Grid-Forming vs Grid-Following BESS: Which Should You Use?

Choosing between a grid-forming and grid-following battery energy storage system is not a matter of selecting the "more advanced" inverter mode. The two approaches perform different electrical roles, and the correct choice depends on the strength of the network, the operating modes required from the battery plant, the applicable interconnection rules, and the performance the complete plant must demonstrate.

A grid-following (GFL) BESS synchronizes to an existing grid voltage and frequency reference. A grid-forming (GFM) BESS maintains an internal voltage and frequency reference and interacts with the surrounding system from that reference.

That distinction is becoming more important as grids add large amounts of inverter-based generation and storage. But it is also frequently oversimplified. Grid forming does not automatically mean black start, unlimited fault current, lower short-circuit-ratio limits, or better economics in every project. Likewise, grid following remains appropriate for many strong-grid commercial, industrial, and utility applications.

Grid-Forming vs Grid-Following BESS at a Glance

The most useful starting question is therefore not "Which technology is better?" but "Where does the inverter get its electrical reference, and what must the plant do when that reference becomes weak or disappears?"

What Is a Grid-Following BESS?

A grid-following BESS uses its power conversion system (PCS) to synchronize with an AC waveform already present at the point of connection.

A conventional GFL controller commonly uses a phase-locked loop, or PLL, to estimate grid phase angle and frequency. The controller then regulates inverter current to deliver the required active and reactive power.

Conceptually, the sequence is:

1. The grid or another voltage-forming source establishes the AC waveform.
2. The inverter measures the waveform.
3. The controller calculates the current needed to meet the power command.
4. The PCS injects or absorbs power while remaining synchronized with the external reference.

This architecture is mature and effective when the grid provides a stable voltage reference. Typical use cases include energy arbitrage, peak shaving, demand management, renewable-energy shifting, capacity services, and many forms of frequency regulation.

The key limitation appears when the voltage reference becomes weak, highly dynamic, or disappears. A pure GFL inverter cannot independently energize a dead AC bus because there is no established waveform for it to follow.

What Is a Grid-Forming BESS?

A grid-forming BESS uses a different control philosophy. The PCS maintains an internal voltage phasor and controls its interaction with the network from that reference. At a simplified level, it behaves like a controlled voltage source behind an impedance.

AEMO's Voluntary Specification for Grid-forming Inverters distinguishes core grid-forming behavior from additional capabilities that may require broader plant functionality. That distinction is crucial: "GFM capable" on a PCS datasheet does not prove that the complete BESS can perform every GFM-related service.

Commercial implementations can use droop-based control, virtual synchronous machine or virtual synchronous generator concepts, power-synchronization approaches, or other algorithms. These control families should not be treated as interchangeable. Their behavior can differ significantly during weak-grid operation, large disturbances, current limiting, and transitions between grid-connected and islanded states.

What Grid-Forming Can Provide

Depending on the inverter and plant design, GFM operation can support:

• voltage-source behavior;
• operation in low-system-strength conditions;
• fast frequency and voltage response;
• synthetic or emulated inertial behavior;
• oscillation damping;
• islanded operation;
• system restoration support;
• stronger system-stability contribution in some high-IBR networks.

However, a GFM label does not automatically guarantee that the complete plant can:

• black-start a dead network;
• operate an island indefinitely;
• meet every TSO or utility definition of GFM;
• provide a specified level of fault current;
• retain ideal voltage-source behavior after current limits are reached;
• switch between operating modes without additional controls or testing;
• resynchronize safely after island operation;
• provide a specified amount of inertia without power or energy headroom.

This separation between inverter capability and plant-level capability is one of the most important issues in BESS procurement.

Detailed Technical Comparison

Weak-Grid Performance and Short-Circuit Ratio

Grid-following controls depend on an external voltage reference. As system strength falls, inverter current can have a larger effect on the voltage waveform the PLL is trying to track. Interactions among PLLs, control loops, network impedance, and nearby inverter-based resources can reduce stability margins.

Grid-forming controls can be advantageous because they maintain voltage-source behavior rather than relying on a stiff external waveform for fundamental synchronization.

But there is no universal short-circuit-ratio value that automatically decides between GFL and GFM. Statements such as "GFL is always stable above SCR 3" or "GFM is required below SCR 1.5" may be useful as rough examples in a particular study, but they should not be used as universal purchasing rules. System strength also depends on network impedance characteristics, operating state, nearby IBR controls, fault levels, control tuning, and the definition of SCR being used.

For a broader discussion of connection behavior, see this guide to energy storage system grid compatibility.

Frequency Response, Fast Frequency Response, and Synthetic Inertia

Grid-following BESS can provide very fast frequency response. A GFL plant measures a frequency deviation or receives a control command and changes power accordingly. This is widely used for frequency-control and balancing services.

Grid-forming control can respond differently. Power can change as a natural consequence of the inverter's internal voltage and frequency control structure, allowing the plant to provide an inertia-like response without waiting for the same measurement-and-dispatch sequence used by conventional GFL services.

These behaviors are complementary, not interchangeable. A contracted fast frequency response service is not the same thing as grid-forming inertial behavior, and the available response can still depend on state of charge, operating point, active-power headroom, converter overload capability, and control settings.

Voltage Support and Oscillation Damping

GFM technology is particularly relevant where the grid needs a resource to contribute voltage-source characteristics rather than only regulate current against an existing voltage waveform.

This can improve stability in some weak-grid or high-IBR conditions and can help damp electromechanical or control-related oscillations. The magnitude of that benefit is project-specific. It must be demonstrated in system studies rather than assumed from the control label.

Fault Current and Current Limiting

Neither GFL nor GFM turns a power-electronic converter into a synchronous machine with large, sustained fault-current capability. Semiconductor and thermal limits remain strict.

Current limiting is especially important for GFM control. During a severe voltage disturbance, the voltage-source controller may request more current than the converter can safely supply. Once the current limit becomes active, the inverter's behavior depends strongly on the selected limiting strategy.

A 2024 IEEE review hosted by NREL on overcurrent limiting in grid-forming inverters explains why current-limiting design affects fault-current contribution, voltage support, stability, and post-fault recovery.

For procurement, do not accept "GFM provides fault current" as a complete specification. Ask:

• How much current is available?
• For how long?
• Under which voltage and frequency conditions?
• How are positive- and negative-sequence currents controlled?
• What happens to voltage-source behavior when the current limit is reached?
• How is the behavior represented in the RMS and EMT models?
• How has protection coordination been verified?

Black Start, Islanding, and Grid Forming Are Not the Same Thing

Grid forming describes how the inverter establishes and controls its electrical reference.

Islanding describes the ability of a plant or local electrical system to operate after separation from the wider network.

Black start describes the ability to energize equipment or a network from a de-energized condition without relying on an already-operating external AC grid.

The U.S. Department of Energy's UNIFI Consortium distinguishes traditional GFL inverters that need an established voltage and frequency reference from GFM controls that can establish the electrical reference needed for inverter-dominated systems and restoration.

A pure GFL inverter cannot initiate black start by itself. GFM control can provide the voltage reference needed for restoration, but a complete black-start-capable BESS may also require auxiliary power, battery availability, transformer energization capability, protection and switchgear logic, plant sequencing, communications, and resynchronization systems.

Projects intended for islanded operation should evaluate the complete microgrid battery storage system design, not only the inverter control mode.

Can Grid-Forming and Grid-Following BESS Work Together?

Yes. A power system does not need every inverter to perform the same role.

Grid-forming resources can establish or strengthen an electrical reference while grid-following resources operate relative to that reference. UNIFI has explicitly discussed systems in which only a subset of inverters operates in GFM mode while other inverter-based resources remain grid-following.

There is no universal rule that a fixed percentage of inverter capacity must be grid forming. The required amount depends on network topology, system strength, operating conditions, control design, disturbance performance, and grid-code requirements.

Can a Grid-Following PCS Be Upgraded to Grid Forming?

Sometimes-but "GFM retrofit" has two very different meanings.

At the inverter level, some modern PCS platforms can change from GFL to GFM through firmware, software configuration, parameter changes, and retuning. At the plant level, however, the project may still need new studies, controller changes, protection modifications, auxiliary-power systems, metering, synchronization equipment, communications, or recommissioning.

A useful real-world example is the Western Downs Battery in Queensland. ARENA reports that the project successfully transitioned from grid-following to grid-forming operation, demonstrating that firmware upgrades and careful tuning can unlock GFM capability on suitable hardware. But that example should not be generalized into "every GFL BESS is only one software update away from GFM."

ARENA's 2026 knowledge-sharing material on the Western Downs BESS transition also illustrates a broader point: technical capability, grid-connection approval, market registration, and service enablement are separate workstreams.

Future-proofing should therefore be written into the plant specification. Ask whether the future GFM upgrade requires only a license and tuning, or whether it also requires plant-controller changes, additional sensors, new auxiliary systems, equipment oversizing, protection redesign, new studies, and recommissioning.

Does Grid-Forming BESS Cost More?

There is no universal GFM cost premium.

Some core GFM functions can be implemented largely through inverter controls on hardware that already supports the necessary current, voltage, and control performance. More demanding services can create additional costs through converter oversizing, overload requirements, power and energy headroom, protection changes, plant controls, studies, model validation, testing, and commissioning.

This is why generic claims such as "GFM costs 20% more" or "GFM costs the same as GFL" should be treated cautiously unless they refer to a defined market, project scope, hardware platform, and service requirement.

NERC's 2025 Reliability Risk Priorities report describes GFM BESS as a potentially cost-effective grid-stabilizing solution while also noting that equipment standards and system understanding are still developing. The commercial question is therefore not whether GFM has one universal premium, but whether the required grid-support capability creates enough technical or market value to justify its project-specific cost.

For a broader breakdown of storage project economics, see this BESS cost analysis.

Is Grid-Forming More Efficient Than Grid-Following?

Do not assume a material round-trip-efficiency advantage from the control label alone. PCS switching losses, transformer losses, auxiliary loads, battery efficiency, operating point, temperature, and dispatch profile normally have a larger direct effect on round-trip efficiency.

GFM services can influence economics indirectly if they require active-power headroom, reactive-power capability, state-of-charge reservations, or additional thermal margin. Compare vendor efficiency maps and the project's real duty cycle rather than treating GFM or GFL as an efficiency class.

What Changed in 2025–2026?

The most important industry development is not that grid-following suddenly became obsolete. It is that grid-forming capability is moving from research and pilot projects into connection requirements, access-standard reviews, large operating projects, and mainstream BESS procurement.

Australia: GFM Is Becoming Mainstream in the Connection Pipeline

AEMO reported in 2026 that standalone battery capacity in the National Electricity Market connection pipeline had grown to 33.2 GW, and that around 74% of battery projects in the pipeline were grid-forming.

This is a strong market signal, but it should be interpreted correctly. It is an Australia NEM connection-pipeline statistic, not a global share of operating BESS.

AEMO is also conducting a dedicated Grid-forming Technology Access Standards Technical Requirements Review. As of August 2026, AEMO's published timetable shows a draft report planned for Q3 2026, a final report in Q4 2026, and a final rule-change proposal to the AEMC in Q1 2027.

This means it is inaccurate to reduce the Australian position to a simple statement that "GFM is universally mandatory." The direction is clearly toward broader GFM deployment and more formal technical requirements, but exact obligations still depend on project type, connection process, and applicable rules.

United States: MISO Has Moved from Discussion to Interconnection Procedures

MISO's current generation interconnection process includes specific GFM control requirements for BESS. Its BPM-015 Section 5.3.7 requires applicable interconnection customers to submit PSCAD simulation test results for BESS grid-forming control procedures by the start of Decision Point 2.

Project teams in MISO should therefore check the current BPM-015 Generation Interconnection requirements and the specific scope applying to their queue cycle and project configuration rather than relying on older headlines describing the policy only as a proposal.

North America: NERC Continues to Encourage GFM BESS

NERC's 2025 ERO Reliability Risk Priorities Report states that grid-forming technology in BESS can support system stability and that its use should be encouraged, while also emphasizing that GFM equipment standards remain under active development and that planning and operating models need to evolve with the technology.

That combination is important: the industry direction is favorable to GFM, but it is not evidence that every GFM implementation is mature, interchangeable, or suitable without detailed studies.

Europe: Technical Requirements Are Advancing, but "EU GFM Mandate from 2026" Is Too Simple

ENTSO-E published its Phase II Technical Report on Grid Forming Requirements in November 2025. The report provides non-binding technical guidance linked to the draft NC RfG 2.0 and its proposed GFM requirements for non-synchronous generation and electricity storage modules.

The important procurement message is that European GFM requirements are becoming more formal and detailed, but developers should not treat "EU grid forming rule from 2026" as one universal, already-finalized requirement for every BESS project. The applicable national rules, connection category, implementation timeline, and system-operator requirements still need to be checked.

What the 2025 Spain-Portugal Blackout Does-and Does Not-Prove

The April 2025 Iberian blackout is frequently used online as evidence that too many grid-following inverters caused the event and that every new BESS should therefore be grid forming. That conclusion is not supported by the final official investigation.

ENTSO-E's final report published on 20 March 2026 concluded that the blackout resulted from a combination of interacting factors, including oscillations, gaps in voltage and reactive-power control, differences in voltage-regulation practices, rapid output reductions and generator disconnections in Spain, and uneven stabilisation capabilities.

The engineering lesson is not "GFL caused the blackout." It is that voltage control, stabilizing capability, inverter behavior, generator behavior, operating practices, protection, and system-level coordination all matter in high-IBR networks.

This distinction is worth including because it replaces a simple narrative with a current, verifiable conclusion from the final investigation.

Real Project Evidence:

Western Downs: GFL-to-GFM Retrofit Is Possible

Western Downs demonstrates that a suitably designed BESS can transition from grid-following to grid-forming operation using existing inverter hardware with firmware changes and careful tuning. That is valuable evidence for future-proof procurement.

It does not prove that every PCS platform can be converted the same way, or that plant-level GFM functions such as black start can be added without other changes.

Broken Hill: Grid-Forming Is a Connection and System-Strength Question

ARENA's 2026 knowledge-sharing material for the 50 MW / 50 MWh Broken Hill BESS focuses on system-strength modelling, differences between GFM and GFL behavior, grid-connection approval, and how expected performance will be verified through studies, commissioning tests, and operational monitoring.

This is the more useful model for buyers: treat GFM as demonstrated plant performance, not as a checkbox on an inverter datasheet.

How to Choose Between Grid-Forming and Grid-Following BESS

Step 1: Check the Grid Code and Interconnection Process

Start with the requirements at the actual point of connection. Determine whether the TSO, ISO, RTO, DSO, utility, or network service provider requires specific voltage-source behavior, GFM controls, EMT models, weak-grid performance, current-injection behavior, inertia, or commissioning evidence.

Do not copy a requirement from Australia, MISO, or Europe into another market without checking the local rules.

Step 2: Define the Required Operating Modes

Ask what the BESS must physically do:

• operate only while connected to a healthy grid;
• support a facility during a utility outage;
• establish and sustain an island;
• black-start a dead bus;
• transition between grid-connected and islanded modes;
• resynchronize with the utility network;
• support restoration of other generation or loads.

A strong-grid peak-shaving project and a remote mining microgrid may use similar battery hardware but require very different control architectures.

Step 3: Assess System Strength

Evaluate SCR together with network impedance, fault level, nearby IBRs, minimum synchronous-generation conditions, planned network changes, and the operating scenarios that create the weakest grid.

If system-strength concerns exist, require the OEM and power-system consultant to demonstrate stable performance using models appropriate to the connection study.

Step 4: Define the Grid-Support Services

Specify the actual performance needed, such as voltage support, frequency response, synthetic inertia, oscillation damping, ride-through, island operation, restoration, or system-strength contribution.

Then determine which control mode and plant configuration can provide those services. Do not buy a control label first and define the requirement afterward.

Step 5: Verify Plant-Level Capability

The inverter is only one element of the BESS architecture. Confirm the plant controller, auxiliary power, protection, instrumentation, communications, transformers, switchgear, and energy-management functions needed for the required operating modes.

For utility-scale systems, the relevant question is whether the complete utility-scale BESS solution can demonstrate the required response at the point of connection.

Step 6: Validate Models and Studies

Depending on the project, this may include:

• RMS dynamic studies;
• EMT studies;
• weak-grid sensitivity studies;
• fault ride-through studies;
• controller interaction studies;
• power-hardware-in-the-loop or controller-hardware-in-the-loop testing;
• model validation against hardware;
• commissioning tests.

For GFM projects, studies should represent realistic current limits and mode transitions instead of modeling the converter as an unlimited ideal voltage source.

Step 7: Evaluate Commercial Value

Compare the cost of the required GFM functionality with the value it creates. This can include avoided system-strength equipment, improved connection feasibility, contracted stability services, reduced future retrofit risk, or black-start and resilience value.

For a strong-grid project with no GFM requirement and no stability-service revenue, GFL may remain the more economical and simpler choice.

Typical Project Scenarios

Strong-Grid Commercial and Industrial BESS

A factory, warehouse, or data center using storage for peak shaving, tariff optimization, energy arbitrage, and conventional grid-connected backup may have no technical reason to require GFM if the network is strong and island formation is not part of the design.

A commercial and industrial energy storage system should be optimized for the site's actual operating objectives rather than selected on the basis of a fashionable control label.

Remote or Islanded Microgrid

If solar, diesel, and batteries must operate without a utility grid, at least one source must establish local voltage and frequency. A system consisting only of pure grid-following resources cannot establish that reference by itself.

A GFM source-often the BESS-therefore becomes a fundamental part of the architecture.

Utility-Scale BESS on a Weak Grid

At a weak point of interconnection, GFM may materially improve stability or connection feasibility. The decision should come from the network studies, grid code, inverter performance, and expected operating range, not from a generic SCR threshold.

Future-Proof Utility Project

If grid-code direction suggests that GFM functionality may be needed during the plant's life, a PCS with a credible upgrade path can reduce retrofit risk. The procurement package should define exactly what "upgradeable" means and what plant-level changes would still be required.

Conclusion

The key difference between grid-forming and grid-following BESS is not that one is new and the other is old. They perform different electrical roles.

A grid-following BESS uses an existing grid waveform as its fundamental reference and regulates current relative to that waveform. A grid-forming BESS maintains an internal voltage reference and can provide voltage-source behavior that becomes particularly valuable in weak grids, islanded systems, restoration applications, and high-IBR networks.

The industry direction in 2025–2026 is clearly toward wider GFM deployment, more formal connection requirements, and larger real-world projects. AEMO's pipeline data, MISO's interconnection procedures, ENTSO-E's developing technical framework, and NERC's reliability guidance all point in that direction.

But the engineering decision still has to be project-specific. Do not choose GFM because of a headline, a universal SCR threshold, or a generic cost claim. Define the operating modes, understand the point of connection, check the current grid code, verify the complete plant design, and require validated models and tests.

For buyers, the most useful question is no longer "Is this PCS grid forming?" It is: "What grid-forming performance can this complete BESS demonstrate under the conditions our project must survive and the services it must provide?"