AI is forcing data center operators to rethink assumptions about power, cooling, density, location and facility lifespan. In this episode of the Moor Insights & Strategy DataCenter Podcast, Matt Kimball and Paul Smith-Goodson examine how infrastructure could evolve by 2036, 2046 and beyond as AI workloads push conventional data center design to its limits.
They explore liquid cooling, higher rack densities, distributed and space-based architectures, onsite power generation and community resource constraints. The discussion then shifts to AI application governance, including platform engineering, Oracle APEX 26.1, citizen development, workload placement and the trade-off between enterprise control and platform lock-in.
Key Topics Covered:
- How AI workloads are driving higher rack densities and accelerating the move toward liquid-first data center designs
- Why traditional 20- to 25-year data center lifecycles may give way to more modular, adaptable infrastructure
- How distributed micro data centers, optical connectivity and eventually space-based infrastructure could reshape capacity planning
- The role of nuclear, natural gas, geothermal, fusion and onsite generation in meeting future data center power requirements
- How excess heat and other data center resources could be redirected back into surrounding communities
- How platform engineering can automate workload placement based on performance, cost, latency, privacy and other requirements
- How Oracle APEX 26.1 and APEX Lang support AI-assisted application development while protecting enterprise runtime, security and governance
- The benefits and trade-offs of enabling citizen developers inside established Oracle environments, including lifecycle management and platform lock-in
- Why Oracle’s support for third-party AI tools such as OpenAI Codex and Claude Code reflects a broader shift toward an open enterprise ecosystem
Watch the full episode:
Disclaimer: This show is for information and entertainment purposes only. While we will discuss publicly traded companies on this show, its contents should not be taken as investment advice.
