About this transcript: This is a full AI-generated transcript of The Entire Renewable Energy Sector Explained: Solar, Wind, Hydro & Beyond from Leo Cui, Ph.D., CFA , published November 20, 2026. The transcript contains 10,041 words with timestamps and was generated using Whisper AI.
"Imagine walking into a big supermarket on a hot afternoon in late summer. The automatic doors slide open, a wall of cool air hits you, and somewhere in the back, rows of freezers hum. The roof above you is covered in solar panels. On an afternoon like this, they can cover a large share of what the..."
[00:00:00] Speaker 1: Imagine walking into a big supermarket on a hot afternoon in late summer. The automatic doors slide open, a wall of cool air hits you, and somewhere in the back, rows of freezers hum. The roof above you is covered in solar panels. On an afternoon like this, they can cover a large share of what the air conditioning and refrigerators need. The supermarket chain's head office has also signed a 15-year contract to buy the output of a wind farm several hundred miles away. The milk in the dairy case comes from a farm where cow manure is piped into a sealed tank and turned into gas that runs a generator. And the delivery truck backing into the loading dock runs on renewable diesel made from used cooking oil and animal fat. Now stay until closing time. The sun drops, the panels on the roof go quiet, and the freezers keep humming without a flicker. Somewhere, a grid operator has asked a dam to release more water, a battery to start discharging, or a gas plant to ramp up. All within minutes, so that nobody in the frozen food aisle notices the sun has set. Every one of those moments sit at the end of a long chain. Quartz is dug out of the ground, and purified into silicon, so clean that a straight item in a billion matters. Wind turbine blades nearly as long as three basketball courts, laid end to end, leave factories on special trucks. Farmers sign land leases. County boards hold hearings late into the night. Banks lend hundreds of millions of dollars against contracts that last 15 or 20 years. And engineers wash screens around the clock to keep supply and demand in balance. Which brings us to a question that sounds simple. Sunlight and wind are free. Nobody sends an invoice for the breeze. So what exactly are we paying for? And what does it take to turn weather into electricity that keeps the freezer cold after sunset? By the end of this video, you'll understand how the main renewable machines work, and see the whole industry as one connected map, from the factories to the customers. You'll follow a project from an empty field to a working power plant, and watch the money move at every step. And you'll be able to look at the real renewable energy company and explain, in plain language, how it earned its money, and what could go wrong. If you want to watch my videos in advance, you can join the channel as a member. Members can also propose research topics for the next videos. Research tier members can also download infographics from my videos, as well as watching ad-free versions of every video. And this is for educational purposes, not financial advice. A few words need pinning down. Renewable energy comes from sources that nature continually replaces. Sunlight arrives every day. Wind keeps blowing, because the sun heats the planet unevenly. Rain refills rivers. The Earth's interior stays hot, and plants regrow. Using them doesn't draw down a stock that took millions of years to form, the way burning coal, oil, and gas does. Low carbon is a different idea. It describes how much greenhouse gas a source releases for each unit of energy it delivers. Counting over its whole life. Mining and manufacturing the equipment, building the plant, supplying any fuel, running it, and taking it apart at the end. On that whole life basis, wind, solar, nuclear, and most hydropower are low carbon, though none is truly zero carbon. Nuclear, however, is generally not classified as renewable, because its fuel is uranium mined from finite deposits. So the two labels answer different questions. Renewable asks where the energy comes from. Low carbon asks how much the whole system emits. They overlap heavily, but not perfectly. Burning wood is renewable, yet whether it's low carbon depends on what's burned, and whether new growth replaces it. The second distinction is between energy and electricity. In 2025, according to the energy think tank Amber, renewable sources produce about a third of the world's electricity, slightly more than coal. But much of the world's energy never becomes electricity at all. Heat, for warming buildings and running factories, makes up almost half of the energy the world finally uses. And renewables, leaving aside traditional wood fires for cooking, supply only about one-seventh of that heat. In transport, the renewable share is smaller still, just a few percent. So a third of electricity is not a third of energy. That's why this video treats renewable energy as three products. There's renewable electricity from panels, turbines, turbines, dams, geothermal plants, and some biomass plants. There's renewable heat from solar water heaters, geothermal hot water, and boilers that burn wood or biogas. And there are renewable fuels like ethanol, biodiesel, renewable diesel, and biomethane. Most of the money and growth sit in electricity, so it gets most of our attention. But heat and fuel have their own customers and economics. Two more words are easiest to understand with a kettle. A typical electric kettle is rated at about 2kW. That's power. How hard it draws electricity at any moment. Run it for half an hour, and it uses 1kWh. That's energy. The total use over time. And what your electricity bill charges for. Power is the rate. Energy is the amount. Power plants work the same way. A solar farm's capacity, in megawatts, is the most power it can produce. What it actually produces over a year depends on how often it runs near the maximum. A ratio the industry calls the capacity factor. For the solar farm built worldwide in 2025, the International Renewable Energy Agency estimates an average of about 16%. For onshore wind, about 36%. For geothermal, about 85%. Weather explains most of the gap between solar and geothermal. But capacity factors also reflect maintenance, breakdowns, curtailment when a grid can't take the power, and decisions about when to run. That helps explain a gap that surprises many people. Renewables made up nearly half of the world's power plant capacity at the end of 2025, but only about a third of its electricity. Lower capacity factors are part of the reason. Timing is another. So much capacity was switched down during 2025, that much of it ran for only part of the year. Two things often lumped in within renewables usually aren't energy sources in their own right. A battery stores electricity that something else produced and give it back later, minus losses. A typical lithium-ion grid battery returns roughly 85 of every 100 kWh put in. Most hydrogen used today is manufactured, usually by processing natural gas or splitting water with electricity. In those pathways, it carries energy supply during production, and its emissions depend on how it's made. Naturally occurring hydrogen also exists underground, but that's a separate emerging resource. With those words settled, we can draw the map. The renewable energy industry is easiest to understand as a journey with six stations, running from left to right. On the far left are the resources: sunlight, wind, falling water, underground heat, and organic material like crop waste and wood. Nobody owns sunlight, but somebody owns the land it falls on, the patch of seabed where the wind is strongest, where the water rides on a river. Access to a good resource is the first thing every project fights for. The second station is equipment and supply chains, mining, refining, and manufacturing. Polysilicon and glass for panels, steel and fiberglass for turbines, copper for cables, and the inverters and transformers that turn raw output into grid-ready electricity. Much of it is concentrated in a few countries. The third station is project development and construction. A developer finds a site, secures the land, studies a resource, wins permits, negotiates a grid connection, and finds a buyer. Then contractors pour foundations, install equipment, and string cables. This is where an idea becomes a physical asset, and where many proposed projects quietly die. The fourth station is the operating asset, a finished solar farm, wind farm, dam, or geothermal plant, producing for decades. Someone owns it, someone maintains it, and someone collects the revenue. The fifth station is delivery and flexibility, transmission lines, local distribution wires, batteries, and the operators who balance the system every second. A plant that can't get its output to customers at the right moment has produced something worth very little. And on the far right are the customers: utilities, factories, tech companies, supermarkets, households, district heating networks, airlines, and truck fleets. Each want something slightly different, and each pays in a different way. Two forces hang over the whole map. One is finance. Almost every station needs large sums long before any revenue arrives. So the price of borrowing touches everything. The other is policy, auctions, tax credits, import tariffs, permitting rules, and market design, which can make a station profitable or unprofitable almost overnight. One warning before we start. A single company can sit at several stations, and two companies that both call themselves solar can have opposite exposures. A company that makes panels wants panel prices high. A company that owns solar farms wants them low. When panel prices crash, one celebrates and the other bleeds. People asking where on the map a business sits. How busy is this map? In 2025, the world added about 690 GW of renewable power capacity, roughly three-quarters of it solar. That same year, according to Amber, solar alone met about three-quarters of the growth in the world's electricity demand. So let's walk the map, starting with the machine that's doing most of the growing. If you have ever used a pocket calculator with a thin dark strip above the screen, you have used the technology behind some of today's largest power plants. That strip is solar cell, a thin slice of extremely pure silicon, treated so its two layers behave slightly differently. When sunlight hits it, the light knocks electrons loose from the atoms. The boundary between the layers creates a built-in electric field that pushes the free electrons in one direction, and wires on the front and back carry them away as an electric current. No fuel, no flame, nothing spinning. One cell produces very little, so manufacturers wear dozens together, seal them behind glass, and frame them in aluminum. That's a module, the rectangle people call a solar panel. A large farm connects modules end-to-end in long chains called strings, which resists the voltage, the way stacking batteries in a flashlight does. But the strings produce direct current, flowing one way, while the grid runs on alternating current, which reverses direction many times a second. So the strings feed an inverter, a box of power electronics designed to convert direct current into great quality alternating current. A good inverter also keeps adjusting how hard it pulls on the panels, hunting for the setting that squeezes out the most power as clouds pass. Finally, a transformer at the plant substation steps the voltage up, so the electricity can travel efficiently over high voltage lines. Cell, module, string, inverter, transformer, grid, that's a whole physical chain. The panels also need something to stand on. Some farms fix them at one angle. Many large farms use trackers, motorized racks that tilt the roads from east to west through the day, catching more morning and evening sun at the cost of more moving parts. Developers also make a choice that surprises people. They usually install more panel capacity than inverter capacity, often around 30% more. On cloudy mornings, the actual panels keep the inverter busy. At noon on perfect day, a little power gas clipped off the top. That's why solar capacity is often coded in two numbers: one for the panel side, in direct current, and one for the grid side, in alternating current. Mixing them up is one of the most common errors in solar statistics. How much a solar farm produces depends on the strength and angle of the sunlight, set by latitude, season, and cloud. Panels also lose some efficiency when they get hot, and dust, snow, and shade cut output. The operator can't choose when to produce. A solar farm can always produce less, by switching sections off, but never more than the sky allows, and nothing at night. So solar is the best at producing lots of cheap electricity during the day, not at meeting peak demand in the evening by itself. Once built, a solar farm is cheap to run, but not free. There's rent on the land, insurance, property tax, monitoring, cleaning, vegetation control, and repairs. Inverters usually need major repair or replacement at least once in the plant's life, and panels slowly lose output, typically plant at around 0.5% a year. Even so, the construction build dwarfs everything that comes after it, and that fact will drive how solar is financed. Now step back one station, to the factories. The silicon journey starts with quartz, a mineral found in much ordinary sand. Quartz is smelted into silicon, then chemically purified into polysilicon, a process that uses enormous amount of electricity. Polysilicon is melting and grown into large crystals, called ingots, which are sliced into wafers, thinner than a fingernail. Wafers become cells, and cells become modules. Separate industries make the glass, frames, inverters, trackers, and cables. This train is concentrated in one country, to a remarkable degree. The International Energy Agency estimates that China holds around 85% of the world's solar manufacturing capacity, and about 95% for wafers. Many governments worry about that dependence, and have answers with tariffs and incentives for local factories, which protect domestic manufacturing, but usually raise project costs at home. And here's what makes solar manufacturing such an instructive business. Demand grew for years, yet the manufacturers lost money. Chinese companies built factories faster than the world could absorb their output. By one industry estimate, reported by Chaixin in 2026, Chinese production capacity had grown to nearly double global demand. When a commodity supply runs that far ahead of demand, prices fall toward the cost of the cheapest producer, and then below it. Longhi, one of the world's largest silicon solar manufacturers, reported a net loss of about 6.4 billion yuan for 2025, its second loss-making year in a row. The cheap panels flow down the map to the people building solar farms. But even that gain was partly absorbed elsewhere. The International Renewable Energy Agency found that the cost of installing solar fell about 6% in 2025. Yet the cost of megawatt hour of the electricity didn't fall at all. Mainly because borrowing became more expensive, and partly because the average new plant produced slightly less for its size. That's the first lesson the map teaches. Growing demand doesn't guarantee profits. When a product is easy to copy and factories are quick to build, the benefits of falling costs flow to the customer. And the factory owner can end up running hard just to stand still. Hold your hands out of a car window, tilt it slightly, and you'll feel lift. That small upward push is the key to every modern wind turbines. A turbine blade isn't a flat paddle that the wind shoves around. It's shaped like an aircraft wing. As air flows over it, the curved blade generates lift. And because the blades are fixed to a hub, that lifts become rotation. Three blades and a hub together form the rotor. On a large modern turbine, the rotor turns slowly, often 10 to 20 rev a minute, but with enormous force. A generator, on the other hand, likes to spin fast. Bridging the two is the drive chain. And here, manufacturers have made genuinely different choices. Some use a gearbox to step the slow rotation up for a compact, fast spinning generator. Others skip the gearbox and use a large slow generator, often built with permanent magnets, connected directly to the rotor, trading actual weight and costly magnets for fewer moving parts to wear out. A third family sits in between. None is universally better. Each trees weight, cost, efficiency, and how often a technician has to climb the tower. All of that sits in the nacelle, the bus size housing at the top. But the wind constantly changes speed. The generator's output isn't naturally in step with the grid. So power electronics convert it into alternating current at the grid's exact rhythm. Motors turn the nacelle into the wind, and twist each blade to control how much energy it captures. The most important fact about wind is the piece of physics. The power in moving air rises with the cube of its speed. Double the wind speed, and the available power rises roughly 8-fold. That's why developers obsess over wind measurements, and why turbines keep getting taller, reaching steadier air. Turbines also have limits at both ends. They need a minimum breeze to start, hold steady at their rated output in strong wind, and shut down in storms to protect themselves. Like solar, a turbine can always produce less, by turning its blades out of the wind. But it can't create wind that isn't there. Unlike solar, it can produce at night. And in many temperate regions, wind tends to be stronger in the winter, and at night. When solar is weakest, that pattern varies a lot by place. But where it holds, the two complement each other. Wind is designed to bulk electricity, wherever the air moves strongly and often. Then there's a choice between land and sea. Offshore, winds are stronger and steadier, and there's room for gigantic machines. IRENA estimates that offshore farms built in 2025 have an average capacity factor of about 41%, against about 36% onshore. But building at sea is a different industry. Turbines need foundations driven into the seabed, or in deeper water floating platforms. Power comes ashore through long export cables. Installation needs specialized vessels, deep water ports, and calm weather. IRENA's 2025 figures put the average installed cost of offshore wind at about three times that of onshore wind per kilowatt. And in the United States, the investment bank Lazar's 2026 analysis put unsubsidized offshore wind electricity at roughly $105 to $167 per MWh. Well above onshore, blades for the biggest machines approach the length of a football field. And towers are huge steel tubes. So much of the manufacturing sits relatively close to the markets it serves. And getting parts to site takes escorted trucks, temporary road widening, and enormous cranes. Once raining, turbines need far more care than solar panels. Gearboxes and bearings wear. Blade edges erode from rain and dust hitting them at high speed. Technicians climb towers, or offshore, cross from boats in rough water. A turbine maker usually sells its machines under contracts signed years before delivery. A price is hard to change later. If steel, shipping, or labor costs jump in between, or a design turn out to have a quality problem, the maker absorbs the heat. So hardware margins are thin and can turn negative. Siemens Gamisa, the wind unit of German Siemens Energy, shows how costly that can be. After quality problems and expensive offshore ramp-ups, its goal for the 2026 fiscal year was simply to break even. But every turbine sold also creates something steadier. A machine that will need maintenance for decades, often under a long service agreement with the company that built it. Vestas, the Danish company that is one of the world's largest turbine makers, shows the difference. In 2025, on its own adjusted measure of operating profit, its service business earned a margin roughly three times that of the company as a whole. And Vestas reported service contracts worth about 38.7 billion euros in expected future revenue. That's the second lesson of the map. In manufacturing, the hardware can be a fierce low margin fight, while the long tail of maintenance is the steadier business. How much a machine wears, and how hard it is to service. Quietly decides where the durable money sits. Sun and wind shares one more treat. They arrive when the weather decides. The next three resources can be stored, scheduled, or run around the clock. And that changes their business entirely. Start with the oldest renewable machine on the grid, the hydroelectric dam. A reservoir dam holds back a river, creating a lake that stores water high above the turbines. When the operator opens the gate, water falls through large pipes, spins turbines, and drives generators. Think of a bus stop with a drain you control. The reservoir lets the operator decide when to produce, within limits set by rainfall, water rights, irrigation, flood control, and the health of the river downstream. That makes reservoir hydro one of the most flexible sources of electricity there is, designed to follow demand up and down. Run of river hydro deliver part of the river flow through turbines without a large reservoir. Some plants have a small pond that lets them shift output by a few hours. Others have almost none, and simply follow the river. Either way, a dry season means less electricity. Pumped storage uses the same machinery for a different job. It pumps water uphill when electricity is cheap, and lets it run back down through turbines when electricity is valuable. Like a battery, it's storage, not a source, and it gets back less energy than it used to, but it can hold a lot of energy for a long time. Dams take massive civil engineering and years to build, then can run for many decades with modest costs. But many of the best sites are taken. New dams can flood valleys and disrupt rivers and communities, and droughts cut output. Only about 18 gigawatts of hydro power were added worldwide in 2025. A small fraction of the solar build that year. Hydro's value today lies less in growth than in flexibility. It can fill the gaps that sun and wind leave behind. Geothermal power taps the Earth's interior. In places like Iceland, Kenya, Indonesia, and parts of the western United States, hot water and steam sit relatively close to the surface. Companies drill wells into those underground reservoirs and bring the hot fluid up. Steam either spins the turbine directly, or in cooler resources, heats the second fluid with a lower boiling point that drives the turbine. The cooled water is usually pumped back down. Geothermal doesn't depend on the weather. Plants can run day and night in every season, which is why IRENA's estimate's capacity factor for geothermal build in 2025 is about 85%. It's designed for steady always-on power. The catch is the front end. You can be sure a field is hot and permeable enough until you drill. And deep wells cost millions of dollars each. Some come up too cool or too dry. That exploration risk makes geothermal expensive to build. IRENA's 2025 figures put its average installed cost at roughly 9 times that of solar per kilowatt. Over time, reservoirs cool and new wells are added, and hot fluids scale and crude equipment. Newer approaches aim to make geothermal work in many more places. Enhanced geothermal targets hold rock that lack natural channels for water and creates them, using horizontal drilling and fracturing techniques borrowed from the shale oil industry. The most watched test is Fervo Energy's Cape Station in Utah. Fervo says its first 33-megawatt unit began commercial operation on September 30, 2026, after a 23-month build, with a further 400 megawatts under construction. Those are company statements, and independent operating data over the coming years will show how the wells hold up. For now, enhanced geothermal is a promising technology at the very start of its commercial life. Bioenergy is the one major renewable that needs a continuous fuel supply. In a biomass plant, wood chips, pellets, crop residues, or sugarcane fiber are burned to boil water, and steam drives the turbine. The leftover heat often warms buildings or factories. Because fuel can be stockpiled, the operator can choose when to run. Biogas works with microbes instead of flames, as on the dairy farm behind our supermarket's milk. Manure, food waste, or sewage sludge goes into steel tanks, where bacteria digests it and release methane-rich gas. That gas can be burned on site, or cleaned into biomethane, chemically almost identical to natural gas, and inject it into gas pipelines. Liquid biofuels go into vehicles. Ethanol, fermented from crops like corn and sugarcane, is blended into gasoline. Biodiesel and renewable diesel are made from vegetable oils, used cooking oil, and animal fats. And a similar feedstock can become sustainable aviation fuel for planes. Because bioenergy needs feedstock, its economics look less like a solar farm and more like a refinery. The businesses live on the gap between what feedstock costs and what electricity, heat, or fuel sells for. Feedstock is bulky and expensive to haul, so plants depend on supplies within reasonable distance. And how much bioenergy help the climate depends heavily on that feedstock. Waste and residues are one thing, clearing land to grow fuel crops is another. Two last technologies make heat directly. A solar water heater is a dark collector that absorbs sunlight and warms water flowing through it, with no electricity involved. Concentrating solar power uses mirrors to make steam for a turbine, and can store heat in molten salt for the evening, though it costs more than alternate panels. And in places like Reykjavik, naturally hot groundwater is piped straight into homes for heating. Heat travels poorly, so these businesses are local, selling to nearby buildings, greenhouses, and factories. We now have the machines. The harder question is how any of those goes from an idea to a working business. For that, we'll follow one project from start to finish. Let's follow one project from an empty field to a working power plant: Sun Valley Solar. It's fictional, but its numbers are built from real benchmarks, listed in the description. It sits in West Texas, on a grid run by the Texas operator known as ERCOT. And when finished, it will deliver up to 200 MW, using about 260 MW of panels. The story starts with the developer, a company whose business is turning possibilities into projects. He looks for strong sunshine, flat and affordable land, and something many people don't expect to matter so much. A nearby stop station with a spare room. A grid site far from the power grid can be less valuable than an average site right next to it. The developer signs options with landowners, small payments for the right to lease the land later, and sets up a weather station, because landers will eventually want an independent estimate of how much electricity the site will produce. Then come permits and the community, environmental studies, county approvals, and public hearings where neighbors ask about views, drainage, and traffic. Developers who adjust their plans and share benefits locally tend to get through. At the same time, the developer applies to connect to the grid. The grid generator studies whether existing lines can carry another 200 MW of the busiest moments. If upgrades are needed, the project may have to pay for them, and a surprise upgrade bill can kill it. This is often the slowest step. In the United States, the median project completed in 2025 took about five years from its grid application to commercial operation. According to Lawrence Berkeley National Laboratory, taxes has generally been quicker than most regions. One reason San Valley is there. None of these steps waits politely for the last one. Land, studies, permits, the grid application, and talks with buyers all run in parallel. And a failure in any one can end the project. Of all the generating capacity that asked to connect to U.S. grids between 2000 and 2020, only about one-eighth were operating by the end of 2025. Early on, the developer is spending a few million dollars on a long shot. And every uncertainty it resolves, a signed lease, an approved permit, a grid agreement, makes the project more likely to be built and therefore more valuable. Setting a fully permitted project for much more than it costs is how many developers earn their living. The most important document comes next: the customer contract. San Valley signs a deal with a large company that wants clean electricity. The company agrees to buy plant's output for 15 years at a fixed $80 per megawatt hour . That's a power purchase agreement, or PPA. The buyer gets a predictable price, and the project's revenue becomes far more predictable, though not guaranteed, because it's still paid only for what the plant produces. In our example, that price is fixed at an agreed market location. Think of it as a reference point on the grid. If congestion makes electricity cheaper where San Valley actually connects, or if the plant produces less or is curtailed, the project can still lose some revenue. After 15 years, the contract ends, and the plants sell into the open wholesale market at whatever the price happens to be, which the industry calls merchant exposure. For those later years, we'll assume a lower price of $45 per megawatt hour to show what happens when the contract ends. That is a teaching assumption, not a forecast. Building San Valley, including panels, trackers, inverters, the substation, the grid connection, labor, and development costs, come to $260 million, about $1.30 per watt, near the lower end of the range Lazard reports for new US solar farms in 2026. That money comes from two kinds of investors. Lenders, usually banks, provide most of it in exchange for fixed loan payments, and they are paid first. Owners provide the rest, take whatever is left after the lenders, and expect a higher return for that extra risk. To decide how much to lend, the bank looks at the cash the plant should produce while the contract runs, and makes sure each year's cash covers the loan payment with a cushion, here 30%. The industry calls that the debt service coverage ratio. With a 15-year loan, time to end exactly when the contract does. The bank lends about $200 million, and the owner put in about $60 million. The size of that loan depends far more on the customer contract than on the panels. Lenders still track the production estimate, the construction contract, the equipment warranties, the buyer's credit worthiness, and what the plan itself would be worth if things went wrong. A contract alone doesn't make a project bankable, but it turns all of those into a stream of cash a bank can count on. Construction goes to a constructor that handles engineering, procurement, and construction, usually for a fixed price, so overruns the land mainly on the contractor. It orders equipment from the factories we met earlier, drives thousands of steel posts, mounts trackers and panels, and builds a substation. After about a year, the plan is tested, synchronized with the grid, and declared commercially operational. Only then does the first dollar of revenue arrive, years after the first dollar was spent. To keep the arithmetic readable, this example leave out taxes and inflation, and uses assumed future prices. It explains how the business works, rather than predicting what a real project will earn. Now let's follow the cash through Sun Valley's first full year. At a 25% capacity factor, typical of US solar farms with trackers, the plan could deliver about 438,000 megawatt hours. We'll assume 4% is a loss to curtailment, leaving about 420,000 megawatt hours sold, multiplied by $80, and the revenue is about $33.6 million. Operation costs, covering maintenance, land rent, insurance, property tax, management, and money set aside to replace inverters, take about $4.9 million, leaving $28.7 million. The loan payment takes about $22.1 million, and about $6.6 million is left for the owners. And that's the whole business in one line. Electricity sold, times the price actually received, minus operating costs, minus loan payments, equal cash for the owners. Three numbers often get blurred here, so let's separate them. The generation cost is the flat price per megawatt, watt hour, held for 30 years. That would just cover building and running the plant, while paying lenders and owners the returns they require. Assuming a blended required return of 7.5%, Sun Valley is about $67. The industry calls it the levelized cost of energy, or LCOE. The selling price is different, $80 for 15 years, then our assumed $45. And the investor return depends on the gap between the two, how much is actually sold, and how the plant was financed. Using the timing of the owners initial investment, and all the cash they received afterward. This example works out to an annualized return of roughly 12% over 30 years. That is not a fixed annual payout, and it depends on the project delivering the cash flows we assumed. Measuring today's money, the construction bill is about 4/5th of everything Sun Valley will ever cost. It's spent before the plan runs, even though most of it is borrowed and repaid out of revenue over the next 15 years. Now let's do a stress test. Suppose it's running, the loan is fixed, and something cuts its revenue by 15%, perhaps more curtailment, or a lower price where the plan connects than at the contract's reference point. Revenue falls by about $5 million, operating costs don't fall, and the bank still takes its full $22.1 million first. So the owner's cash in year 1 drops from about $6.6 million to about $1.6 million, a fall of more than three quarters. The lender's 30% cushion shrinks to about 7%, close to the point where many loan agreements block payouts to owners. A 15% revenue problem becomes a 75% problem for the owners, because fixed debt payment magnifies everything. Other risks run through the same machinery. Higher interest rates when the loan is signed mean a smaller loan and more money from the owners, so their expected return falls even though the plan is identical. If the cost of money rises for lenders and owners alike, two extra percentage points would add roughly a 6th to some value generation cost, and the construction or grid delay piles up interest while nothing is sold, pushing every dollar of revenue a year further out. Policy can change all of these numbers. In the United States, a 2025 law ends federal tax credits for new wind and solar projects placed in service after 2027, unless construction began by early July 2026, and many other countries run auctions for long-term contracts. Sun Valley has no subsidy, because the subsidized price should never be set against unsubsidized costs. Once this has run smoothly for a few years, the developer might sell it to an infrastructure fund or utility happy with lower returns for steady contracted cash. At the end of its life, the owner can re-equip the same site and grid connection, which may by then be the most valuable part, will restore the field. Other technologies change the story. Offshore winds start with winning a seabed lease and years of ocean service, then multi-billion dollar construction that depends on ships, ports, and permits. So political and financing shocks hit much harder. Geothermal carries its biggest risk at the very start, when drilling reveals whether the resource is really there. A rooftop system skips most of this, and its customer is the building itself. We'll come back to the rooftops when we meet the customers. First, let's follow the electricity after we leave the plant. A megawatt hour is not just a megawatt hour. Its value depends on when it arrives and where. Think about strawberries. In June, when every farm nearby is harvesting at once, the price at the market collapses. In January, the same strawberry is a luxury. Solar electricity behaves the same way. Solar farms in the same region produce at roughly the same hours, because they share the same sky. When there are only a few, midday solar is valuable, because it replaces expensive power. As the more are built, each new farm adds output to hours that are already well supplied, and the price in those hours falls. The industry has a way to measure this. Take the wholesale prices in the hours a solar farm produces, weigh them by how much it produces in each, and compare that average with the average price across all hours. The ratio is called the capture rate. In October 2025, Spain's solar output fetched an average wholesale price of about €42 per megawatt hour, while the average across a month was about €74, a capture rate of roughly 57%. That's what the market values solar output at, not necessarily where every owner was paid. A long-term contract can protect the owner from some movements in wholesale prices, but the protection depends on its terms. Sun Valley still bears production, curtailment, and location price risks. Captured rates will also shape what buyers will offer for new contracts, and what the plant earns after its existing contract expires. Push far enough, and midday prices drop below zero. In Germany, the wholesale price fell below zero in 573 hours during 2025. About one hour in 15. Why would anyone keep producing at a negative price? Some producers are paid per unit, through contracts or subsidies, regardless of the market. Though rules in several countries are changing, so plants simply stop in those hours. Location matters too. Electricity travel over transmission lines, the grease highways, and each line has a limit. A region full of wind or solar can produce more than its lines can carry out, even when cities a few hundred miles away could use the power. Then some plants are told to cut back. That's curtailment. And much of it is a local traffic jam, rather than a sign a whole grid has too much power. It's also why electricity can cost very different amounts, at two points, on the same grid, at the same moment. The risk we met in Sun Valley's contract. So the system has to move energy across space and time. Across space, the answer is more and better transmission, which lets a calm day in one region borrowed from a windy day in another. Across time, the answer is storage. A battery has three numbers worth knowing. Its power, in megawatts, is how fast it can charge or discharge. Its energy, in megawatt hours, is how much it can hold. Divide the second by the first and you get its duration. A 100 megawatt battery holding 400 megawatt hours can discharge at full power for 4 hours. Most grid batteries built today hold 2-4 hours. Designed to soak up cheap midday solar and release it into the evening. Battery prices have fallen fast. IRENA estimates that the install cost of 4-hour utility-scale battery fell close to 30% in 2025, to around $140 per kWh on global average. But the battery's profit isn't simply the gap between cheap and expensive hours. The owner pays for electricity it charges with, loses roughly 15% of it on the round trip, pays to operate the system, watches capacity fade at sales age, and must recover a large upfront investment. Revenue comes from that gap. From payments for being available at peak, and from grid-stabilizing services. And as the more battery trace the same evening gap, they narrow it. A battery can rescue a solar project's economics, or it can be an expensive bet, depending on the market. Demand can move too. Electric cars can charge at noon instead of midnight. Water heaters and air conditioners can shift by an hour or two. And factories can schedule energy-hungry work for cheap hours. And complementary generation, reservoir hydro, geothermal, biomass, gas, and nuclear can run when the sun and wind are low. Two questions often get mixed together here: what it costs to build a plant, and what it costs to run a whole reliable system. A cheap solar farm doesn't make a cheap reliable grid by itself, because the system also pays for wires, storage, flexible plants, and controls. But it's equally wrong to say each solar or wind farm needs its own backup plant of the same size. Grids plan for the whole system's toughest moments, and credit each resource for how much it can be counted on then. Those toughest moments differ by place. In much of Northern Europe, the worst stretch may be a cold, dark, windless week in winter. In Texas, it may be a scorching evening at the sunset, or a deep freeze. In February 2021, a winter storm shut down power plants across Texas, mainly because equipment froze and natural gas supplies failed. Reliability is about the system's worst week, not its average day. And every kind of power plant has failure modes. That's why the same solar farm can be very valuable on one grid, and struggle to earn a living on another. We have followed the electricity to the edge of the grid. Now let's meet the people who pay for it. At the far right of our map see the customers. And the most useful thing to know about them is that they're buying different things, even when the product looks identical. The biggest buyers of renewable electricity are utilities and retailers, the companies that send bills to homes and businesses. Some own plants, others buy in wholesale markets or sign long-term contracts with plants like Sun Valley. What they need is enough electricity in every hour of the year, at a price they can defend to regulators and customers. Which is why utilities increasingly ask for solar paired with batteries, or a mix of resources that cover the evening. Then come corporate and industrial buyers, like Sun Valley's customer, or our supermarket train with its wind contract. A large company signs a long-term deal for two reasons: to lock in predictable price, and to be able to say its operations run on renewable electricity. The second part is tracked with renewable energy certificates. Each certificate represents the renewable attributes of one megawatt hour of electricity generated. To use it for renewable electricity claim, a buyer retires it, so nobody else can claim the same certificate. Subject to the relevant accounting rules, a company can match its annual electricity use with eligible certificates. That is an accounting match, not proof that renewable power reaches buildings every hour. On a calm evening, the electricity reaching our supermarket freezers came from whatever the grid was running. That's why some of the largest buyers, especially technology companies, now try to match their use hour by hour. Which raises the value of resources that run at night, like hydro and geothermal. The supermarket's rooftop is a different kind of deal, and so is the house with panels. Here the owner isn't many selling electricity, it's avoiding buying them. Every kilowatt hour used straight from the roof is one not bought from the utility, and the retail price includes the cost of wires, meters, and taxes on top of the electricity itself. That's why rooftop solar can make financing sense, even though it costs much more per unit than a utility scale farm. It competes with the price on the bill, not the wholesale price a solar farm receives. Not every charge disappears though. Many bills include fixed monthly fees, and a larger customer often pay charges based on their peak demand, so buildings with panels still pay to stay connected. Any extra power the panels produce goes back to the grid, and how much you get paid for it depends on local rules. Some places credit exports at the full retail price, which is called net metering. Others pay something closer to the wholesale value, which can be a small fraction of that. Cutting export payments makes owners lean on their own power more, often by adding a battery. That makes rooftop solar one of the most policy sensitive corners of the industry. It includes installers whose biggest costs are sales and labor, lenders, third-party owners who install systems for no upfront cost, and then lease them or sell the power. And the utility, which shapes the export rules. In the United States, the federal tax credit for homeowners who buy their own systems ended for expenses after 2025. A reminder of how quickly household economics can shift. Then there are customers who don't want electricity at all. District heating networks, factories that need steam and hot water. And greenhouses buy heat, usually under long-term contracts with a nearby producer. Because switching heat suppliers is rarely practical. Remember the scale. Renewables, excluding traditional cooking fires, supply only about one-seventh of the world's heat. And most of the rest still come from burning fossil fuels. Fuel customers are different again. Fuel distributors blend ethanol and biodiesel into gasoline and diesel. Often because laws require a minimum share. Truck fleets and city buses buy renewable diesel. And airlines are starting to buy sustainable aviation fuel, pushed by mandates such as the European Union's minimum share in jet fuel. These fuels usually cost more to make than the fossil version they replace. So their market depends heavily on mandates and credits. And their prices tend to move with the fossil fuel they replace, which ties the renewable business to the oil market. Put those customers side by side. And there's no single renewable business model. A solar farm in a wholesale market earns whatever its hours are worth. A rooftop earns the retail bill it avoids. A biomass heating plant earns the local contract price for steam. A renewable diesel refinery earns the gap between fee stock and fuel prices, propped up by policy. We have now walked the whole map, from quartz and seabed to loading dock. It's time to put it together. Every business we have met can be sorted into four questions. What determines its revenue? What consumes its capital? What protects its returns? And what could break the business case? Manufacturers earn units sold times price, and their capital goes into factories. And for turbine makers, contracts price years ahead. Longi showed what overcapacity does to them, and investors showed how a fleet that needs servicing can protect them. Developers earn by selling projects once the risks are cleared, spending risk money across many sites, most of which fail, as we saw at Sun Valley. A project that dies late, or a sudden shift in rates or policy, is their nightmare. Builders, the construction contractor and rooftop installers, earn a margin on each job. A fixed price drop that runs over, or a tariff changes that drives up rooftop demand, can wipe that margin out. Assay owners, independent power producers, infrastructure funds, and pension funds, earn electricity delivered times the price actually received, finance mostly with debt. Long contracts and cheap capital protect them. Curtailment, falling market prices, and contracts expiring into weak markets are what our stress test showed can hurt them. Operations and maintenance providers earn service fees with modest capital, protected by skills and data, and squeezed where equipment is simple and competitors are many. And the network businesses split in two. Companies that own transmission and distribution lines, often regulated utilities earn an approved return on what they have invested. So a hostile regulator or a backlash over bills is their main risk. System operators like ERCOT in Texas, PJM in the eastern United States, and CAISO in California, run the grid and its market, but own few or none of the wires, and fund themselves mainly through fees. In some countries, one company does both, as Red Electrica does in Spain. But the incentives differ. A wire owner earns more by building more, while an operator succeeds by keeping its lights on at the lowest cost. Put together, a pattern appears. The more a business depends on commodity prices, the more its performance rises and falls with the industry. The closer it sits to a long contract, or a regulated monopoly, the more it depends on the cost of money and on regulators. Now let's test a map on real companies. Here are six companies in three pairs. Each comparison shows a different way to make money from renewable energy. The same news can affect them differently, because their customers, contracts, costs, and assets are different. None of this is a recommendation. It's practice in reading a business. The first pair is Longhi and First Solar, both selling solar panels to developers and builders. Longhi, the Chinese silicon giant we met in the price war, sells into a global market where panel prices are set by their overcapacity. First Solar, based in the United States, makes thin-field panels from cadmium, telluride, so it doesn't rely on the poly-silicon chain. It has been expanding its American factories and sells much of its output to American developers. In 2025, its sales were about $5.2 billion, and a large part of its profit came from a US manufacturing tax credit paid for each panel made domestically, which is sold to other companies for well over a billion dollars in cash. Now imagine wall panel prices fall another 10%. That puts pressure on the prices Longhi can secure for new sales, though existing contracts influence how quickly the effect reaches earnings. First Solar also faces competition, but its exposure differs because of the technology, customer contracts, US trade protections, and domestic manufacturing incentives. If Washington reduced the manufacturing credit, first Solar would be directly exposed. If it eases the import restrictions, that could change competition for both companies. They share a market, but they do not share the same mix of price, contract, and policy risks. The second pair is Vestas and Ersted, both tied to wind, from opposite ends of the map. Vestas sells and services turbines. Ersted, majority owned by the government of Denmark, is the developer and owner. It wins offshore seabed leases, manages construction, keeps stake in the finished farms, and often sells part of each farm to partners such as pension funds. That puts billions of its own capital at risk before a farm earns anything, which is why its American projects have been so painful. In 2023, it canceled two New Jersey projects and recognized large write-downs. In 2025, federal stop-work orders hit its US projects under construction, and after co-investors withdrew from one of them, Ersted raised about 60 billion Danish kroner, roughly 9 billion dollars, from shareholders, much of it to fund that project. Now imagine a big offshore farm is delayed by a year. For Ersted, that's a year of interest and standby costs with no revenue. For Vestas, if it's a turbine supplier, it means the delayed delivery and perhaps disputed costs, while its existing service contracts keep earning. Or imagine financing becomes cheaper. That can improve the economics of Ersted's new projects, or future refinancing. Although existing fixed-rate debt does not automatically become cheaper, Vestas can benefit as more customers can afford to place turbine orders. The third pair goes beyond panels and turbines. Brookfield Renewable is a long-term owner and operator of hydroelectric dams, wind and solar farms, and batteries across many countries. In 2025, its hydro business produced about 607 million dollars of its roughly 1.33 billion dollars in funds from operations. A cash flow measure the company reports, and Brookfield says about 90% of its output is contracted, on average for about 14 years. In 2025, it also signed 20-year contracts to supply Google with power from two hydro plants in Pennsylvania, and a framework for up to 3,000 MW, a direct payoff from hydro's ability to run when the sun doesn't. It also holds an interest in the nuclear company Wessinghaus, a reminder to check what a company with renewable in its name actually owns. Nest, from Finland, is a fuel refiner that calls itself the world's leading producer of renewable diesel and sustainable aviation fuel, made mostly from waste and residues like used cooking oil and animal fat. Its revenue depends on fuel prices plus the value of policy credits in places like California and the European Union, minus what it pays for fee stock, and that's price swings. Nest's renewable products margin, on its own comparable measure, fell to about $242 per ton in the last quarter of 2024 for more than $800 a year earlier, then average about $411 across 2025. And its oil refining business still earned slightly more than its renewable one in 2025, measured by comparable earnings before depreciation. Now compare their direct exposures. A drought can reduce Brookfield's hydro generation, with the cash impact depending on its contracts, power prices, and performance elsewhere in the portfolio. That is not Nest's central operating exposure. A rise in used cooking oil prices matters much more directly to Nest. If product prices and policy credits do not rise enough to offset it, refining margins shrink. Brookfield's generating assets are much less directly exposed to that fee stock price. One business depends heavily on operating assets, resource conditions, and contracts. The other depends heavily on the relationship between fee stock costs, fuel prices, and policy incentives. So which change matters most for the decade ahead? Forecasts disagree widely because they rest on different assumptions about forces we have already met. What matters is how each could move, who feels it first, and how you notice. Start with equipment. Innovation and skill keep lowering the cost of making panels, turbines, and batteries. But raw material prices, trade rules, shipping, and the health of the manufacturers can offset those gains or reverse them for a while. If the price war continues, developers and owners keep buying cheaply while factories keep losing money. If manufacturers consolidate, prices may recover, helping companies like Longhi, and costing buyers. Manufacturers' price and margins are the signal to watch, then money and policy, because the construction bill is most of what a wind or solar farm will ever cost. Borrowing costs hit developers and owners first, and they depend far more on the country than on the technology. Irene's modeling suggests that conditions in a country, its interest rates, currency risk, and the reliability of its buyers. It explains more than half the variation in what renewable projects pay for capital, more than twice as much as the choice of technology. Policy can move faster still. In China, after the government announced that new projects would shift from fixed prices to market-based pricing from June 2025, developers rushed to connect first, adding about 93 gigawatts of solar in May 2025 alone. In the first 9 months of 2026, China's solar installations then fell by more than half from the year earlier. Neither Swing said anything about weather panels work. Borrowing costs where projects are built, and the rules in the biggest markets are worth to watch. Next, the grid, storage, and flexibility. If transmission, batteries, and flexible demand keep pace, new solar and wind keep their value, and asset owners keep building. If they lack, curtailment rises, capture rates fall, and profit shifts toward whoever owns grid connections, storage, and flexible plants. Watch grid connection times, the miles of new transmission actually completed, battery installations, and their hours of storage, and a capture rate for solar where you live. Demand has changed direction. Electric cars, heat pumps, factories, and data centers are pushing electricity use up again in many countries, and growing demand absorbs new supply instead of simply pushing prices down. And then there's the ground itself, permits, neighbors, and physical limits. Projects that share their benefits with neighbors tend to move faster. Land near cities, water for hydro in a changing climate, sustainable biomass, and minerals like copper rarely stop the industry outright, but they shape where it grows and what it costs. Two countries show what it looks like when those forces line up. Uruguay, a country of about 3.5 million people, generated about 98% of its electricity from renewable sources in 2025, with hydro the largest source and wind and biomass together supplying more than half. After the 2000s exposed how dependent it was on rainfall and imported oil. It adopted a long-term energy policy that the commission of all its parliamentary parties endorsed in 2010, and its state-owned utilities signed long-term contracts with wind developers. Its dams work like a giant battery, saving water when the wind blew. The limits are real. In the drier 2025, fossil generation doubled from a year before, though only to about 2%. And because transport still runs largely on oil, renewables supply a little under two-thirds of Uruguay's total energy. The transferable lesson is that stable rules and reliable buyers lower the cost of capital. Pena shows a similar principle underground. In the year to June 2026, geothermal plants in the Rift Valley produce about 41% of the electricity, according to the National Energy Regulator, more than any other source. Stay back the companies carry much of the riskiest early drilling, so later plants can be built on proven fields. The grid still has limits. Hydro depends on rain, imports have grown, and on some nights, when demand is low and wind is strong. Geothermal output has had to be cut back. You can't copy the geology, but you can copy the strategy of using public funding to reduce the risk of early drilling. So the future won't arrive as one global number. It will show up place by place, in capture rates, borrowing costs, connection queues, battery hours, and factory margins. Those are the dials worth reading. Let's go back to the supermarket, just before closing. The panels on the roof has gone quiet. Each began as quartz, was purified and sliced in a factory that may well have lost money making it, and was installed by a contractor for whom winning the drop was one of the biggest costs. All afternoon, they saved the store from buying electricity at its full retail price. The wind farm under the chain's contract is turning in the evening breeze. Its turbines came from a manufacturer that also earns recurring revenue servicing them. And the farm is owned by investors repaying a bank out of every megawatt hour. The milk came from a farm whose manure became few. The delivery truck's tank holds diesel made from waste fat, by a refiner living on a spread that exists partly because of laws someone voted for. And the freezers keep raining after sunset, not because one power plant back up the solar panels, but because the entire system is designed to handle the hours when power is needed most. Dams that can release water, batteries charged at noon, flexible plants ready to start, and the wires that bring power from places where the weather is different. So back to our question. Sunlight and wind are free. What we pay for is the machinery that catches them, the factories that make it, the money that builds a plant years before it earns anything, the lands and the permits, the wires, and the storage and flexibility that deliver energy at the hour where someone needs it. Not every renewable works without an ongoing feedstock supply. The dairy farm collects and processes its own manure, while the renewable diesel producer sources oil and fats. Both have costs that the passing breeze does not. But for sun and wind, the fuel costs nothing. Making it dependable is the industry. If you enjoyed this video and want to see more deep dive like this, please subscribe, like this video, and turn on the little bell icon. And if you want to watch new videos in advance, you can join the channel as a member. Members can also propose a research topic for the next videos. Researcher members can also download infographics from the videos, as well as watching ad-free version of every video. Thank you so much for watching. I'll see you in the next one.