"Renewables" are not Renewable

The fossil fuel foundation of wind, solar, and batteries

Note: I’m writing from Houston this week while I am attending CERAWeek. Today’s post is one that I’ve been contemplating for a while. I was motivated to finish it after hearing remarks by U.S. Secretary of Energy Chris Wright earlier this week on the importance of process heat in industry. I often include a line in my talks about how wind turbines and solar panels cannot be made using energy generated by wind and solar. Today I formalize that argument, and it is eye opening. —RP

Today’s post starts with a simple question: Can wind turbines and solar panels be created from a supply chain powered by wind turbines and solar panels?

The answer is no.

Wind turbines and solar panels come from supply chains that are fossil fuel intensive and technological options to replace those fossil fuels in their production do not yet exist, and may never exist. This post unpacks the details.

To be absolutely clear, what follows is not an argument against wind and solar. THB readers will know that I am bullish on solar and not so much on wind. I’ve long argued that the lowest hanging fruit for large emissions reductions is dirty coal plants, which can be replaced with natural gas, nuclear, as well as wind and solar with storage.

Today’s post is an exercise in understanding quantitatively the true challenges of an energy transition and move beyond the claim that we have all the technology we need for deep decarbonization — typically emphasizing extensive deployment of wind and solar energy generation, accompanied by battery storage.

So-called “renewables” are not remotely renewable. To be sure, solar and wind technologies, coupled with storage, can contribute to the decarbonization of electricity. However, they are each built on a deep foundation of fossil fuels.¹

Let’s look at some numbers.

The arguments in this post apply to nuclear power as well, however, nuclear is less resource intensive than wind and solar, plus battery storage. Source: Our World in Data.

The IEA’s Net Zero by 2050 roadmap calls for solar PV capacity to increase 20x and wind power 11x. These increases require that annual solar additions must reach 630 GW per year by 2030 and wind must see annual increases of 390 GW. Battery storage must increase 14x to 1,200 GW by 2030.

These numbers imply an unprecedented mobilization of materials and industrial production. For example:

• The world currently produces about 1.9 billion tonnes of steel per year. Wind turbines are 71–79 percent steel by mass. The Energy Transitions Commission estimates that a net-zero energy transition requires 6.5 billion tonnes of end-use materials between 2022 and 2050, of which 95 percent is steel, copper, and aluminum — which is about three and a half years of current total global steel production.

• The world produces about 23 million tonnes (Mt) of copper per year. A January 2026 S&P Global study projects a 10 Mt copper shortfall by 2040. Mine development averages 17 years from discovery to production — That means that projects started today will not produce copper until the early 2040s.

The manufacturing of wind turbines, solar panels, and batteries at scale is not a niche activity in a few high-tech factories. It requires the sustained output of the entire global heavy industrial base — steel mills, cement plants, copper smelters, aluminum refineries, petrochemical complexes, glass furnaces, and the shipping networks connecting them. Every one of those industries currently runs on fossil fuels, with no commercial zero-carbon alternatives widely deployed in its most energy-intensive processes.

Making primary steel from iron ore — about 70 percent of global production — requires metallurgical coking coal in a blast furnace at around 1,500°C. Coal is not simply burned as fuel to create very high heat, it is also used in the chemical process that removes oxygen from iron ore to make iron. In 2023, less than 1 Mt of near-zero emission steel was produced globally, of a total global production of 1,889.2 Mt.

In its Net Zero 2050 scenario, the IEA projects that steelmaking in 2050 would still use significant coal — for ~22 percent of energy input — and theoretically paired with carbon capture and storage that does not yet exist at commercial scale.

The foundation under a wind turbine is reinforced concrete. Cement kilns run at around 1,450°C, and about two-thirds of cement’s CO₂ comes not from burning fuel but from a chemical reaction that happens regardless of what source heats the kiln. Full decarbonization of cement has been projected to double its cost and also requires industrial-scale carbon capture and storage that does not yet exist.

Solar panels are similarly carbon intensive. Producing solar-grade polysilicon requires smelting quartz at 1,500–2,000°C, followed by chemically intensive purification. According to the IEA’s Special Report on Solar PV Global Supply Chains, coal generates more than 60 percent of the electricity used in global solar manufacturing and in China, which dominates solar manufacturing, that figure exceeds 75 percent.

The glass covering a solar panel — about 75 percent of its weight — is made in furnaces at around 1,100°C fueled by natural gas or coal. The aluminum frame requires fossil-fuelled smelting. The silver contacts come from diesel-powered mines. Other materials come from petrochemicals. Then, panels are shipped around the world on vessels burning heavy fuel oil.

There is another category of fossil fuel dependency in solar panel and wind turbine supply chains: chemical feedstocks, necessary to create the many components necessary to assemble the final products.² Wind, solar, and battery manufacturing necessarily depend upon the petrochemical industry, which the IEA projects will continue growing through 2050 in every scenario.

Batteries, necessary to store electricity when the wind does not blow and the sun does not shine, are fossil fuel intensive as well.³ Batteries last ~10–13 years, which means they need replacing two or three times over the life of wind or solar generation assets they are paired with, which have lifespans of ~25-30 years. Every replacement cycle is a full repeat of mining, smelting, and manufacturing.⁴

Wind turbines, solar panels, and batteries are products of the entire global industrial base. That base accounts for about 37 percent of global energy-related CO₂ emissions, with five heavy industries — cement, steel, oil and gas, chemicals, and coal mining — accounting for 80 percent of all industrial emissions.

The figure below shows an estimate of the carbon dioxide (CO₂) emissions from manufacturing supply-chains for new wind, solar, and battery capacity. Annual emissions have grown from ~4 Mt in 2000 to ~470 Mt in 2023 — about 1.3 percent of global energy CO₂, and comparable to the total annual emissions of South Korea or Canada. That growth is a pure volume effect: manufacturing carbon intensity per GW has fallen substantially, but absolute emissions have risen because deployment scale has grown much faster than intensity has declined.

We can get a sense of the technological challenge of decarbonizing supply chains for wind and solar by looking at net zero scenarios and backing out what they imply in terms of needed resources. In a 2008 paper in Nature with Tom Wigley and Christopher Green, we called this a “frozen technology baseline” — If we freeze technologies at today’s level and then look at what projections imply for the future, that then tells use how much technological improvement is actually assumed in the scenarios. We argued that “it is only with a clear-eyed view of the mitigation challenge that we can ever hope to adopt effective policies.”

In the exercise, manufacturing carbon intensity is frozen at 2024 levels, and I explore implied carbon dioxide emissions implied to 2050. The point is not to predict the future. It is to isolate the effects of assumed technological innovation within scenarios.

Advances in technology do not occur on predictible schedules, however scenarios of deep decarbonization often assume JITTI — Just In Time Technological Innovation.⁵ JITTI allows scenarios to assume technologies necessary for deep decarbonization will appear at global and industrial scale just when the world needs them to transform the global energy system. Convenient!

The figure below shows projected CO₂ emissions from wind, solar, and battery supply chains projected to 2050 under a frozen technology baseline for the IEA’s net zero scenario (NZE), its stated policies scenario (STEPS), and a simple extension of the historical trend.⁶ The historical data in the figure is the same found in the figure above, which gives a sense of scale.

The results are incredible — and described in more detail below.

In the IEA’s Stated Policies Scenario (STEPS), annual supply-chain manufacturing emissions are ~870 Mt by 2030 and ~1,600 Mt by 2050. That 2050 figure exceeds Japan’s entire national CO₂ output today — with a population of 125 million and a $4 trillion economy — and approaches the combined annual fossil CO₂ of Germany, France, the United Kingdom, Italy, and Spain.

In the IEA’s Net Zero Emissions Scenario (NZE), supply-chain emissions are ~1,540 Mt by 2030 alone — similar to the combined emissions of Germany, France, and the United Kingdom. By 2050 in the NZE, the figure is ~4,000 Mt — comparable to the current annual fossil CO₂ of the United States, or ~10% of today’s total global emissions of carbon dioxide from energy.

The NZE scenario requires the most new infrastructure, so it generates the most supply-chain emissions under frozen technology assumptions. For deep decarbonization to occur, both the massive hardware build-out and the assumed decarbonization of the global industrial base must happen simultaneously.

Consider that the IEA NZE roadmap requires that every month from 2030 onwards, ten heavy industrial plants are equipped with carbon capture and storage, three new hydrogen-based industrial plants are built, and 2 GW of electrolyser capacity is added at industrial sites. That is the minimum background rate of industrial transformation required just to keep the scenario on track, independent of the deployment of wind, solar, and batteries across global grids.

The narrow focus on wind, solar, and batteries by many climate advocates obscures the fact that these technologies do not emerge spontaneously from zero carbon industrial processes. The steel industry accounts for roughly 7–9 percent of global CO₂ annually. Cement accounts for another 6 percent. Copper, aluminum, chemicals, and the petrochemical feedstocks woven through every component add more. These are industries with capital stock turning over only once every 25–40 years, where investment decisions made today lock in emissions profiles for decades.

Wind and solar do reduce overall emissions when they displace fossil generation on the grid. But the energy transition is not simply a story of replacing electricity generation from fossil fuels system with lower carbon alternatives. Far more importantly, it is a story of transforming the foundations of the global industrial base — and today, that transformation is a long way off.⁷

Scenarios of deep decarbonization have long assumed that technological progress would achieve what is required on schedules that align with political targets. The next time you hear numbers on the deployment of wind, solar, and batteries, acknowledge that reality, and then ask about rates of decarbonization in steel, cement, copper, aluminum, petrochemicals, glass, shipping and the other foundations of the modern world.

Before you go — If you enjoy deep dives like this one, let me and everyone else know by clicking that “❤️ Like”. Thanks!

Comments, questions, discussions, debate, challenges — all welcomed!

Announcement: The annual subscription for new subscribers to THB is increasing to $100 on 1 June 2026! Lock in your support at a lower level, at a price that will never increase!

1. I think that the demonization of fossil fuels (and by extension the global industrial base which powers the modern world) by many activists in the climate movement — notably some academics, scientists, journalists, and politicians — creates strong pressure to simply avoid discussing their essential role in the global economy.

2. Claude AI provides some details: “Wind turbine blades are built from thermoset epoxy resins derived from petroleum-based bisphenol-A; a large offshore blade uses roughly 12–15 tonnes of epoxy. Solar panel backsheets are fluoropolymers — primarily PVF (Tedlar) and PVDF — produced from petroleum feedstocks. Every kilometer of power cable connecting wind farms and solar installations to the grid is insulated with cross-linked polyethylene (XLPE), another petroleum derivative. Polyurethane foam and structural adhesives appear throughout nacelles and blade assemblies. Gearbox lubricants and hydraulic pitch-control fluids require periodic replacement across a turbine's 25-year life.”

3. More details from Claude AI: “The graphite anode is manufactured by calcining petroleum coke at 2,500–3,000°C. Petroleum coke is a refinery by-product; there is no commercially available non-fossil feedstock for synthetic graphite. The electrolyte is organic carbonate solvents derived from petroleum. Lithium is extracted from hard rock mines in Australia, refined in China on a coal grid, at 15–35 tonnes of CO₂ per tonne of lithium depending on route and location. Cell manufacturing — about 20 percent of a battery’s carbon footprint — requires formation cycling consuming 30–50 kWh per kWh of capacity produced, plus continuous dry-room operation. China produces about 85 percent of global cells on a grid still heavily coal-powered.

4. Recycling of materials does not significantly reduce fossil fuel dependency. Claude AI offers details: “Recovering silicon from an end-of-life solar panel requires thermal delamination, acid etching, and re-purification back to solar grade — a process that essentially repeats the most energy-intensive steps of the original manufacturing chain. Currently about 90 percent of end-of-life solar panels go to landfill because it is cheaper. The dominant commercial route for wind turbine blades is feeding shredded material into cement kilns, where the resin burns as supplementary fuel — downcycling, not recovery. Pyrolysis can carry a carbon footprint up to 19 percent higher than landfilling. Solvolysis requires 21–91 MJ per kilogram of blade material, potentially exceeding the energy to produce virgin glass fiber from scratch. Battery recycling recovers cobalt and nickel via high-temperature smelting, but lithium and graphite are lost in pyrometallurgical processing. The dominant grid storage chemistry — lithium iron phosphate — contains no cobalt or nickel, so the economic incentive to recycle it is structurally weak.”

5. I just made that up — JITTI. I like it.

6. Notes on the charts: The historical figure (2000–2024) shows estimated annual CO₂ from manufacturing new wind, solar, and battery capacity: annual additions multiplied by manufacturing CO₂ intensity derived from published LCA experience curves. Sources: Louwen et al. 2016 (Nature Communications); NREL LCA Harmonization 2012/2024; IEA Solar PV Supply Chains 2022; Emilsson & Dahllöf 2019; PNAS Nexus 2023 (Llamas-Orozco et al.); Peiseler et al. 2024 (Nature Communications). Uncertainty on intensity estimates is approximately ±25–30%. The figures cover combustion-derived supply-chain emissions; petrochemical feedstock carbon (epoxy resins, cable insulation, backsheets, lubricants) is not included and likely adds 5–12% to the values shown. The projection figure applies the frozen technology baseline introduced in Pielke, Wigley & Green, Nature 452, 531 (2008) — manufacturing intensity held constant at 2024 levels — to three IEA deployment scenarios: Stated Policies (WEO 2025), Net Zero Emissions, and a current-trajectory extrapolation. Country emissions comparators drawn from EDGAR 2025 and IEA CO₂ Emissions in 2023. Capacity data: IEA-PVPS/IRENA (solar); GWEC/WWEA (wind); IEA Global EV Outlook 2024 (batteries).

7. For some, it may be uncomfortable to realize that the global economy will likely require fossil fuels for the foreseeable future, and perhaps forever. The implication is that to achieve any net zero target will necessarily require advancements in technologies of carbon capture and storage and direct air capture. It would be an interesting exercise to create scenarios of the required magnitude of this fossil fuel foundation under various assumptions of technological progress. My guess is that carbon dioxide reductions of 70-80% from today’s levels would be achievable while maintaining a required fossil fuel foundation, leaving considerable emissions that society would have to choose to remove to achieve a net zero target. But that is just a guess — For a future post!