Gearing / leverage

What it is (precise)

A utility-scale BESS is funded from a capital stack: senior debt at the bottom, sometimes mezzanine or subordinated debt above it, and sponsor equity on top. Gearing expresses the debt portion as a fraction of total funding — a debt-to-equity split (70:30), a loan-to-cost percentage, or a pure ratio, where 70% gearing equals 2.33x debt-to-equity. 'Leverage' is the same idea framed as a multiplier on equity returns. Watch the convention: 70% of total cost and a 0.7x debt-to-equity ratio describe very different capital structures, and mixing them is the classic modelling error.

Lenders are paid first from project cash flows and accept a lower, fixed return; equity is residual and absorbs the volatility. Spreading the same profit over a smaller equity base lifts the equity Internal Rate of Return, provided the unlevered return exceeds the cost of debt.

The arithmetic is simple: a project earning roughly 8% unlevered, funded 70:30 with debt costing 6%, delivers an equity return in the low teens, because equity pockets the two-point spread on every borrowed dollar. Flip the sign — unlevered return below the debt cost — and the same mechanism accelerates losses and stretches the Payback period.

Gearing is also the weighting inside the Weighted Average Cost of Capital, which blends the after-tax cost of debt and the cost of equity in proportion to the capital structure. That blended figure often feeds the Discount rate used for Net Present Value in a Discounted Cash Flow model. Change one gearing assumption and you move the discount rate, the NPV, and the equity return at once — which is why the debt-sizing tab sits at the heart of every project-finance model, and why a learner should treat it as the first cell to interrogate.

Why it matters in a real grid-scale project

BESS revenues — energy arbitrage, frequency response, capacity payments, and other ancillary services — are more volatile and less contracted than a solar PPA. Lenders price that merchant exposure conservatively, so achievable gearing is set by how much stable, bankable cash flow the project can demonstrate.

A 200 MW / 800 MWh asset with a long-tenor tolling agreement supports far more debt than an equivalent merchant-only asset chasing arbitrage spreads. On identical hardware, that difference in contracting can move the equity return by several points — the contract, not the battery, sets the leverage.

Gearing also drives engineering and contracting decisions. The lender's independent engineer scrutinises round-trip efficiency, the warranted capacity-retention (degradation) curve, augmentation strategy, PCS and grid-code compliance, and safety design — including UL 9540 system certification, UL 9540A fire-propagation test data, and NFPA 855 installation compliance — because every assumption that erodes throughput or availability erodes the cash flow that services the debt.

This is the concrete link learners miss: a weaker technical package does not just cost a point of efficiency on a datasheet, it costs gearing, and so costs equity return.

The binding constraint is usually the Debt Service Coverage Ratio, not the headline percentage. Over-gearing leaves no headroom for degradation surprises, augmentation capex, or a soft revenue year, and a DSCR breach can trip a cash sweep, lock up distributions to equity, or push the project into default and lender step-in. Sponsors therefore trade equity return against resilience: the maximum debt the model supports is almost never the debt a prudent sponsor actually takes, and the gap between those two numbers is where experienced developers live.

Typical values and standards

Senior gearing on contracted utility-scale storage commonly lands around 60-80% of project cost; fully tolled assets with strong offtakers reach the top of that range or slightly beyond in competitive markets. Merchant or partly-merchant assets are geared lower, often 50-65%, reflecting revenue risk.

Lenders size debt to a minimum DSCR — often about 1.20-1.40x for contracted cash flows, and materially higher, frequently 1.7-2.0x or more, for merchant revenue lines — tested on a P50 / P90 revenue case rather than the sponsor's upside. Memorise those two bands; they anchor almost every storage debt conversation.

Debt structure varies by market. In the US, 5-7 year mini-perm loans with a refinancing assumption are common for storage, while fully amortising debt of 15 years or more generally requires a matching contracted revenue term. Amortisation is often sculpted to the revenue profile, and degradation matters here: usable energy declines over life, so late-year cash flows carry less coverage unless augmentation is planned and funded. Treat every figure above as an indicative range that shifts with market, contract, the interest-rate environment, and lender appetite.

Standards enter through due diligence rather than the loan agreement itself. Independent engineers and insurers look for UL 9540 certification of the ESS product, UL 9540A fire-propagation test data, NFPA 855 installation compliance, and cell-level certifications such as IEC 62619 or UL 1973.

Weak or missing safety documentation raises insurance premiums and deductibles, and insurance terms feed straight into the lender's risk view — so the safety file measurably affects the cost and quantity of debt. LFP's higher thermal-runaway onset and cycle life are one reason it is generally viewed as more bankable than NMC for stationary storage.

How it shows up in specs, studies and contracts

In the financial model, gearing lives in the debt-sizing module: target DSCR, tenor, interest rate, and the resulting maximum debt. In the term sheet and credit agreement it reappears as covenants — minimum DSCR thresholds for distributions (lock-up tests), cash-sweep triggers, and reserve accounts for debt service and augmentation.

Ask any model three questions: what DSCR was the debt sized to, on which revenue case (P50 or P90), and how much coverage headroom survives the worst contiguous years — which, for a battery, are usually the late years when degradation bites. Thin answers on any of the three flag an over-geared case.

Engineers meet gearing indirectly but constantly. The warranty capacity-retention table, the availability guarantee, round-trip-efficiency test protocols, and liquidated damages for underperformance are all negotiated with one eye on the lender's technical advisor.

Two checks pay off: confirm the warranted degradation curve matches the cycling the revenue model assumes — a warranty written for 365 cycles per year does not protect a merchant asset dispatching 500 — and confirm how augmentation is funded, from reserves, future equity, or refinancing. An unfunded augmentation plan is a hidden increase in leverage that no covenant flags until coverage has already thinned.

Common pitfalls

The most common numerical slip is mixing conventions: '70:30' means debt is 70% of total cost, a debt-to-equity ratio of 2.33x, not 0.7x. A related trap is comparing the equity IRR of a highly geared case against the unlevered IRR of another and declaring one project better — leverage changes the risk equity bears, so returns are only comparable at like gearing, or on an unlevered basis. When someone quotes a headline equity return, the first questions are always: at what gearing, and against which DSCR?

The second trap is treating maximum bankable debt as free money. Every extra point of gearing shrinks the buffer between P50 cash flow and the covenant, and BESS-specific risks — faster-than-warranted degradation, an ancillary-market price collapse, delayed augmentation — land on that buffer. Storage markets have repeatedly shown revenue lines halve within two or three years as competitors saturate a service, so a capital structure sized to yesterday's frequency-response prices is how batteries end up in distribution lock-up.