Powering AI When the Grid Can’t - Data Center Frontier Show Podcast скачать в хорошем качестве

Powering AI When the Grid Can’t - Data Center Frontier Show Podcast

This is a video presentation of the recent episode of the Data Center Frontier Show Podcast entitled: "Powering AI When the Grid Can’t: Inside the New Behind-the-Meter Playbook" In this special edition of the DCF Show Podcast, Editor in Chief Matt Vincent hosts a recast and 2026 update of a pivotal DCF Trends Summit 2025 conversation. Utilities aren’t just “behind.” In many markets, they’re years behind—colliding head-on with AI’s buildout schedule. With interconnection queues stretching into the late 2020s and early 2030s in some regions, the industry is moving beyond “backup gensets” and into a new era where onsite and behind-the-meter power becomes foundational—built for schedule certainty, resilience, and, increasingly, strategic control. Moderator Fengrong Li, a Senior Managing Director at FTI Consulting with more than 20 years of experience in energy and infrastructure, leads a wide-ranging discussion on how integrated power + data center campuses are reshaping site selection, redundancy, and the economics of AI infrastructure. Li has provided expert analysis before FERC, the FCC, and state regulators and commissions, and previously spent 13 years at Siemens on grid issues and 8 years at Mitsui in commodity trading and infrastructure investment. Joining the conversation: Brian Gitt (Oklo) on modular nuclear generation and emerging “additionality” procurement models David Blank (Siemens Energy) on gas turbines as near- and mid-term backbone—and the reality of turbine lead times Marty Trivette (AlphaStruxure) on microgrid integration, financing structures, and designing for AI load volatility Yuval Bachar (ECL) on distributed, modular architectures and hydrogen/flexible fuel pathways, including urban inference constraints Topics covered include: • Why behind-the-meter power is accelerating—and why standard designs still don’t exist • The rise of modular, phased buildouts that mirror how AI campuses scale • AI “transient” load swings and why buffering/smoothing is becoming a first-order design problem • Fuel, permitting, water, noise, and community acceptance as gating factors • How contracts (take-or-pay, deposits, credit support, longer PPAs) increasingly function as “infrastructure” that unlocks financing and supply chains • Whether behind-the-meter sites eventually reintegrate with the grid—or stay islanded for good If you’re tracking the industry’s shift from grid dependence to integrated onsite power, this is the conversation you want.