The reason is that the inverter architecture is not really a line item on a bid sheet. It is a decision about what you want the plant to be good at, and the honest way to make it is to weigh three things at once: cost per watt, availability, and visibility. Central wins the first. String wins the other two. Which of those matters most depends entirely on the site.
The two architectures
The difference comes down to size and count. A central inverter is a large machine, often a megawatt or several, that sits in a station and takes the direct current from a whole field of strings gathered together through combiner boxes. A plant built this way runs on a handful of big inverters.
A string inverter is small, a few tens to a few hundred kilowatts, and it sits out among the arrays, converting the output of just a few strings. A plant built this way carries dozens or hundreds of small inverters spread across the site.
Everything else about the trade follows from that one difference: a few big machines, or many small ones.
Cost per watt, where central wins
Fewer, larger units are cheaper to buy per watt. There is less hardware, fewer enclosures, fewer things to wire, and the economies of a big machine are real. Procurement is simpler too, a handful of large inverters against a bill of hundreds of small ones. On a large, flat, uniform site where the ground is predictable and labour is cheap, that capital saving can be the whole argument, and central is the right answer.
If the only column that mattered were the price of the watt, this article would end here. It does not, because the watt you buy cheaply still has to run reliably for thirty years, and you still have to see what it is doing.
Availability, where string wins
Concentration cuts both ways. Gathering a whole block of the plant into one big inverter makes it cheap, and it also makes that inverter a single point of failure. When a central inverter goes down, it does not take a panel offline. It takes a block offline, often several megawatts, and it stays down until a crew and the right parts arrive. On a hot afternoon with a good price on the grid, that is a large, visible hole in the day’s production.
A string inverter failure is a different kind of event. One unit drops a sliver of the plant, a fraction of a percent, and the fix is usually to unbolt the box and bolt in a new one. No block goes dark. The plant barely notices.
Spread that difference across the operating life of a plant and it stops being an abstract reliability argument and becomes energy, which is to say revenue. A plant that loses less production to outages, and recovers from the outages it does have more quickly, simply sells more electricity. How much that is worth depends on the value of uptime on your particular site, but it is never zero, and it does not appear anywhere in the cost per watt.
A central failure takes a block offline. A string failure takes a sliver. Across thirty years, that difference is measured in energy, not reliability.
Visibility, where string wins again
The second thing cost per watt hides is how much of the plant you can actually see.
A central plant reports at the block level. You know what each big inverter is producing, but beneath it, the output of hundreds of strings is summed into a single number. A block that should be making a hundred units and is quietly making ninety-two looks, on the dashboard, like a block making ninety-two. The eight percent that went missing has nowhere to show itself, and it can bleed for months before anyone notices.
A string plant reports per string. When one string drifts below its neighbours, the data says so, and the loss surfaces while it is still small. The many small inverters also bring many maximum power point trackers, each working the output of just a few strings rather than one averaging a whole field. On a site with shade, with rolling terrain, with rows that do not all see the same sun, that granularity recovers mismatch losses a central inverter simply averages away.
It is the same reason a plant instrumented at the string level finds its underperformance faster. You cannot fix what you cannot see, and central architecture, by design, sees in coarser grain.
The part that complicates it: operations and maintenance
None of this makes string strictly better, and anyone who tells you it does is selling something. The distributed architecture that gives you availability and visibility also gives you far more units to own. More inverters means more connection points, more firmware, more warranties, and more individual things that can fail, even if each failure is small and each fix is simple.
Central and string also ask for different maintenance models. Central concentrates the work: a few large, serviceable machines, repaired in place by specialists, backed by a stock of major spares. String distributes it: many small units, replaced rather than repaired, by crews who can swap a box without a factory technician. Neither is obviously cheaper to run. They are cheaper in different ways, and the right one depends on your crew, your spares strategy, and how far apart your sites sit.
The honest summary is that string moves cost out of the capital column and into the count of things you operate, and buys availability and visibility with it. Whether that trade is worth making is a site question, not a universal one.
How to actually decide
The mistake is deciding on the bid sheet. The capital cost is real and it favours central, but it is one of at least four things that matter, and the other three do not show up until the plant is running.
Model the lifetime energy, not just the capital. Put a number on the production lost to outages under each architecture, and value it at the price your plant actually earns. Look hard at the site: flat, uniform ground with stable geology and simple sun leans central; complex terrain, partial shade, or high-value uptime leans string. Decide the monitoring strategy at the same time, because if surfacing a quiet underperformer quickly matters to you, string-level data is how you get it. And match the maintenance model to the team and the spares you will actually have.
Do that, and the answer stops being a reflex and becomes a judgement, which is what it should have been all along.
The reframe
Central inverters buy you the cheapest watt. String inverters buy you availability and the ability to see what your plant is doing. Neither is the right answer in the abstract, because they are answers to different questions.
So the real decision is not which is cheaper. It is what you want the plant to be good at, and then picking the architecture that is good at that. The bid sheet will always point at central. Whether it is right depends on everything the bid sheet leaves out.
Zigron’s engineering teams work on solar plant design, from inverter architecture and electrical layout to the geotechnical work beneath it. To talk through a project, reach us at info@zigron.com.
