How to Add 300 Gigawatts to the Grid Without Building a Single Power Plant

Episode Summary

Executive Summary: The episode centers on Gridcare CEO Amit Narayan’s claim that the U.S. grid is not truly out of power, but out of visibility: modern AI can reveal underused capacity already in place, speeding interconnection for data centers and other loads while lowering rates. The hosts frame this as a major affordability and competitiveness breakthrough, alongside a Q&A segment on solar installer bankruptcies and hydro’s stagnation.

Main Topics: Grid utilization as the core bottleneck (Priority: 5/5): The hosts and Amit argue that grid constraints are often a visibility and planning problem, not just a physical capacity problem. AI-based analysis can identify unused headroom in existing infrastructure. Gridcare’s business model and value proposition (Priority: 5/5): Gridcare positions itself as a neutral bridge between utilities, customers, regulators, and ISOs, helping prove where capacity exists and how to unlock it without compromising reliability. AI demand, speed-to-power, and infrastructure urgency (Priority: 5/5): The conversation ties explosive AI load growth to the need for faster power solutions. Existing infrastructure optimization is presented as the fastest way to meet demand while new transmission and generation take years. Rates, affordability, and utility incentives (Priority: 4/5): A major theme is that higher utilization can spread fixed grid costs across more electrons, lowering rates for everyone and helping utilities earn sooner without relying solely on rate-based new builds. Trust, contracts, and stakeholder alignment (Priority: 4/5): Amit argues the challenge is ultimately trust: utilities, customers, regulators, and ratepayers need transparent evidence that flexible load additions are fair and reliable. Ask Jigger: residential solar bankruptcies and hydro’s challenges (Priority: 3/5): The Q&A segment explains why retrofit solar installers are fragile in high-interest environments and why hydro remains underinvested due to site exhaustion, permitting delays, and ownership friction.

Key Arguments: The grid is often underutilized because utilities plan conservatively for worst-case scenarios and lack modern visibility tools. Modern AI/compute can analyze quadrillions of grid scenarios, exposing capacity that legacy methods missed. Gridcare can unlock capacity in months rather than years because it uses software, data, and flexible demand instead of new concrete and copper. Every new gigawatt placed on existing infrastructure can lower rates roughly 5% for a typical mid-size utility, according to Gridcare’s white paper. Data centers are willing to pay for speed-to-power, especially when alternatives like new transmission, nuclear, fusion, or space-based compute are too slow. Gridcare acts as a neutral intermediary that aligns planners, operators, customers, regulators, and ISOs by quantifying reliability, flexibility, and rate impacts. Utilities have incentives to build, but today they also face real competitive pressure to attract growth and affordability backlash, making utilization gains attractive. VPPs are finally valuable at scale because data centers now have enough economic upside to justify flexibility, unlike earlier periods when VPP economics were too small. In Ask Jigger, residential solar installer bankruptcies are framed as a balance-sheet and financing cascade amplified by high rates and weak counterparties. Hydro’s lack of momentum comes from limited viable sites, slow FERC approvals, Army Corps constraints, and entrenched asset owners.

Data Points: Potential unlocked capacity: 100,000 MW to 300,000 MW - Estimated existing-grid capacity that could be unlocked with better visibility and software-based optimization. Portland General found capacity: 450 MW - Gridcare reportedly uncovered capacity in Portland General’s territory that was previously unknown. National Grid New York found capacity: 650 MW - Gridcare reportedly found additional capacity in National Grid’s New York territory. Rate reduction estimate: 5% - Gridcare claims each new gigawatt of capacity added on existing infrastructure can reduce rates for a typical mid-size utility by about 5%. Existing grid utilization: About 30% - Amit argues the grid is only about one-third utilized when examined at full granularity. AI load economics: $10M-$20M per year per megawatt - Estimated annual value of unlocked megawatt capacity for AI-related demand. Data center value example: Billions per year - A 100-megawatt data center is described as generating billions of dollars of value annually. Latest funding round: $64 million - Gridcare’s recent financing round is cited as validation of the market opportunity. Electric vehicles parked: 3 million - A sponsor spot notes three million EVs in American driveways, parked most of the day. EV parked time: 22 hours/day - Used in the Octopus Energy ad to illustrate stranded battery capacity. Scenarios in NERC-style study: 100 million - Example of two simultaneous outages across 10,000 grid elements, illustrating computational complexity. VPP deployment: 8 GW in 20 countries - Amit references prior VPP experience to show the technology has been deployed at scale. Generation queue backlog: 2.6 terawatts - Amit says a large amount of generation is waiting to connect to the grid. Current/near-term data center training load: 500 MW - Training a new AI model today is framed as roughly a 500 MW load. Predicted training load by decade end: 4,000 MW - Projected growth in model-training power demand. Example utility project count: 6 data centers - Gridcare says six flexible connection offers in Portland were all accepted and energized. Residential installer bankruptcies: 100+ installers - Jigger cites widespread distress in the retrofit solar market, including Freedom Forever, SunPower, and others.

Pivotal Quotes: "America doesn't have a power shortage. We have lack of visibility in what already exists." — Amit Narayan: Core thesis for Gridcare’s approach to unlocking capacity. "If we can do it in code, not copper and concrete, we can do it in six to 12 months, not six to 12 years." — Amit Narayan: Explains why software-based grid optimization is a faster alternative to traditional infrastructure buildout. "It's like you're kind of like looking in the couch cushions and finding a million bucks." — Jamie Nolan: Hosts’ analogy for the value of discovering hidden grid capacity.

Implications: The episode suggests a major near-term path to cheaper, faster power for AI and other loads is better use of the grid we already have. If Gridcare’s model scales, utilities could cut rates, speed interconnections, and reduce pressure for immediate mega-builds.