Methodology

Routing the end of a first life

A storage asset does not have an end-of-life date; it has a sequence of decisions, each of which forecloses the others. This is how we decide between them: what is measured, what is scored, what is priced, what is excluded outright, and what we refuse to claim.

One principle

Every route is expressed as the same quantity: net value in US$ per kWh of nameplate, realised at the retirement decision. Not payback, not internal rate of return, not a sustainability score. One number, four times, comparable.

The reason is uncomfortable but simple: these four options are usually argued by four different people with four different metrics. The operations team argues years of service, the sustainability team argues tonnes of CO₂, the finance team argues capex deferral and the procurement team argues scrap value. Until they are converted into one unit, no one is disagreeing about the same thing.

Suitability — whether the route is technically defensible at all — is scored separately and never blended into the money. A route can be worth more and still be the wrong answer. When the two rankings disagree, we say so and hand the decision back.

The four valuations

  • Keep in service. Each additional year of operation avoids one year's share of the replacement capex. Value = extra years × (capex ÷ design life of the replacement).
  • Refurbish. The same, plus the life recovered by replacing the modules that are dragging the pack down, less the refurbishment cost.
  • Second life. Resale of the remaining capacity — state of health × resale price per kWh — less repurposing and logistics.
  • Material recovery. The recycler's gate price less logistics. Negative when the recycler charges, which for LFP is the normal case today.

The inputs to all four are visible and editable in the Route Grader. If you disagree with a price, change it — the model has no attachment to our defaults.

The mechanism

Each route has a window, and the windows barely overlap

This is the whole argument in one chart. State of health falls along a modelled trajectory; each route is available only inside a band of that trajectory; and the bands are narrower than the interval between the decisions people actually make.

Route availability over the modelled health curve SAMPLE
Illustrative asset4-hour LFP, ~350 EFC/yr
Modelled, not measuredcalendar + cycling ageing
Bands are conventions80 / 70 / 55 % are defaults you can move

Above ~80 %

The asset is still good at its first job. Repurposing here destroys value: you sell capacity that was still earning at first-life rates in order to buy capacity you already had.

80 % → 55 %

The window. Modules are worth substantially more repurposed than recycled, because a second-life buyer is buying usable energy and a recycler is buying mass. This is the only interval in which the difference is large.

Below ~55 %

No integrator recovers the cost of testing, sorting, re-certifying and warranting the pack. What remains is the metal — and, for LFP, frequently a disposal cost rather than a payment.

The thresholds are conventions, not physics — 80, 70 and 55 % are defaults drawn from where second-life practice currently clusters, and they move with resale prices and repurposing costs. What does not move is the shape: a bounded window, with value falling off on both sides.

What actually decides

Four routes, and what each one is really asking for

State of health gets the attention, but it is rarely the binding constraint. In most real assets the answer is decided by cell-to-cell spread, by resistance rise, or by whether anyone kept the records.

Route 01

Keep in service

Re-rate the duty cycle — shallower cycles, tighter state-of-charge limits, lower C-rate — and run the asset longer at reduced output.

  • Binding constraint: usually resistance rise, not capacity
  • What it earns: deferred replacement capex, year by year
  • Hidden cost: the revenue given up by cycling gently
  • Fails when: the pack can no longer hold its rated power
Route 02

Refurbish

Replace the worst decile of modules and rebalance. A pack delivers what its weakest modules deliver, so the cheapest capacity available is often the capacity you already own but cannot reach.

  • Binding constraint: module-level data — without it you cannot target
  • What it earns: years of life recovered per unit spent
  • Best case: high spread with good average health
  • Fails when: fade is uniform across the whole pack
Route 03

Second life

Repurpose into an application that asks less: backup, commercial and industrial peak shaving, EV charging buffers, off-grid and telecom.

  • Binding constraint: traceability, then spread
  • What it earns: resale of remaining capacity
  • Regulatory gate: the EU passport from February 2027
  • Fails when: nobody can warrant what is in the crate
Route 04

Material recovery

Into the regulated waste chain. Always available, which is exactly why it becomes the default when nobody priced the other three in time.

  • Binding constraint: chemistry — no nickel, no economics
  • What it earns: metal, minus the process
  • Often negative: for LFP the recycler charges
  • Never fails: it is the floor, not the goal

Score your own asset →

Before any scoring

What we exclude outright

Weighted scores are good at ranking options and bad at refusing them. Four conditions remove routes from consideration entirely, before a single weight is applied — and an excluded route is shown as blocked, not quietly down-weighted into fourth place.

This matters because the failure mode of a scoring model is a plausible middle answer. A pack with a thermal event on record will still score respectably on capacity, age and cycling. It should not be offered for resale on any score.

Thermal event on record

All reuse routes excluded. No third-party market, no insurer, no defensible resale — regardless of how well the remaining modules test.

State of health below 50 %

Below the floor at which the fixed cost of testing, sorting and re-certifying a pack can be recovered from what is left of it.

Resistance more than doubled versus beginning of life

Capacity may look acceptable while the pack can no longer deliver its rated power. Energy without power is not a product.

Cell-to-cell spread above 15 %

Refurbishment or recycling only. Selling a pack this unbalanced does not solve the problem; it transfers it to someone with less information than you had.

The artefact

A route recommendation nobody can verify
is worth what an opinion is worth

Every party downstream of the decision — a second-life buyer, a collector, a customs officer, an EPR scheme, an auditor — needs to check the claim without trusting the party that made it. So the output is not a PDF with a logo on it.

Signed and hash-verifiable

Reports ship as evidence packages under an open manifest format, signed with Ed25519. Anyone can check the signature and the file hashes offline, with the public key we publish — no account and no call to us.

Reproducible end to end

The engine is deterministic and seeded: the same configuration produces the same numbers. A third party who disagrees with an assumption can change it and re-run rather than argue about the output.

Passport-shaped

The fields a battery passport has to carry — chemistry, state of health, lifecycle history, recycled content, due diligence — are the fields the record already holds. We are the independent source, not the registrar.

Verify a sample evidence package →

The clock

Lifecycle data stops being optional

Dates and percentages below are taken from the legal texts and official summaries, not from interpretation. Where a rule is still a draft, it says so.

18 February 2027

EU battery passport becomes mandatory

Every EV battery, light-means-of-transport battery and industrial battery above 2 kWh placed on the EU market must carry a digital passport reachable by QR code, covering chemistry, carbon footprint, recycled content, state of health and supply-chain due diligence. There is no grace period for units already in the production pipeline.

Regulation (EU) 2023/1542 verified
End of 2025 → end of 2030

Overall recycling efficiency

65 % of the total mass of lithium-based batteries must be recycled by the end of 2025, rising from the end of 2030. This is a mass target, not a metal target — and it is decided by the fractions with no market value.

Regulation (EU) 2023/1542 verified
End of 2027

First material recovery targets

50 % of contained lithium and 90 % of contained cobalt, copper, nickel and lead must be recovered from waste batteries.

Regulation (EU) 2023/1542 verified
End of 2031

The targets that bite

Lithium recovery rises to 80 %; cobalt, copper, nickel and lead to 95 %. From 18 August 2031, new batteries must contain minimum recycled content: 16 % cobalt, 6 % lithium, 6 % nickel and 85 % lead — rising again in 2036.

Regulation (EU) 2023/1542 verified
Chile · draft in consultation

Ley REP 20.920 — collection and valorisation targets for batteries

Chile's extended producer responsibility framework lists batteries among its priority products. The supreme decree that sets the actual collection and valorisation targets is still a draft in public consultation, with lithium phased in more slowly than lead to allow treatment capacity to be built. The environmental assessment guidance already places batteries in BESS projects inside the law's scope, and BESS operators are most likely to qualify as generators of priority-product waste.

Ley 20.920 · Ministerio del Medio Ambiente verified
Chile · already in force

Hazardous waste and reporting

Independently of the pending decree, handlers of waste batteries operate under the hazardous-waste rules (DS 148/2003), maintain technical records and report to the national emissions and waste register.

DS 148/2003 · RETC verified
Ongoing

Basel Convention

Waste lithium-ion batteries crossing a border are a controlled transboundary movement requiring prior consent. The condition declaration is what decides whether a consignment is waste or a functioning product — which makes an independent grade a shipping document as much as a commercial one.

Basel Convention · national implementations verified
What it adds up to

The same missing artefact, three times

A second-life sale, a cross-border shipment and a producer-responsibility declaration all fail on the same thing: an independent, dated, tamper-evident statement of condition. Building it while the asset runs costs almost nothing. Reconstructing it afterwards is usually impossible.

Our reading, not legal advice
Our standard

What we will not claim

The circular economy attracts round numbers and confident percentages. Here is where our model stops.

We do not publish a single accuracy figure

Every health and remaining-life output is a p05–p95 band whose width depends on the data behind it. A vendor quoting one accuracy number for degradation forecasting across chemistries, climates and duty cycles is quoting a marketing figure.

We do not certify batteries

We produce an independent, verifiable grade. Certification, homologation and placing a battery on the market are obligations that sit with accredited bodies and with the party doing the placing. Conflating the two would be convenient and wrong.

We do not claim recycled-content compliance

Recycled content is a supply-chain attribute of a new cell, evidenced through the manufacturer's chain of custody. What we can evidence is what came out of your asset, and with what confidence.

We do not sell the routes we recommend

No equity in asset owners, no operated assets, no trading desk, no hardware, no collection, no recycling, no brokerage. If we earned a margin on the recycling route, our recommendation to take it would be worth nothing.

Bring us the asset you are least sure about.

One asset approaching the end of its first life, anywhere in the world. We will come back with all four routes scored and priced, the exclusions that applied, and what your data would sharpen.

Request an assessment