← All Signals
Home batteries · July 2026

Battery rubric v2.3: three structural changes and the evidence behind them.

The Solar Analytica battery rubric version 2.3.0 is effective 1 July 2026. No pillar weights changed. Three structural changes: grounded in peer-reviewed research and confirmed operational data: update the cell chemistry bands, the EV charging integration bands, and the round-trip efficiency criteria.

Overview

Three changes. All evidence-driven.

Rubric changes at Solar Analytica are triggered by material evidence shifts, not by manufacturer announcements, not by trade show claims, not by category marketing. The v2.3 changes follow a targeted research session (29 June 2026) that commissioned three parallel research streams: V2G regulatory and operational status in Australia; HTW Berlin / AQUU Research testing methodology; and the commercial and electrochemical status of LMFP and sodium-ion chemistries as of mid-2026.

All three streams returned actionable findings that clear the bar for rubric revision. The changes below explain what changed, what evidence drove each decision, and what was deliberately left unchanged.

Change 1: Cell chemistry

LMFP and sodium-ion are not the same technology.

In rubric v2.2.0, lithium manganese iron phosphate (LMFP) and sodium-ion (Na-ion) shared a single band at 85 points. This grouping was provisional: it acknowledged both as post-LFP alternatives without enough differentiated field data to separate them. That provisional grouping is no longer defensible.

LMFP moves to 90 points. LMFP (LiMnxFe(1-x)PO4, typically x = 0.6) is a commercial lithium chemistry. Chak et al. (Journal of The Electrochemical Society, 2025) measured directly matched pouch cells: LiMn0.6Fe0.4PO4/graphite versus LiFePO4/graphite, and found:

PropertyLMFPLFP
Gravimetric energy density169 Wh/kg154 Wh/kg
Volumetric energy density360 Wh/L355 Wh/L
Thermal onset temperature125°C140°C

Source: Chak et al., JES 2025 (DOI: 10.1149/1945-7111/ae08f2). Direct measurements in matched pouch cells, not manufacturer claims.

LMFP's edge over LFP is about 10% GRAVIMETRIC ENERGY DENSITY · MATCHED POUCH CELLSLMFP's edge over LFP is about 10% 154 Wh/kg LFP 169 Wh/kg LMFP
Directly matched pouch cells, LiMn0.6Fe0.4PO4/graphite vs LiFePO4/graphite. Volumetric density and thermal onset are in the table above - the gravimetric gain doesn't carry over to thermal stability. Source: Chak et al., JES 2025.
Chart data
SeriesValueNote
LFP154 Wh/kg
LMFP169 Wh/kg

The ~10% gravimetric advantage is peer-reviewed and real. LMFP earns the higher placement. But the same paper records a meaningful thermal stability disadvantage: LMFP's self-heating onset temperature is 125°C versus LFP's 140°C. That 15°C gap is not trivial in Western Australian or Queensland conditions, where battery enclosures in non-climate-controlled garages or utility rooms routinely exceed 40°C ambient in summer.

More significant is what temperature does to LMFP cycle life. Argonne National Laboratory tested LMFP/graphite pouch cells at 40°C: the most relevant ambient condition for Australian residential storage:

ElectrolyteRetentionCycles at 40°C
Standard carbonate~80%165
+ VC additive~87%390
+ VC + DTD additives~85%610
LFP (reference)~80%6,000+

Source: Zuo/Xu et al., EES Batteries Jan 2026 (DOI: 10.1039/D5EB00228A). C/3 rate, 40°C. LFP figure is typical residential rated life.

Cycle life collapses, even with additives LMFP CYCLE LIFE AT 40°C · BY ELECTROLYTECycle life collapses, even with additives 165 cycles Standard carbonate 390 cycles + VC additive 610 cycles + VC + DTD 6,000+ LFP (reference)
Cycles to ~80-87% retention at 40°C, C/3 rate. The gap is the point: even LMFP's best additive package doesn't close a tenfold difference against LFP at realistic Australian summer ambient. Source: Zuo/Xu et al., EES Batteries, Jan 2026.
Chart data
SeriesValueNote
Standard carbonate165 cycles
+ VC additive390 cycles
+ VC + DTD610 cycles
LFP (reference)6,000+

The mechanism is Mn dissolution via Jahn-Teller distortion at the Mn3+/Mn4+ transition, confirmed by Argonne (JES, December 2025), and the dissolution rate doubles without electrolyte additives. This is an intrinsic feature of Mn-containing cathodes, not a solvable manufacturing defect.

The rubric handles this correctly: the Cell chemistry criterion (weight 0.28 within Safety & Chemistry) now scores LMFP at 90: 10 points below LFP, reflecting the thermal onset gap and the elevated-temperature cycle life penalty. The separate Longevity pillar's cycle life criterion will also score LMFP products lower when cycle life evidence is available. A product using LMFP cells faces a double discrimination in the rubric. That is intentional, because the risks compound in practice. LMFP's commercial home is EVs: where higher gravimetric density matters and cycle requirements are lower (~800–1,500 cycles over vehicle life). No BNEF Tier-1-listed residential home battery product uses LMFP cells as of mid-2026: Tier-1 is a bankability/financing listing, not a quality rating, so this reflects manufacturer scale rather than a performance judgement.

Sodium-ion moves to 75 points. The Na-ion picture is more nuanced. Utility-scale BESS Na-ion cells have achieved remarkable cycle life: CATL TENER at 15,000 cycles, Hithium HiF at 20,000 cycles, HiNa at 10,000 cycles. These figures exceed LFP's residential cycle life claims by a factor of two or three. Na-ion is also structurally superior in thermal stability, no Mn dissolution, no Jahn-Teller distortion, no lithium-metal plating risk.

But those headline numbers are from large-format prismatic cells operating in controlled BESS enclosures. The residential Na-ion products available as of mid-2026: Biwatt PowerNest, Salzstrom Natrium Home, Eleven Energy: carry manufacturer-published cycle life claims of approximately 3,000 cycles, none of which are independently verified. That is comparable to lower-end LMFP compositions. Na-ion's lower cell-level energy density (~95–175 Wh/kg) is also a genuine disadvantage at residential scale, where system size is constrained by cost and physical footprint.

The utility-scale numbers haven't reached homes yet NA-ION CYCLE LIFE · UTILITY VS RESIDENTIALThe utility-scale numbers haven't reached homes yet 10,000 HiNa (utility) 15,000 CATL TENER (utility) 20,000 Hithium HiF(utility) ~3,000 Residential Na-ion unverified, manufacturer-published
Utility-scale figures are from large-format prismatic cells in controlled BESS enclosures - not yet demonstrated at residential scale. Residential Na-ion products: Biwatt PowerNest, Salzstrom Natrium Home, Eleven Energy.
Chart data
SeriesValueNote
HiNa (utility)10,000
CATL TENER (utility)15,000
Hithium HiF (utility)20,000
Residential Na-ion~3,000unverified, manufacturer-published

At 75 points, sodium-ion sits above NMC (60 pts) and below LMFP (90 pts) and LFP (100 pts). The score reflects its thermal and materials advantages while applying appropriate conservatism to the residential product reality. If a major manufacturer launches a validated Na-ion home battery with demonstrated cycle life at residential scale, this band will be revisited.

Change 2: EV charging integration

V2G is operationally live in Australia.

The EV charging integration criterion previously ran four bands. The gap between 80 points (smart OCPP V1G) and 100 points (V2G dispatch confirmed in AU) was always a live question: as V2G moved from regulatory framework to operational reality, the space between "hardware that could do V2G" and "hardware that has done V2G in Australia" was going to need its own tier.

That tier is now warranted. Amber Electric's May–June 2026 trial confirmed 655.5 kWh of V2G grid export from 33 vehicles over 30 days in the National Electricity Market. V2G is commercially operational in Australia at small scale. At the same time, operational V2G has clarified a new distinction: products with V2G-capable hardware and a confirmed DNSP connection pathway are meaningfully different from products that merely support OCPP 2.0.1 for charging.

PointsBand definition
100V2G bidirectional dispatch: AU confirmed: live AU VPP/trial with confirmed grid export, CEC-listed or DNSP-approved charger, EV manufacturer warranty backing
90V2G-ready hardware: CEC-approved bidirectional charger + DNSP connection pathway confirmed, OCPP 2.0.1/2.1 or ISO 15118-20, AU V2G commercial product not yet off-the-shelf
80Smart OCPP V1G EV charging: solar surplus tracking, OCPP 1.6 or 2.0.1 (charge-only), paired or native integration
60Basic EV charger scheduling (timer / relay)
35No EV integration

The regulatory infrastructure for these distinctions exists. AS/NZS 4777.2 Amendment 2 (mandatory 23 August 2025) created the bidirectional charging standards pathway. The AEMC's Unlocking CER Benefits rule (August 2024) and Integrating Price-Responsive Resources rule (December 2024) cleared the legal framework for V2G NEM dispatch. From 1 July 2026, new V2G chargers must meet AS 5438 / OCPP 2.1 demand-response requirements.

Product mapping as of 1 July 2026: 100 pts: Sigenergy SigenStor (CEC-approved; confirmed AMBER energy-trading integration; first residential system to achieve AU V2G dispatch classification under the v2.3 definition). 90 pts: V2Grid Numbat (first CEC-certified standalone bidirectional charger in Australia, Ausgrid-approved November 2025, CHAdeMO + CCS2); StarCharge Halo (OCPP 2.0.1/2.1, ISO 15118-2/20, AS 4777.2 certified, used in Amber Electric trial). 80 pts: GoodWe HCA G3; Fox ESS A Lite (both OCPP 2.0.1, smart V1G solar-surplus charging confirmed, no AU V2G dispatch as of this update).

Change 3: Round-trip efficiency

HTW Berlin RSP/SPI is now the preferred efficiency source.

HTW Berlin's Stromspeicher-Inspektion study (conducted annually by HTW Berlin and AQUU Research) tests residential battery storage systems at a standardised protocol using the Relative System Performance (RSP) and System Performance Index (SPI) metrics. These are system-level measurements: they capture actual inverter, wiring, and management losses in a way that manufacturer-stated round-trip efficiency figures structurally cannot.

Battery rubric v2.3 now recognises HTW Berlin RSP/SPI results as: (1) the preferred data source for round-trip efficiency, superseding manufacturer-stated RTE where a result is available for the specific product and configuration; and (2) equivalent to DNV/PVEL/RETC testing for the purposes of the verified reliability criterion: the same tier as the established third-party testing bodies Solar Analytica already recognises for panel assessment.

The trigger was the Fox ESS PQ3-H3-Ultra achieving a 97% SPI result at Smarter E Europe 2026. A 97% RSP maps directly to the 100-point band in the rubric's round-trip efficiency criterion (≥96% RTE). For products with an RSP result below 96%, the score maps to the band the RSP figure falls within. HTW Berlin's study is predominantly European in scope; where a tested product's Australian configuration differs meaningfully, the EU result should be applied with a qualifier. A Solar Analytica methodology note covering the HTW Berlin testing protocol and current AU product coverage is scheduled for publication in July 2026.

What didn't change

Pillar weights and criterion weights are unchanged.

PillarWeight
Capacity & Usability0.11
Power & Efficiency0.16
Longevity0.22
Safety & Chemistry0.21
Energy Ecosystem0.15
Guarantee & Backbone0.15
How the rubric weighs its six pillars BATTERY RUBRIC V2.3 · PILLAR WEIGHTSHow the rubric weighs its six pillars 11% Capacity &Usability 16% Power &Efficiency 22% Longevity 21% Safety &Chemistry 15% EnergyEcosystem 15% Guarantee &Backbone
Pillar weights sum to 100%, unchanged in v2.3. Longevity carries the most weight of any single pillar. Full criterion-level breakdown: review.solar/methodology.
Chart data
SeriesValueNote
Capacity & Usability11%
Power & Efficiency16%
Longevity22%
Safety & Chemistry21%
Energy Ecosystem15%
Guarantee & Backbone15%

The v2.3 changes operate entirely through band reclassification and criterion note updates, not through weight adjustments. Weight changes are the most significant interventions in the rubric and require a separate review cycle. The stability of the pillar weights across v2.2 → v2.3 is intentional.

v2.4 candidates

Two signals flagged for the next revision cycle.

AFCI (Arc Fault Circuit Interrupter)

Fox ESS PQ3-H3-Ultra includes AFCI as an active arc-fault detection layer at the electrical connection level: the same failure mode responsible for the 2025 Sigenergy SigenStor EC ACCC recall. AFCI is not currently differentiated in the Thermal & fire mitigation criterion. If AFCI normalises across premium products, it warrants a discrete band between Passive thermal management (75 pts) and Active thermal management + fire suppression / cell isolation (100 pts). Proposed position: 85 pts.

IP66 with flood rating

IP66 is available on the Fox ESS PQ3-H3-Ultra and provides additional protection margin over IP65. However, IP66 without a certified flood rating provides insufficient differentiation to justify a new band. The existing 90-pt band (IP65+ outdoor-rated) is appropriate until IP66 becomes the market norm or a product achieves a certified flood-rating test result that can be cited.

Battery rubric v2.3.0 | Solar Analytica | 1 July 2026. Research basis: Rubric Shifters session 29 June 2026. Full rubric methodology: review.solar/methodology.

Sources12 references

Read our ownership, funding and editorial disclosures.

Stay across the benchmark.

New scorecards, score changes and market signals, published as they happen.