Efficiency numbers sell modules. They do not sell debt. For institutional lenders, insurers, and project financiers evaluating the first wave of perovskite-on-silicon tandem projects, the question was never whether tandem cells could clear 30% conversion efficiency in a lab. It is whether that performance can be underwritten with the same confidence as 25 years of crystalline silicon field data. That gap, not the cell architecture itself, is the actual bottleneck standing between tandem and utility-scale capital.

Why Bankability, Not Efficiency, Is the Real Bottleneck

Bankability is a narrower question than “does the technology work.” It asks whether a lender can size debt against a 20 to 25 year revenue stream with an acceptable probability of underperformance. Silicon earned that confidence the slow way: multiple manufacturers, multiple geographies, and multi-decade field deployments that let independent engineers build degradation models nobody seriously disputes anymore.

Perovskite-on-silicon tandem has none of that history yet. A supplier can show a Fraunhofer-certified lab record and a promising pilot line, but a lender is being asked to price a risk that has, at most, two to three years of real field exposure. That asymmetry between claimed performance and verified performance is what actually gates capital, independent of how compelling the underlying physics is.

Translating a 10% LCOE Reduction Into Debt Sizing

Oxford PV’s public claim of a 10% LCOE reduction versus standard silicon, at a commercial module efficiency of roughly 25%, is a reasonable industry reference point, but it is a vendor claim, not an underwriting input. Lenders should treat it as a hypothesis to be tested against the specific project, not a number to be plugged directly into a model.

Mechanically, the pathway from higher efficiency to improved debt-service coverage ratio (DSCR) runs through energy yield, not through the efficiency percentage itself:

• Higher energy density per panel increases annual kWh production from the same footprint, without a proportional increase in fixed costs such as land lease, interconnection fees, or O&M contracts that are typically priced per site rather than per watt.
• Lower balance-of-system cost per watt (fewer racking, wiring, and inverter components at a given target capacity) reduces the capital stack a lender needs to size debt against in the first place.
• The combined effect, if the claimed 10% LCOE reduction holds up in practice, flows through to either a higher DSCR at the same debt sizing, or headroom to size additional leverage at the lender’s existing DSCR threshold.

The critical underwriting question is not whether a 10% LCOE improvement is plausible. It is what discount a lender should apply to that figure given the technology’s limited field history, and how that discount should be reflected in the DSCR floor or the debt-to-equity ratio required to close. Most financiers evaluating early tandem projects today are applying exactly this kind of haircut rather than accepting supplier LCOE claims at face value, which is the correct posture until independent field data closes the gap.

What Insurers Actually Demand: Degradation Curves and Damp-Heat Testing

Before any insurer will back a 25-year performance warranty, they need degradation data that silicon has accumulated for decades and tandem is only beginning to generate. Two things matter most:

1. Degradation curves. Silicon modules degrade in a well-characterized, roughly linear pattern that underwriting models have priced for years. Early perovskite formulations degraded faster than silicon under sustained heat, humidity, and UV exposure. Compositional engineering and improved encapsulation have closed much of that gap in recent commercial designs, but “closed most of the gap” is not the same as “matched silicon’s 25-year track record.” Insurers price the difference as residual risk until multi-year field data says otherwise.

2. IEC 61215 and IEC 61730 certification, specifically damp-heat testing. IEC 61215’s damp-heat protocol (1,000 hours at 85 degrees Celsius and 85% relative humidity) is the standard stress test the industry uses as a proxy for long-term field durability. For silicon, decades of correlation between damp-heat results and real-world performance make this a reliable proxy. For perovskite-silicon tandem, damp-heat performance is improving quickly, but the correlation between a passed test and 25-year real-world behavior is still being established. Insurers and independent engineers should treat a passed IEC 61215/61730 certification as a necessary, not sufficient, condition for warranty backing, and should ask suppliers directly for accelerated aging data beyond the minimum certification threshold.

Until field data accumulates, expect insurers to require additional mitigants: shorter initial warranty terms with step-up provisions, degradation reserve accounts sized more conservatively than silicon precedent, or manufacturer parent-company guarantees that do not rely solely on the project-level warranty.

Actionable Guidance for Lenders Evaluating 2026/2027 Pipelines

For institutional lenders and financiers underwriting tandem projects entering the pipeline over the next 18 months, four things are worth doing before term sheets are issued:

• Ask for supplier-specific field data, not industry averages. A 10% LCOE claim from one manufacturer does not transfer to another’s product without independent verification. Request accelerated aging data and any available field deployment results specific to the exact module being financed.
• Apply a bankability discount to vendor LCOE and yield claims, sized to the supplier’s actual field track record rather than their lab record. A manufacturer with two years of commercial shipments warrants a smaller discount than one still in pilot production.
• Treat IEC 61215/61730 certification as a floor, not a finish line. Request the underlying test data and any extended damp-heat or thermal cycling results beyond the certification minimum, and have an independent engineer review it directly rather than relying on the supplier’s summary.
• Size degradation reserves conservatively relative to silicon precedent until at least three to five years of comparable field data exists for the specific technology and manufacturer in question.

Tandem’s underlying economics are compelling enough that the technology does not need to be oversold to be worth financing. The lenders who get ahead of this market will be the ones who build underwriting frameworks now that correctly price the gap between lab claims and field-proven performance, rather than waiting for that gap to close on its own.