For a century, grid stability was a free by-product of how we made electricity: big spinning turbines resisted frequency changes with their physical inertia and fed fault current into short circuits without being asked. Retire the coal and gas plants, and those services leave with them. That’s the problem grid-forming inverters exist to solve — and batteries are the machines delivering it.

For a visual companion, explore the interactive Grid-Forming Droop diagram on BESS.Engineer.

Grid-following vs grid-forming

Almost every inverter on the grid today is grid-following: it measures the grid’s voltage waveform and injects current synchronized to it. It’s a very fast follower — but a follower. If the grid gets weak or disturbed, a follower has nothing to lock onto.

A grid-forming (GFM) inverter behaves as a voltage source: it establishes its own voltage waveform — magnitude and frequency set locally — and holds it steady through disturbances. AEMO’s working definition captures it well: a grid-forming inverter maintains a constant internal voltage phasor over short timeframes, which is exactly what a synchronous machine’s spinning mass does electrically. The result is that the inverter responds instantly and inherently to grid events, before any control loop has measured anything.

What that buys the grid

  • Synthetic inertia. When frequency starts to fall, a GFM battery injects power immediately — resisting the rate of change of frequency the way rotating mass does.
  • System strength / short-circuit level. During a fault, GFM inverters contribute fault current that helps protection relays see and clear the fault, and keeps voltage waveforms coherent in weak parts of the network.
  • Voltage stability in weak grids. Renewable-rich, remote areas often have low system strength; GFM plants let more inverter-based generation connect where grid-following plants would become unstable.
  • Restoration potential. A voltage source can, in principle, energize a dead network — which is why black-start from batteries is an active area of trials.

Proof on the grid: Great Britain

The flagship is Blackhillock in Scotland — Zenobē’s 300 MW / 600 MWh site (200 MW live in March 2025, the remaining 100 MW following in 2026). It was the first transmission-connected battery in the world contracted to deliver stability services, procured under NESO’s Stability Pathfinder Phase 2: ten contracts awarded in 2022 worth £323 million in total — five synchronous condensers and five grid-forming batteries — to secure 11.55 GVA of short-circuit level in Scotland and 6.75 GVA·s of inertia for the GB system. Blackhillock also went through the first compliance process under GB’s grid-forming grid code provisions (GC0137).

One number from that project tells you where the engineering frontier is: per IEEE Spectrum’s reporting, the site’s inverters were programmed to deliver a fault-current pulse of roughly 2.5 times nominal current for 140 milliseconds to meet NESO’s requirement. Power electronics can overload — but only briefly and only as far as their silicon allows, which is the fundamental difference from a synchronous machine that can deliver several times rated current from sheer physics.

Proof on the grid: Australia

Australia is the world’s deepest grid-forming market. As of AEMO’s 2025 system security planning, ten grid-forming BESS were operating in the National Electricity Market with about 1,070 MW combined, and the development pipeline held 94 projects. AEMO’s 2026 Integrated System Plan reports grid-forming inverters featuring in 74% of the NEM battery storage pipeline, and explicitly acknowledges that batteries and synchronous condensers can now perform many stability services previously delivered by coal plants. ARENA seeded this market deliberately, committing AU$176 million across eight large grid-forming batteries totalling 2.0 GW / 4.2 GWh, and operating examples now span the market — Neoen’s Western Downs stage one in Queensland (270 MW / 540 MWh) among them.

The landmark asset is the Waratah Super Battery in New South Wales: 850 MW / 1,680 MWh at a former coal plant site, operating as a System Integrity Protection Scheme — a contracted 700 MW / 1,400 MWh “shock absorber” that lets the network run closer to its limits because the battery stands ready to absorb the consequences of a sudden fault.

The honest caveats

Grid-forming is real, but three caveats belong in any serious conversation:

  1. Fault current is bounded. Typical GFM continuous overload headroom is only tens of percent, though a purpose-built asset like Blackhillock can deliver a brief pulse of ~2–2.5× rated for a fraction of a second — still short of a synchronous machine, which supplies ~5–7× rated current from its low subtransient reactance, several times more and sustained by physics. Advisory work for Australian network operator Transgrid rated heavy reliance on GFM inverters for short-circuit current as high risk, and the operator’s procurement paired grid-forming batteries with synchronous condensers. Hybrid portfolios are the current state of wisdom.
  2. Headroom costs money. Inertial response requires the battery to hold energy and power headroom — capacity you can’t sell into the energy and ancillary markets. Stability contracts have to pay for what they reserve.
  3. Standards are young. AEMO’s grid-forming specification is voluntary, with a two-tier structure of core and additional capabilities, and pre-commissioning test procedures (including hardware-in-the-loop testing) are still being standardized. Expect specifications to keep hardening through 2026–2027.

What to put in your spec

If you’re procuring a GFM battery today: define the overload capability (current magnitude and duration), the voltage and frequency response envelope, the energy headroom policy, protection coordination studies with the network operator, and factory plus site testing that actually exercises grid-forming behavior — not just grid-following compliance with a firmware flag.

FAQ

Is grid-forming a hardware or software feature? Mostly control software on capable power electronics — but genuine GFM performance depends on hardware overload ratings, DC-side energy availability, and validated controls. A firmware label alone is not a capability.

Does grid-forming replace synchronous condensers? Partially. Batteries earn market revenue while providing stability, which idle condensers can’t; but condensers still deliver stronger fault current. Operators are deploying both.

Can a battery black-start a grid? Technically demonstrated in trials and increasingly contracted; a grid-forming inverter can energize a dead network section and other plants can synchronize to it.

The controls, the physics, and the market products behind all of this get full modules in my Grid-Scale BESS: Complete Guide.