On a napkin
Frequency response when a generator trips Hz 60.0 59.5 trip nadir 59.6 Hz inertia + primary FR ~0–10 s regulation (AGC) seconds–min spinning + non-spin reserves 10–30 min −5s +30 min Stacked service products that catch the frequency drop Inertia spinning mass automatic, ~0s Regulation (AGC) 4–6 sec dispatch batteries dominate Spinning reserve ≤10 min response online generation Non-spin reserve 10–30 min start offline + ready Each layer catches a faster timescale than the nextElectricity cannot be stored at scale in the wires themselves. At every instant, total generation must equal total consumption plus losses, or the system frequency drifts away from 60 Hz — and equipment starts to trip. Grid operators maintain this balance through a layered system of services operating at different timescales: inertia in seconds, primary frequency response and regulation in seconds-to-minutes, spinning and non-spinning reserves in tens of minutes. As inverter-based wind, solar, and battery resources displace synchronous generators, the underlying mechanics of how this balance is maintained are changing fundamentally.
Frequency is the universal signal. Above 60 Hz means too much supply. Below means too little. Sustained deviation of even a fraction of a hertz tells operators that something is wrong somewhere on the grid — and triggers automatic responses long before humans can react. Frequency control is the most fundamental service the grid provides, and it works the same way across the Eastern Interconnection, Western Interconnection, ERCOT, and every other synchronized grid in the world (just at 50 Hz in much of the world outside North America).
In an interconnected AC grid, every connected generator must rotate at synchronized speed. In North America the standard is 60 cycles per second — 60 Hz — corresponding to generators rotating at 3,600 RPM (for 2-pole machines) or 1,800 RPM (for 4-pole machines). When load increases or generation decreases, the rotating generators effectively slow down as their kinetic energy is drawn upon to meet the imbalance, and frequency drops. When generation exceeds load, frequency rises. The relationship is essentially instantaneous and identical across the entire interconnection — frequency in Boston, Toronto, and Florida are all the same at the same moment in the Eastern Interconnection.
This makes frequency the ideal signal for distributed control. Every generator can independently measure local frequency and respond — speeding up when frequency drops, slowing down when it rises. Governors on synchronous generators do this automatically through droop control. The result is that the entire grid balances itself, second by second, without requiring centralized coordination of every change.
How fast frequency drops in response to a sudden imbalance depends on system inertia — the total kinetic energy stored in rotating mass connected to the grid. A 1,000 MW generator trip in a high-inertia system might cause frequency to drop 0.2 Hz over several seconds. The same trip in a low-inertia system might drop frequency 0.5 Hz in under a second, exceeding under-frequency relay settings and causing cascading trips. The metric is ROCOF (rate of change of frequency), measured in Hz per second.
Synchronous generators — coal, nuclear, gas turbines, hydro — provide inertia automatically because their rotating masses are physically coupled to the grid. Inverter-based resources (wind, solar, batteries) connect through power electronics that decouple the generation from grid frequency. They produce no inherent inertia. As the resource mix shifts, total system inertia declines, ROCOF for a given disturbance increases, and the grid becomes more sensitive to sudden imbalances. Some grids — particularly Ireland and the UK — are already operating with minimum-inertia constraints that limit how much wind and solar can be online at any moment without conventional plants keeping inertia available.
Modern grids procure layered reserve products through ancillary services markets. Primary frequency response operates in the first seconds after a disturbance — automatic governor response on synchronous generators, fast frequency response from batteries. Regulation reserves operate continuously through Automatic Generation Control (AGC), with generators receiving dispatch signals every 4-6 seconds to correct small ongoing imbalances. Spinning reserves are generators online and producing but with capacity available to ramp up within 10 minutes. Non-spinning reserves are offline resources that can start within 10-30 minutes.
The economics matter for commercial buyers. Each ancillary service has its own market clearing price, and these prices spike during scarcity — a generator trip, extreme weather, or unusually high renewable variability can drive regulation prices from $5/MWh to $500+/MWh in minutes. Resources with the right capabilities — particularly battery storage, where fast response is a structural advantage — earn substantial revenue in these markets. In CAISO and ERCOT, batteries now provide more than half of regulation up service. The cost flows through to capacity charges, ancillary service uplift, and ultimately to retail rates.
Three concurrent shifts are reshaping how the grid stays balanced. First, declining synchronous inertia as coal and nuclear retire faster than gas plants are added. Second, growing demand variability from data centers (highly stable load) and EVs (variable but increasingly managed). Third, the rise of inverter-based generation that requires different stability control. The grid is adapting through grid-forming inverters (which can mimic synchronous behavior), synthetic inertia services from batteries, new fast frequency response products in markets (ERCOT's Responsive Reserve Service, ISO-NE's Energy Imbalance market), and stricter interconnection standards (IEEE 2800 for inverter-based resources).
For commercial and industrial buyers, the most relevant implications are in ancillary service market design and grid reliability standards. Capacity prices, ancillary service charges, and reliability event frequency are all sensitive to how successfully the grid manages this transition. Markets where the transition is happening smoothly (CAISO with batteries, ERCOT with diverse resource mix) tend to maintain reliability with manageable cost increases. Markets where the transition is rushed (rapid retirements without inertia replacement) tend to see higher reliability event frequency and rising scarcity costs.
Common questions
Related reading on The Outlet