Explained, Technology
Solar PV
Solar photovoltaics is the simplest machine on this platform, it has no moving parts, and the most consequential: 2,383 gigawatts installed, half a terawatt added in a single year, and the cheapest electricity most of the world has ever built. Its simplicity is the point. A panel is a semiconductor sandwich that turns photons into direct current, silently, for thirty years, and everything interesting about solar happens either in the factory that made it or in the electronics that connect it to the grid.
The system
A utility-scale solar plant is four layers: the modules that generate, the structure that aims them, the inverters that make the power grid-shaped, and the collection system that carries it away. The panel gets the photographs. The inverter does the work.
Crystalline silicon cells, today overwhelmingly n-type TOPCon at 22 to 24 percent efficiency, laminated behind glass, increasingly glass on both sides so the rear face harvests reflected light. A modern utility module is rated around 600 to 700 watts and warranted to lose only a fraction of a percent of output per year for 25 to 30 years.
Fixed-tilt racks are cheapest; single-axis trackers, which follow the sun east to west through the day, lift energy yield by 15 to 25 percent on good sites and now dominate utility builds in sunny markets. The tracker is also the plant's main moving part, which makes it the plant's main mechanical failure source.
Power electronics converting DC to grid-frequency AC, either as central units the size of a shipping container or as string inverters distributed through the array. The inverter sets the plant's grid behaviour: voltage support, fault ride-through, and increasingly grid-forming operation, holding frequency the way spinning generators once did. It is the most intelligent, most connected, and most replaced component on site.
DC cabling to the inverters, medium-voltage AC to the onsite substation, then the grid. Plants are routinely built with a DC capacity 20 to 30 percent above the inverter rating, deliberately overplanting so the plant holds its full export level for more of the day and clips the rare peaks.
Why it got so cheap
Solar's price collapse is a manufacturing story, not a physics one. The technology follows a learning curve, every doubling of cumulative production has cut costs by around a fifth, and China industrialised the entire chain, polysilicon to module, at a scale that pushed factory-gate prices below ten US cents a watt in 2025, frequently below the cost of making them. The result is the strangest market structure in energy: the module, the part that actually generates, is now often among the cheaper line items in a finished project, undercut by land, labour, steel and the grid connection.
The same story has a second edge. Manufacturing is concentrated in one country to a degree unmatched by any other energy technology, which is why solar sits at the centre of the trade-policy argument: US tariff walls holding module prices near three times the Chinese level, European dependence debates, and export controls moving in both directions. And the era of ever-falling prices has paused, China's removal of export VAT rebates and consolidation among loss-making manufacturers pushed prices upward through 2026. Cheap is not the same as stable.
What limits it
The panel's weakness is the sun's schedule. Solar output is zero at night, low in winter at high latitudes, and maximal at exactly the same hour for every plant on the grid. That correlation is the technology's defining system problem: each additional gigawatt earns less than the one before it, because they all sell into the same saturated midday hours. High-penetration markets now see routine negative midday prices and rising curtailment, which is why solar's economics are increasingly inseparable from storage, the technology that moves its output to when it is wanted.
The physical limits are gentler. Capacity factors run from around 10 percent in northern Europe to 30 percent for tracked desert plants, land use is real but flexible, agrivoltaics and rooftops both work, and the resource is the most evenly distributed of any renewable. Solar's binding constraints are the grid connection queue and the shape of its own generation, almost never the sunlight.
What breaks
A solar plant degrades more than it fails, but the failure list is specific and mostly electrical.
- Inverters. The dominant source of lost energy in fleet statistics. Power electronics running hot for decades fail long before the panels they serve, and a utility plant plans for inverter replacement or major overhaul at least once in its life.
- Module degradation and defects. Gradual output loss of roughly 0.4 percent a year is the design assumption; the exceptions, cell cracking from rough handling, potential-induced degradation, connector faults that become hotspots and occasionally fires, are found by thermal drone imaging long before a human would spot them.
- Trackers. Motors, gearboxes and control faults, plus the storm problem: a tracker fleet caught at the wrong angle in extreme wind or hail is a write-off risk, which is why stow strategy is now a core part of plant control.
- The weather itself. Hail is the insurance industry's headline solar loss, and soiling, plain dust, quietly costs more energy worldwide than most failure modes. Cleaning schedules are yield engineering.
Control systems, and keeping them safe
Everything a solar plant does on the grid, it does through inverter firmware. That sentence should be read twice, because it is the technology's least-discussed property: the transition is installing hundreds of millions of internet-adjacent power electronic devices, from utility central inverters to rooftop units, most of them remotely updatable by their manufacturer, and a large majority of them made by a handful of vendors in one country. Researchers and, in 2025, mainstream reporting documented undocumented communication hardware in some imported inverters, and several governments now treat inverter firmware provenance as a strategic question rather than a procurement detail. Whatever weight you give any single report, the structural fact stands: aggregated inverter fleets are gigawatt-scale controllable load and generation, and whoever can command them, legitimately or otherwise, holds grid-relevant power.
At plant level the discipline is the same as every other technology here: segmented networks between the plant SCADA, the vendor cloud links and the corporate estate, controlled and logged remote access, and an honest inventory of every device that can speak. The difference with solar is scale and anonymity, ten thousand identical devices, one firmware image, one supply chain. The machinery of securing it is explained at OT Cybersecurity.
Where it fits in the system
Solar is the transition's volume engine: the cheapest marginal electricity in most of the world, deployable at any scale from a roof to a desert, and the reason storage has a business case at all. Its pairing with wind is genuinely complementary, wind in most climates delivers more in winter and at night, precisely when solar does not, and its pairing with batteries is now so routine that hybrid solar-plus-storage is becoming the default project shape in leading markets. How the six technologies compare, and where the 2025 record build went, is on Global Markets.