Solar Scaling Losses Shape the Perovskite and 210mm Panel Race

The watts race has a hidden cost

For most of the past decade, the solar industry sold progress as a single number: watts per panel. That number has risen sharply. Trina Solar says its first 210mm module in 2020 helped push module power from 500W to 600W and then to 700W.7 Lab tandem modules have gone further, with Trina reporting an 829W perovskite-silicon module at 30.6% efficiency, verified by Fraunhofer ISE.22 Much of that headline power, however, comes from making panels physically bigger rather than making each square centimetre convert more sunlight. Engineers know the difference well, but buyers often miss it.

The idea that bigger is not automatically better applies to two very different technologies right now. In crystalline silicon, larger wafers bring electrical and mechanical penalties that the industry has had to engineer around. In perovskites, the problem is harsher: record efficiencies on fingernail-sized cells drop steeply once the material is spread across a commercial-size panel. Recent research suggests both gaps are narrowing. Neither has closed.

How silicon got big

Some observers noticed the shift as early as 2019. Peak module power was rising not mainly because cells were improving, but because modules were getting larger.23 Wafers grew from the old 156mm standard to two competing formats, 182mm and 210mm.24 The jump was dramatic. Annual cell-efficiency gains usually add 10W to 15W per module, while moving from M6 (166mm) to M10 or M12 wafers can add 100W to 150W.

The market followed. In 2023, TrendForce estimated that large-format modules would make up about 93% of module production capacity, with 210mm products alone at about 70%.4 By December 2024, the industry had shipped a cumulative 380GW of 210mm modules. Trina accounted for more than 170GW of that.7

The main reason was economics, not physics. One modelling study of a German utility plant found that M10 and M12 modules cut balance-of-system costs by roughly 5% to 6% compared with M6. Most of the savings came from shipping and installing fewer modules, and from needing less substructure and cabling.21 Trina has argued that a 210mm string design can mean 62 fewer piles and a kilometre less cable per megawatt.27

Where efficiency leaks out

The physics pushes back. A larger cell produces more current, and resistive losses rise with the square of the current. A cell-to-module analysis covering cells from 156.75mm up to 210mm found that larger cells did raise module power by as much as 77%. Over the same range, though, the cell-to-module power ratio fell by about five percentage points.25 For full, uncut M12 cells, interconnection losses reached around 7.4% of total module power. Those losses outweighed the gain from a larger active area and actually lowered module efficiency.25 The authors' conclusion was blunt: large full cells should be avoided.25

The fix was to cut the cells. Splitting cells into halves or thirds divides the current, reduces resistive and thermal losses, and turns the size penalty into a gain.2429 With split cells, module efficiency rose with cell size by up to 1.1 percentage points absolute, and M12 designs delivered the highest power density of any layout studied.25 Size only helped once the electrical design changed to handle it.

Heat is also a factor. The same study found that modules with smaller cells ran up to 2.8K cooler, and split-cell modules ran cooler than full-cell ones.25 Field modelling points the same way. In the German plant study, M12 modules ran at higher operating current, lost more to ohmic losses, and produced less energy per installed kilowatt than M10 modules, which run at currents similar to M6.21

Disagreement over the payoff

The two sides do not agree on how much of this reaches the bottom line. TrendForce, in research Trina has promoted, said 600W+ modules raise efficiency by 0.3 to 0.5 points over 500W+ products and can lower levelized cost of energy by up to 4.1%.1 An engineering contractor speaking to pv magazine in 2021 saw it differently. Large-format modules had not delivered higher module efficiency, it said, and longer trackers required wider row spacing that could cancel out the extra wattage.26

Mechanical risk is another point of dispute. One industry analysis notes that manufacturers often make glass and frames thinner to keep weight down on modules with 20% to 30% more surface area. The result is a weaker panel that is more prone to flexing, cell cracks and hotspots in high wind or heavy snow. Suppliers selling 700W+ products counter that dual-glass construction and better cell-cutting methods deal with those fragility concerns.3 Location matters too. The German study found M10 saved more on balance-of-system costs where freight is expensive, as in Europe, while M12 came out ahead where freight is cheap, as in China.21

The split-cell research and the plant-level modelling support a clear reading. Larger silicon formats are a genuine engineering win for utility-scale projects, but the gains come from reworking the module's electrical design and from logistics, not from size itself. A 700W figure on a datasheet says little about how well a panel converts light per square metre. As one UK trade body pointed out, wattage used to be a reliable proxy for cell quality only because every panel was the same size.29

Perovskites face a steeper version

If silicon's scaling penalty can be managed, perovskites face a much harsher one. A 2023 review estimated that perovskite efficiency falls with area at about 1.4×10⁻² percentage points per cm². That is roughly two orders of magnitude worse than crystalline silicon, and well above other thin-film technologies.17 The causes are well known. Spin-coating, the standard lab method, leaves uneven film thickness and more defects when scaled up.16 Larger panels also add sheet resistance in the transparent electrode, dead zones from laser scribing, and film non-uniformity.19 On larger substrates, uneven drying of the precursor solvent distorts how crystal grains grow.15

Recent figures show the size of the gap. Small tandem cells have reached 34.85%, a LONGi record certified by NREL.18 Oxford PV's certified 60-cell module reached 26.9%, which one analysis places 2 to 8 points below cell-level results.18 Reports differ on details. One puts LONGi's record in 2024 and another in April 2025.1814 Estimates of what perovskite products will actually deliver also vary widely, from 18–22% for single-junction modules and 22–26% for early tandems12 to 24–29% for commercial modules in 2026.19 The more cautious estimates fit better with reports that only a handful of perovskite lines are running at tens-of-megawatts scale.12

The scaling gap starts to narrow

The strongest clean-energy results this year are about reducing scaling losses rather than setting cell records. In March, Chinese Academy of Sciences researchers reported a crystal-seeding method combined with slot-die coating, an industrial-style deposition process. It produced a 49.91 cm² mini-module at 23.15% efficiency, losing less than 3% compared with small cells.20 In August, a team led by Nanjing University and including Renshine Solar and Three Gorges-affiliated partners reported a certified 24.0% on a perovskite module with an 810 cm² aperture.15 They tailored the film's surface chemistry before passivation, which matters because ammonium-halide passivators are sensitive to moisture and spread unevenly at large scale.15 Other work has used non-destructive X-ray diffraction mapping to find lattice distortions across large modules, reaching 22.75% efficiency.15

These results are important because they target what actually holds perovskites back. Analysts describe scale-up, yield and reliability, not peak efficiency, as the real bottlenecks.12 Stability is still an open question. The longest publicly reported operating data is around 1,000 hours, against warranties that run 25 years.18 Silicon is also still improving. TOPCon and HJT modules are expected to reach 26–27%, which would shrink the efficiency lead tandems need to justify their extra cost.18

What to watch

The lesson across both technologies is the same. The useful measure of progress is how much efficiency survives the move from a small cell to a full module, not total wattage or a record set on a square centimetre. Silicon solved its version of the problem by splitting cells and reworking module design, and it now has to manage the mechanical side effects of very large panels. Perovskites are at an earlier stage, and this year's results suggest scaling losses can be engineered down rather than simply accepted. The breakthrough worth watching is a full-size module that keeps its cell-level efficiency and holds it for decades, not the next 800W press release.