Dietreich Farquharson has spent more than twenty-five years in Australian electrification, moving into solar and battery storage from the mid-2000s as the technology shifted from early grid-tied inverters to today's integrated storage systems. His writing draws on that hands-on history: a preference for the measured retention curve over the marketing claim, and a habit of checking a warranty clause against its own datasheet before writing about either. He writes the Signals essays and other authored interpretation, where a personal read of the data is the point. The scores and leaderboards elsewhere on the site are the method's work, not his.
Published under this byline
Four makers on this benchmark did not stay the same company. Only one of those is an insolvency.
Seven of 122 active scorecards carry a corporate note. Behind them sit four different outcomes: a liquidation, a court-supervised restructuring, a company that left the category while still servicing claims, and a listed manufacturer reporting losses and still trading. A buyer is told to check the warranty. Nobody explains that these four are not the same risk.
The same filing that carries SolarEdge's loss also carries six quarters of margin expansion.
SolarEdge's Q1 2026 result is cited on this site as a corporate-status note, because a 25-year warranty depends on the entity behind it. Read the whole document and it also reports 46% revenue growth, a sixth consecutive quarter of margin expansion, positive operating cash flow, and guidance approaching breakeven. Both readings come from one filing.
Six architectures, one recombination chain: how silicon solar cells closed in on their physical ceiling.
Al-BSF, PERC, PERT, TOPCon, HJT, back-contact: each generation didn't invent a new idea so much as chase the same electron loss to a new location in the cell. The chain, the physics ceiling it's closing in on, and a correction we had to make to our own source material before publishing this: even a same-month technical reference had already been overtaken by a newer record.
The chemistry that won wasn't the one that launched the category.
Tesla's original Powerwall ran on NCA. A decade later, LFP owns the residential market almost outright, a national fire crisis rewrote the standards nobody had written yet, and the battery management system quietly became the product's real safety layer. Four patterns from ten years of residential battery history, and what they predict for the next one.
Mono-silicon's 2030 mix: three scenarios, and six roadmap calls the data already overtook.
TOPCon, back-contact, HJT and tandem's shares of the 2030 module mix are pulled from ITRPV, CPIA, Exawatt and NREL roadmaps, not invented. The catch: those roadmaps keep revising themselves, sometimes within a single year. Why the model runs three scenarios instead of one number, and which 2025-vintage calls are already wrong.
One name, three companies: SunPower's long unravelling, and its unlikely survival.
SunPower is no longer one company. It is three: a US installer with no factory, a Singapore manufacturer under judicial management holding the actual IBC patents, and a TCL-owned brand selling outside the US. If you own SunPower- or Maxeon-branded panels, including in Australia, your warranty sits with one specific entity among those three, and working out which one now needs your attention.
The electrification shift, projected: three scenarios to 2035.
Not our model. Three named scenario tracks, IEA and BloombergNEF's own Low, Base and High cases, run out to 2035 for EVs, grid storage, solar capacity and electricity demand. The spread is wide (52 to 73% global EV sales share by 2035 depending on which track you believe), and the risk register underneath it, mineral supply, sodium-ion commercialisation, AI demand competing for the same grid capacity, US policy reversal, matters as much as the headline numbers.
The warranty says 0.25% a year. The desert measured 8.7% in three.
The HJT-versus-TOPCon story is usually told as an architecture race: records, the 2030 mix, the tandem endgame. This is the narrower finding underneath it: independent field and lab data show both n-type families degrading in ways their warranties don't price, and the variable that predicts it is bill-of-materials and manufacturing discipline, not the passivation chemistry the marketing argues about.
The glass beneath the glass: why 'dual glass' stopped describing one thing.
Dual-glass construction genuinely outperforms glass-backsheet on moisture, thermal cycling and cell protection, when it's well executed. But the label now spans everything from SolarWatt's 2mm-plus-2mm POE format with a 26-year field record to cost-reduced 1.6mm EVA-encapsulated modules with no comparable data. Four variables (glass thickness, glass treatment, encapsulant chemistry and edge sealing) decide whether the format delivers over 30 years, and the marketing word for all of them is the same.
The degradation trap: what your battery warranty isn't telling you.
An end-point warranty lets a battery lose most of its usable capacity in years one to five, then stabilise, and the manufacturer still complies. Here's which home batteries actually protect you across the full term.
The global electrification shift: a decade of cause and consequence, 2015 to 2025.
In 2015 the global EV fleet was 1.2 million cars, grid-scale battery storage barely registered as a category, and solar supplied a rounding error of world electricity. A decade on: one in four new cars sold is electric, storage additions crossed the 100 GW mark in four years (solar took eight, wind fifteen), and battery pack prices fell 93%. What actually caused that, decade by decade, region by region, and why it's the demand context every panel, battery and inverter on this benchmark was built into.
The Panasonic EverVolt: fifty years of HJT, and the structural economics that ended them.
Panasonic discontinued EverVolt solar and battery storage in April 2025, ending a heterojunction programme that ran, under two company names, for fifty years. The technology was not the problem: HIT delivered a world-record temperature coefficient and, in 2016, a world-record module efficiency. What ended it was an expired patent moat, two failed strategic pivots, and boutique-scale manufacturing competing against Chinese continental volume. Why review.solar carries a flag on this product, and the warranty position for owners of an installed system.
Two 400 W panels, one hot roof, 14 watts apart: what temperature coefficient predicts that efficiency alone doesn't.
Every module datasheet lists a power rating measured at 25°C. Real cells run 30 to 50°C hotter, and how much power that costs is written into one figure most buyers skip past: temperature coefficient of Pmax. What it measures, why HJT cells respond to heat differently at a physical level, what independent PVEL testing and published field studies show across PERC, TOPCon and HJT, and the checks worth making on your own datasheet before the technology label does the deciding.