Catalyst with Shayle Kann

AI scaling pathways: on grid, on edge, off grid, off planet

30 snips
Mar 12, 2026
Jake Elder, SVP of Research and Innovation at Energy Impact Partners, specializes in built environment and data center energy issues. He walks through four compute pathways: on grid, on edge, off grid, and off planet. Short takes cover grid transmission limits, edge scalability doubts, engineering reliability for off-grid sites, and the physics and feasibility of orbital data centers.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

Transmission Speed Is The Primary Grid Constraint

  • The grid-connected hyperscale path is chiefly constrained by transmission speed and permitting delays.
  • Jake Elder notes interconnection upgrades and new transmission lines often take five to seven years, while major new transmission projects can be effectively stalled for decades.
ADVICE

Bridge Transmission Gaps With Onsite Power

  • Use hybrid designs: stay grid-connected but bring behind-the-meter generation to bypass transmission bottlenecks.
  • Shayle points out many current projects already add on-site generation as a bridge while retaining eventual grid ties.
INSIGHT

Edge Wins On Speed Not Latency Or Cost

  • Edge compute's value is mainly about speed to market, not obvious latency or cost advantages.
  • Shayle Kann and Jake Elder argue latency benefits are limited and fully scaled 300 MW hyperscale sites usually have better levelized costs than many small edge sites.
Get the Snipd Podcast app to discover more snips from this episode
Get the app