories and data center evolution
- 5:42Training vs inference infrastructure lifecycle
- 6:48Sustainability and liquid cooling in AI
- 8:34Liquid cooling standards and future
- 9:36Open source and domain-specific AI models
- 10:58AI ROI and enterprise investment trends
- 12:03Future of AI investment and super cycle
related talks
01h 13mEngineering the Cloud for Tomorrow: Performance, Speed and Scale for AI Workloads
14mWhy We Don’t Need More Data Centers
18mThe Geopolitics of AI Infrastructure
14mFireside
11mthe future of cloud
31mHow to build a sovereign European AI compute infrastructure
18mArchitecting Multi-tenant Data-center Networks for GPU Customers
05mFrom Open Source to Real-World Scale