Microgrid modeling software in 2026: what changed, what matters, and which tool should you use?

Key takeaways

• The 2022 buying framework still works, but a 2026 procurement decision needs a fifth criterion: can another analyst reproduce and challenge the result?
• Public list prices changed materially: HOMER Pro now spans $1,575–$4,650 per user per year, HOMER Grid is $4,200 per year on annual billing, and Xendee DESIGN starts at $440 per user per month billed annually.
• A longer feature list does not make a tool better for every study. Renewable performance, facility optimization, storage value stacking, hybrid feasibility, network design, and live operation are different modeling boundaries.
• MicrogridModeler’s strongest 2026 claim is low-friction, auditable single-bus feasibility—not universal superiority. Historical multi-year NSRDB/user-record ingestion is now implemented with a feasible-year/P90-cost sizing gate; subhourly simulation, multi-node power flow, and several multi-energy technologies remain outside scope.
• Run a controlled bake-off with the same load, weather, costs, tariff, constraints, horizon, and acceptance tests before trusting any cross-tool ranking.

The 2022 framework was right—and incomplete

In 2022, Mayfield Renewables compared five widely used product families: NREL’s System Advisor Model (SAM), NREL’s REopt platform, EPRI’s DER-VET, HOMER Pro and HOMER Grid, and Xendee. The article organized the buying decision around price, modeling capabilities, utility tariffs and incentives, and ease of use. Those four questions remain practical because they connect software features to staff time, project economics, and the client deliverable.

What changed is the consequence of the answer. Microgrid studies now travel through more hands: developers, owners, utilities, lenders, engineers, procurement teams, and operators. A result that cannot be reconstructed from its load, resource, tariff, component assumptions, constraints, model version, and time-series outputs is difficult to review and expensive to update. So our 2026 framework adds auditability and reproducibility as a fifth buying criterion.

This is also why we reject a simple “most checkmarks wins” ranking. SAM is not trying to be an enterprise microgrid operating platform. DER-VET is not a one-click sales proposal builder. Xendee is not priced or scoped like a public research tool. A fair comparison begins by naming the decision each product is designed to support.

Sources for this section: Mayfield Renewables: 2022 microgrid modeling software comparison

Price and access in 2026

The free-versus-paid split still exists, but the old numbers should not be carried into a 2026 budget. SAM remains free desktop software. REopt remains available through a free public web tool and open model/API paths. DER-VET remains publicly accessible with desktop GUI and Python command-line options. MicrogridModeler’s complete local workflow is unlocked at no cost during development, and its published Solo plan is $199 per user per month.

HOMER Pro now publishes three standard single-user annual plans: Base at $1,575, Professional at $3,100, and Expert at $4,650. The included modules differ substantially, so “HOMER Pro price” is not one number. HOMER Grid publishes an annual plan at $350 per month, or $4,200 per year, and a monthly plan at $665 per month. Xendee DESIGN publishes Gold at $440 per user per month billed annually and Platinum at $540 per user per month billed annually; Platinum can be billed monthly for an additional $110 per month.

License cost is only the visible part of total cost. Add onboarding, training, analyst setup time, data preparation, tariff work, proposal production, review cycles, and the cost of rebuilding a project in another tool when the first model reaches its boundary. A free tool that needs three specialist days can cost more than a paid tool; a broad enterprise platform can be wasteful when a transparent feasibility screen is all the decision requires.

Sources for this section: SAM current downloads · DER-VET current release · HOMER Pro pricing · HOMER Grid pricing · Xendee DESIGN pricing

Modeling power starts with the boundary, not the feature count

“Can it model a battery?” is no longer a useful differentiator. The harder questions are how the tool represents chronology, network topology, technology coupling, degradation, uncertainty, controls, reliability, economics, and hard constraints. A model can include the same component names and still answer a different decision because its mathematical boundary is different.

For a remote clinic, the decisive evidence may be hourly state of charge, generator minimum loading, fuel use, unmet load, and the least-cost feasible PV-battery-diesel size. For a grid-connected hospital, the decisive evidence may be tariff demand charges, outage survival, CHP heat recovery, and lifecycle cost. For a campus, energy transfer among nodes, cables, transformers, thermal networks, and one-line structure may matter. For an operating fleet, forecast-driven control and live data become a separate product category.

Time resolution matters too. An hourly model can be excellent for early energy feasibility while missing 15-minute demand peaks or short controller events. A deterministic typical-year run can be reproducible while understating weather and load uncertainty. A MILP can find a formal optimum inside its formulation while still depending on simplified performance curves and perfect-forecast assumptions. The right question is not “Is it accurate?” but “What is represented, what is simplified, and what evidence proves the result inside that boundary?”

SAM in 2026: the renewable performance and finance specialist

SAM remains the strongest choice in this group when detailed renewable technology performance and project finance are the center of the study. The current release is SAM 2025.4.16 Revision 2. Its desktop workflow covers multiple renewable technologies, storage, detailed financial models, parametric studies, weather-file collections, scripts, plots, tables, and exports.

The P50/P90 distinction is especially important. A credible probability-of-exceedance result is not a decorative percentile attached to one typical year. SAM documents statistical workflows and simulations across collections of weather files. That makes it the current leader here for measured multi-year solar variability studies, provided the analyst supplies quality-controlled records and an appropriate uncertainty method.

SAM is not automatically the best microgrid optimizer. Its broad input surface and technology-specific depth can require specialist setup, and a planner may need to assemble microgrid control, resilience, tariff, and cross-technology questions carefully. Use it when PV, storage, renewable performance, and finance deserve deeper treatment than a focused site-feasibility workflow provides.

• Best fit: technology-specific renewable performance, finance, sensitivity, and measured weather-year analysis.
• Watch for: setup complexity, the exact dispatch and grid assumptions, and whether the study needs formal cross-DER capacity optimization.
• 2026 verdict: category leader for detailed renewable performance and measured solar uncertainty—not the universal microgrid winner.

Sources for this section: SAM current version · SAM parametric and statistical analysis · SAM battery storage models · SAM financial models

REopt in 2026: the open facility DER optimization reference

REopt remains the most compelling free reference when the decision is which distributed energy resources a building, campus, community, or microgrid should install and how they should dispatch to minimize lifecycle cost or pursue resilience and clean-energy goals. NREL’s current manual describes a deterministic mixed-integer linear program that selects, sizes, and dispatches candidate technologies while meeting loads at each time step.

Its integration layer is a material advantage. The web tool automatically queries the Utility Rate Database for tariffs and uses NREL resources including PVWatts, SAM, and the Wind Toolkit. The technology set extends beyond PV and batteries to wind, CHP, geothermal heat pumps, electric and thermal storage, absorption chilling, existing heating and cooling plants, and backup diesel for outage analysis.

That formal breadth does not eliminate modeling judgment. NREL states that REopt is a decision-support model rather than a procurement design and that investment decisions should not rely on its results alone. Analysts must verify utility rates, actual loads, project costs, incentives, financial inputs, interconnection, permitting, and other project facts. REopt also relies on deterministic forecasts and simplified technology formulations appropriate to its optimization problem.

• Best fit: U.S. facility and campus DER sizing, lifecycle economics, resilience, clean-energy goals, and API-backed studies.
• Watch for: tariff selection, simulated versus measured loads, formulation assumptions, and the boundary between opportunity screening and design.
• 2026 verdict: category leader for open, formal facility-DER optimization and automatic public-data integration.

Sources for this section: REopt web tool user manual · REopt public web tool

DER-VET in 2026: the open storage value-stacking specialist

DER-VET remains the strongest open tool in this set when the core question is how storage and other DERs create value across customer, reliability, and market services. EPRI’s latest public release is DER-VET Version 1.3, with GUI version 1.3.0 released December 18, 2024. It is available as a desktop application and as Python command-line source.

The reference cases show its center of gravity more clearly than a generic feature list: battery sizing for bill reduction, storage paired with PV for guaranteed outage coverage, CAISO energy and frequency-regulation value, EV-fleet charging mitigation, degradation-aware market operation, and long-duration energy shifting across U.S. ISO/RTO regions. EPRI also publishes hourly sample data structures for loads, prices, PV generation, import/export limits, ancillary services, resource adequacy, and controllable demand.

The tradeoff is analyst workload. DER-VET provides inspectable methods and unusually rich value-stream modeling, but the user is more responsible for sourcing and validating market, tariff, cost, and time-series inputs than in a maintained commercial data service. Use it when transparent storage valuation matters more than one-click onboarding or client-ready proposal automation.

• Best fit: storage valuation, stacked services, bill reduction, reliability, EV charging, market participation, research, and regulatory analysis.
• Watch for: input-data quality, market-rule interpretation, specialist setup, and the distinction between valuation and construction design.
• 2026 verdict: category leader for open storage-centric value stacking.

Sources for this section: DER-VET software and release history · DER-VET reference cases · DER-VET help and sample data

HOMER in 2026: the mature hybrid-system workbench

HOMER still deserves its incumbent status. HOMER Pro combines chronological simulation, capacity optimization, and sensitivity analysis for remote power systems, islands, and microgrids. Its current public material supports time steps from one minute to one hour and a base technology set that includes PV, wind, batteries, diesel generators, and simple grid connections. Higher plans or modules add biomass, hydro, CHP, advanced loads and grids, hydrogen, advanced storage, multi-year analysis, and MATLAB-linked dispatch.

HOMER Grid targets behind-the-meter and grid-connected economics. It includes EV charging, demand-charge reduction, resilience, sizing, sensitivity, CHP, and value stacking. Its tariff proposition remains strong: UL advertises access to more than 35,000 commercial and industrial tariffs in the United States, Canada, and Mexico, plus an advanced builder for time-of-use, seasonal, contracted-demand, and tiered rates. It can also generate customized client-facing proposals.

The buying caution is packaging. A Base HOMER Pro license and an Expert license do not provide the same modeling scope, and HOMER Grid is a separate product. Compare the exact modules, training, automation, tariff geography, and delivery workflow your team needs. HOMER’s maturity is valuable, but it should not be confused with proof that every system boundary or input is automatically correct.

• Best fit: broad hybrid-system configuration, subhourly-to-hourly simulation, sensitivity analysis, training, and established professional workflows.
• Watch for: product and module selection, license cost, desktop workflow, and preserving enough model evidence for independent review.
• 2026 verdict: category leader for mature hybrid breadth and sensitivity analysis.

Sources for this section: HOMER Pro capabilities · HOMER Pro pricing · HOMER Grid capabilities · HOMER Grid pricing

Xendee in 2026: the enterprise multi-energy and network platform

Xendee has moved furthest beyond a stand-alone feasibility model. Its platform spans portfolio discovery, proposal development, detailed design, and operation. Xendee DESIGN publishes support for more than 25 DER technologies, multi-objective and multi-year optimization, 15-minute analysis, detailed finance, incentives, mobility, multi-node design, power flow, one-line diagrams, and automatic cable and transformer sizing.

The multi-energy boundary is a genuine differentiator. Xendee documents CHP, thermal storage, chillers, waste-heat recovery, hydrogen electrolyzers and fuel cells, EV charging, export and interconnection limits, battery degradation, tariffs, demand response, resilience, and shared projects. Its OPERATE product extends the workflow into forecast-driven live optimization, which is categorically different from a planning-only simulator.

That scope comes with enterprise cost and process. A single Gold DESIGN seat is $5,280 per year at the published annual rate; Platinum is $6,480 per year before monthly-billing premiums or added services. The value can be rational for teams that need network engineering, multi-site programs, collaboration, branded reports, API access, and operations. It is harder to justify when the decision is one early, single-bus feasibility screen.

• Best fit: campus and multi-site programs, multi-energy systems, network-aware design, EV infrastructure, collaboration, proposals, and live operation.
• Watch for: total subscription and service cost, implementation scope, plan-dependent features, and whether the project really needs an enterprise platform.
• 2026 verdict: category leader for enterprise multi-energy, multi-node, and design-to-operation breadth.

Sources for this section: Xendee platform overview · Xendee DESIGN pricing and feature matrix · Xendee platform FAQ

MicrogridModeler in 2026: our focused case for category leadership

We built MicrogridModeler around a narrower proposition: another analyst should be able to open a browser, run a defensible single-bus feasibility study, see why a design passes or fails, and receive the evidence needed to reproduce the decision. The complete local workflow requires no installation or account. It runs full-year chronological dispatch across PV, wind, batteries, diesel generation, and the grid; applies explicit operating policies and hard feasibility constraints; sizes alternatives; calculates lifecycle economics; and exposes outage and component-failure resilience.

The audit trail is the product advantage. A run can export selected-window and full hourly dispatch, lifecycle cash flow, a complete JSON package, input and source provenance, an evidence grade, and a deterministic content fingerprint. The modeler imports timestamped custom loads, PVWatts modeled TMY data, complex OpenEI Utility Rate Database JSON, and complete historical weather or normalized PV-production years. The default historical fetch downloads 10 NSRDB GOES Aggregated PSM years. Missing, duplicate, and invalid hours fail; leap day is removed explicitly; no gap is silently interpolated.

The historical record changes sizing, not only reporting. Every fixed design and promoted optimizer finalist is re-dispatched against every exact year. The user sets a minimum feasible-year probability, and passing finalists are ranked by 90th-percentile high lifecycle cost. The export preserves the complete resource series, fill flags, evidence grade, year outcomes, threshold, and optimizer evidence. A 40-year seeded distribution remains only as an explicitly synthetic fallback when no historical record is attached.

That candor still defines our boundary. Historical NSRDB is satellite-derived/model data, not a ground meter, and a planning probability is not automatically bankable. MicrogridModeler does not yet equal SAM’s full loss-uncertainty and finance workflow, REopt’s formal open MILP and automatic tariff discovery, DER-VET’s storage-market depth, HOMER’s subhourly hybrid breadth, or Xendee’s multi-node power flow, thermal and hydrogen systems, collaboration, proposals, and operations. Our credible number-one claim is no-install, no-account, reproducible single-bus feasibility with unusually transparent historical risk—not “best at everything.”

• Best fit: fast, transparent PV-wind-battery-diesel-grid feasibility, sizing, economics, and resilience that must survive technical review.
• Current lead: the lowest-friction complete audit package in the focused single-bus planning category.
• Current gaps: a complete finance-grade loss-uncertainty budget, maintained tariff discovery and versioning, subhourly simulation, multi-node power flow, thermal, hydrogen, CHP, flexible EVs, shared cloud projects, and branded proposals.

Sources for this section: MicrogridModeler methodology and limits · MicrogridModeler live comparison scorecard

The honest 2026 leadership scorecard

The table below is the conclusion we would want before spending our own software budget. It does not award one trophy to unlike tools. It names the strongest current fit for each difficult decision and records why that lead exists.

Tariffs and incentives: database access is not the same as bill accuracy

Tariff convenience remains one of the largest differences among products. REopt automatically searches the Utility Rate Database and accepts custom detailed rates. HOMER Grid advertises a large maintained commercial and industrial tariff catalog and an advanced builder. Xendee includes tariff and incentive workflows, while higher tiers add custom tariff support and NEM 3.0 access. MicrogridModeler now imports full URDB JSON structures for seasonal and weekday schedules, tiered energy, monthly and time-of-use demand charges, fixed and minimum charges, and export rules. DER-VET and SAM can represent economic inputs, but the analyst generally owns more of the rate preparation.

No database selection should be treated as self-verifying. Utility eligibility, rider applicability, coincident peaks, lookbacks, ratchets, standby charges, export compensation, taxes, minimum bills, incentive deadlines, and interconnection limits can change the answer. Preserve the source, label, effective date, eligibility class, and any manual transformations. Then reconcile a baseline bill against actual invoices before using avoided cost as an investment claim.

The 2026 winner is not the tool with the most tariff names. It is the workflow that combines convenient discovery with the exact applicable rate, a versioned source, transparent schedules and tiers, and a bill-validation test. No product removes the need to confirm the tariff with the utility.

How to run a software bake-off that produces evidence instead of screenshots

A fair evaluation should use a small real project and freeze every overlapping input. Vendor demos are useful for learning the interface; they are poor substitutes for a controlled comparison because each demo can quietly change the load, resource, costs, constraints, and acceptance criteria.

Do not begin with the headline LCOE. Begin with the time-series energy balance and the assumptions that generate it. If two tools disagree, map converter treatment, storage efficiency, state-of-charge boundaries, generator fuel curves, minimum load, reserve, replacement logic, tariff schedules, degradation, escalation, and salvage before declaring that one engine is wrong.

1. Write the decision and the excluded scope in one paragraph: feasibility, investment sizing, storage valuation, network design, proposal delivery, or operation.
2. Freeze the same timestamped load, renewable resource, technology ratings, costs, tariff, financing, incentives, outage assumptions, and study horizon.
3. Define hard acceptance tests before running: annual energy balance, zero or bounded unmet load, minimum SOC, generator operating limits, bill reconciliation, and outage survival.
4. Run a baseline without new DERs and reconcile energy, demand, fuel, and utility cost before comparing optimized designs.
5. Inspect chronological dispatch at peak load, low renewable output, minimum SOC, generator starts, curtailment, outage windows, and tariff peaks.
6. Export inputs, model version, solver or search settings, time-series outputs, financial cash flow, warnings, and source dates—not only the summary report.
7. Record which product reached its boundary and what downstream engineering is still required before procurement, interconnection, protection, controls, and construction.

Our final recommendation

If renewable performance and finance are the hard part, start with SAM. If a U.S. facility needs formal DER sizing around lifecycle cost, resilience, and public data, start with REopt. If storage value stacking or market services drive the investment, start with DER-VET. If broad hybrid simulation, subhourly work, sensitivity, and an established commercial workflow matter, evaluate HOMER. If the program needs multi-node power flow, multi-energy design, portfolios, team delivery, or live operation, evaluate Xendee.

Start with MicrogridModeler when the question is a single-bus feasibility, historical-weather risk, sizing, economics, or resilience decision and the team values speed, visible constraints, and a complete reproducibility package. Then escalate the shortlisted design when finance-grade uncertainty, network, thermal, market, subhourly, protection, controls, or construction engineering becomes decision-critical.

The best microgrid software in 2026 is the smallest defensible model that answers the real decision and makes its limits impossible to miss. That is the standard we are building toward—and the standard buyers should demand from us and everyone else.

Keep exploring

Sources and review notes

This article is grounded in the cited technical sources and the stated modeling assumptions. Recheck project inputs, equipment data, and local requirements before using it for design.

FAQ