Grid-Forming Inverters Market (2026-2036)

In volume terms, shipments are expected to reach 577.0 GW by 2036 from an estimated 43.2 GW in 2026 and 33.8 GW in 2025, registering a CAGR of 29.6% during the forecast period, with the average selling price declining from USD 62.0 per kW in 2026 to USD 38.0 per kW by 2036 as grid-forming control moves from a specified option to a default firmware capability on volume power conversion platforms.

The growth of this market is mainly driven by a published code specification that defines grid-forming capability as the ability of a plant to behave similarly to a conventional synchronous generator and sets the power and fault current responses required of it, by reliability standards being developed for inverter-based resources across data sharing, model validation, planning studies and performance requirements with the final milestone falling inside the forecast period, by the retirement of synchronous generation that removes the inertia and system strength the network previously received without procuring it, by battery storage that is being built at scale and can take grid-forming control at marginal additional cost, and by islanded and microgrid applications in which no grid-following inverter can operate at all. A grid-following inverter needs a voltage waveform to synchronise to, so as synchronous plant leaves the system the question becomes which device establishes that waveform.

However, factors such as a code specification that is non-mandatory and whose intended short-term operational stability market has not yet been developed, which leaves capability procured without a revenue stream attached to it, model validation and compliance demonstration that is slower and costlier for grid-forming control because the behaviour cannot be represented in the models network operators have historically accepted, a current-carrying and thermal margin requirement that raises the cost of hardware able to behave as a voltage source during a fault, and interoperability uncertainty where multiple grid-forming devices from different suppliers share a network and no supplier will warrant collective behaviour are expected to restrain the growth of this market to some extent. Capability that is specified but not paid for is installed only where it is cheaper than the alternative, and that is the commercial position in most markets today.

A grid-forming inverter establishes and regulates the voltage waveform of the network it connects to, rather than measuring an existing waveform and injecting current in step with it, which is what a grid-following inverter does. The distinction is defined in code for the first time by Grid Code modification GC0137, on which the Authority decision was published on 31 January 2022 with implementation on 14 February 2022, which introduces a high-level minimum specification for GB Grid Forming capability, formerly described as Virtual Synchronous Machine capability, defining that capability as the ability of a plant to behave similarly to a conventional synchronous generator and so to help maintain electricity system stability, and setting minimum technical, compliance and data requirements with detailed guidance to be provided in a separate GB Grid Forming Best Practice Guide. Defining the capability as behaving similarly to a conventional synchronous generator is what makes it procurable, because a network operator can specify a behaviour rather than a technology, and it is the reason the market is described in terms of services delivered rather than in terms of inverter architecture.

The commercial weakness of that specification is as important as its existence. Specifically, the modification defines the types of power including reactive power and the fault current responses required without prescribing a minimum magnitude for them, the specification is non-mandatory for generators, and it is intended to underpin a future short-term operational stability market in which plant holding the capability could be procured and paid according to that capability, a market that has not yet been developed. Capability that is defined, specified and non-mandatory, with no settled market to pay for it, is purchased only where it costs less than the alternative or where a connection cannot be obtained without it, and both of those conditions are met in a growing number of places rather than everywhere. This is why deployment is concentrated in weak-grid locations, in islanded systems and in jurisdictions where a network operator has made the capability a connection condition, rather than distributed evenly across new capacity.

The United States has approached the same exposure through reliability standards rather than through a capability specification. Under Order No. 901 of the Federal Energy Regulatory Commission, issued on 19 October 2023, which directs the North American Electric Reliability Corporation to develop new or modified reliability standards addressing the risks presented by inverter-based resources, covering both utility-scale resources and behind-the-meter and distributed energy resources, across four subject areas comprising data sharing, model validation, planning and operational studies, and performance requirements including ride-through, momentary cessation and ramp rates, and that order set a sequence of milestones under which an initial work plan was filed on 17 January 2024, standards on disturbance monitoring data sharing and post-event performance validation were filed on 4 November 2024, standards on data sharing and model validation were due on 4 November 2025 and standards on planning and operational studies are due on 4 November 2026, which places the final milestone inside the forecast period of this report. The order does not require grid-forming control, and reading it as though it does would misstate the position, but it establishes the data, modelling, study and performance regime that any inverter-based resource must demonstrate compliance within, and the model validation and post-event performance validation requirements fall hardest on control schemes whose behaviour is not well represented in existing models. Grid-forming control is in that category, so the order raises the compliance burden on it while also creating the measurement framework through which its benefit can eventually be credited.

The physical driver behind all of this is the retirement of synchronous generation. A spinning generator supplies inertia, short-circuit current and a stiff voltage reference as inherent properties of its construction, at no incremental cost and without being asked, and a system that replaces it with inverters loses all three at once. System strength and inertia then have to be procured deliberately, through synchronous condensers, through contracts with remaining thermal plant, or through inverters capable of providing the same behaviour. The last option is the cheapest where an inverter is being installed anyway, which is why storage and renewable projects have become the vehicle for a service the network used to receive for free.

Battery storage is the natural host for the capability and the reason the market exists at scale. A storage plant already contains a four-quadrant power conversion system with headroom in both directions, sized for full output at any state of charge, and adding grid-forming control is largely a matter of firmware, controller bandwidth and current-carrying margin rather than of new hardware. The incremental cost is therefore small relative to the project, while the service delivered is the one the network is short of, and that asymmetry explains why grid-forming penetration among new storage projects has risen far faster than among solar or wind projects where the power source itself is intermittent.

Procurement differs from conventional inverter buying because the specification is behavioural and the proof is a study rather than a datasheet. A developer must demonstrate through electromagnetic transient simulation that the plant behaves as specified under fault, phase jump and islanding conditions, using a model the network operator accepts, and the model, the study and the commissioning tests frequently cost more and take longer than the price difference between a grid-forming and a grid-following inverter. Suppliers that hold validated models and prior approvals with a given network operator therefore enjoy an advantage unrelated to hardware performance, and that advantage is one of the strongest barriers to entry in this market.

The premium for grid-forming control over a comparable grid-following unit has narrowed considerably and continues to narrow, because the differentiating content is control software, controller computing capacity and a degree of current and thermal margin rather than a different power stage. Suppliers have accordingly moved from charging a feature premium toward including the capability across a platform and charging for the engineering, modelling and compliance support around it, which converts hardware margin into services revenue and makes the reported average selling price decline faster than the value delivered to the customer.

Interoperability remains unresolved and is a genuine constraint rather than a transitional inconvenience. Multiple grid-forming devices from different suppliers sharing the same network interact through their control loops, and the collective behaviour is not the sum of the individual behaviours, yet no supplier will warrant how its control scheme behaves alongside a competitor scheme it cannot inspect. Network operators therefore carry the system-level risk, which they manage by limiting how much grid-forming capacity they will connect in one area, by requiring area-wide studies before approval, and in some cases by preferring a single supplier across a region, each of which slows deployment.

Microgrids and islanded systems are the one application where the capability is not optional. An islanded network has no external voltage waveform to synchronise to, so at least one device must establish it, and a system built entirely of grid-following inverters cannot start or run without a diesel generator present to form the grid. Replacing that generator is a direct and quantifiable saving in fuel, maintenance and emissions, which gives this application a business case independent of any grid service market, and it is the reason remote and islanded deployment preceded utility-scale deployment by several years.

Geographic distribution follows the pace at which synchronous generation has left individual systems rather than the size of those systems. Asia-Pacific holds the largest share on the scale of Chinese storage deployment and on Australian system strength and inertia requirements, which have made the National Electricity Market the most demanding grid-forming environment in the world relative to its size. Europe follows on the British code specification, on Irish and Nordic system non-synchronous penetration limits and on island systems. North America is substantial but weighted toward Texas and California, and the Middle East & Africa grows fastest on large new-build renewable and storage systems in Saudi Arabia and the United Arab Emirates and on weak and islanded African networks.