A forecast landed this week that looks like good news for anyone who has ever wondered what happens to a composite part after the car it belonged to stops existing. Read it next to the forecast for ordinary, brand-new carbon fiber, and it stops looking like good news at all.
Mordor Intelligence puts the global recycled carbon fiber market at $224.62 million in 2026, rising to $427 million by 2031. Compounded, that works out to roughly 13.7 percent a year. Healthy business. Earlier this year the same research house published its outlook for virgin carbon fiber: 245,370 tonnes in 2026, climbing to 562,770 tonnes in 2031, a compound rate of 18.06 percent.
Those two figures are not in the same units. One is money, the other is mass, and that mismatch is worth noticing on its own. But the growth rates are directly comparable, and they are moving apart. New carbon fiber is entering the world faster than the industry built to deal with it afterward is growing. Every year of that forecast, the gap widens.
That is the story here, and it is not a story about recycling technology. The technology works better than most people assume. What does not exist is a loop.
The Fiber Survives Recycling. The Length Does Not.
Start with what recycling actually does to carbon fiber, because the popular version is wrong in an interesting way.
A Department of Energy-funded study published through OSTI took expired epoxy prepreg scrap reinforced with Toray T800S intermediate-modulus fiber, dissolved the resin chemically, cleaned the fiber and applied fresh sizing. The recovered filaments gave up about 3 percent of their Young’s modulus. Tensile strength and failure strain came back statistically unchanged. The authors’ conclusion was blunt: the inherent filament mechanical and density properties remain intact after recycling.
So the fiber is fine. What happens to it next is not.
The recovered material came out as bundles 6 to 10 millimeters long. After compounding and pelletizing, average fiber length had collapsed to roughly 150 microns. After injection molding, about 100 microns — near enough the width of a human hair.
Now the part that does not make it into a press release. The virgin commercial compound the researchers benchmarked against finished at 93 microns. Same process, effectively the same answer. The twin-screw compounder and the injection molder shred everything to the same length regardless of where the fiber started life.
That cuts both ways, and it is the most useful thing in the dataset. In injection-molded parts, recycled carbon fiber is not a compromise at all. It is genuinely equivalent, because the manufacturing process has already destroyed the one property that made virgin fiber worth its price. Which is exactly why recovered fiber wins brackets, housings, covers and battery-tray reinforcement, and never wins a floor pan. It is not being punished for being recycled. It is being sold into the only applications where its handicap does not apply. Motorsport learned the same lesson from the other direction decades ago, where the value of a composite structure was always in the weave, not the fiber.
The feedstock in that study deserves a second look too. Expired prepreg. Not a crashed car, not a worn-out part — material that aged out in cold storage before it was ever built into anything. A share of the carbon fiber being recycled today never made it as far as being a product.
Almost None of It Comes From Cars
The release names its supply segments plainly: aerospace scrap, automotive scrap, and other. It flags retiring wind turbine blades as a significant emerging feedstock. It credits North America’s lead to established aerospace operations and existing recycling infrastructure, and Europe’s acceleration to regulatory tightening and wind-turbine decommissioning.
Read that again from an automaker’s side of the table. Automotive appears on the demand side — battery trays are the example given — far more than it appears on the supply side.
The scale of the aerospace stream turns up in DOE’s own project records. One prepreg recycling program in Washington State was aimed at nearly 900 tonnes a year of material that had been going to landfill. One region, one industry, one scrap stream. The number of end-of-life cars feeding structural carbon fiber into that same supply is, practically speaking, close to zero, because almost no mainstream vehicle ever had a structural carbon fiber body to begin with — and the one automaker that built its identity on one has since moved on.
Which leaves an awkward arrangement that the word “recycled” papers over nicely. The recovered carbon fiber in a car part almost certainly was never in a car. It was an aerospace offcut, an expired roll, or a wind blade. Automotive is not closing a loop. It is taking delivery at the end of somebody else’s.
And that supply is hostage to somebody else’s efficiency. The same DOE-backed project describes blending roughly 85 percent virgin prepreg with 15 percent recycled, in service of a stated goal of zero-waste composite manufacturing. Aerospace is actively working to stop producing the scrap this market runs on. A recycling industry whose feedstock shrinks when its supplier gets better at its job is not a circular economy. It is a salvage trade with a sustainability slide deck.
Europe Just Wrote Thermosets Out of the Arithmetic
The regulatory picture is where this gets pointed.
The European Council gave its final green light in June 2026 to the Regulation on circularity requirements for vehicle design and management of end-of-life vehicles, after striking a provisional deal with Parliament in December 2025. The consolidated legislative text is dated 8 July 2026, and it retires both the 2000 end-of-life vehicles directive and the 2005 recyclability type-approval directive in one move.
Three provisions matter for composites. Not one of them says “carbon fiber.”
First, the old type-approval bar carries straight over: new vehicle types must be reusable or recyclable to a minimum of 85 percent by mass, and reusable or recoverable to a minimum of 95 percent. Nothing new there — except that it is a mass-based test, and a material without a demonstrated recycling route sits on the wrong side of the ledger while its weight still counts in the total.
Second, and this is the one to circle: the regulation’s recycled-plastic content targets — 15 percent within six years, 25 percent within ten, with a fifth of that drawn from closed-loop end-of-life vehicle material — explicitly exclude thermosets. The text gives its reasoning in the open: thermosets other than seat polyurethane foams and elastomers are not counted in the target calculation because they are particularly challenging to recycle. Epoxy-matrix carbon fiber is a thermoset. Brussels looked at the material and wrote it out of the circularity math on the record.
Third, extended producer responsibility. Producers now finance part of the cost of collecting and treating end-of-life vehicles, and every passenger car and van manufacturer must publish a circularity strategy and refresh it every five years. The company that specified the material helps pay to dispose of it.
Stack those together and an epoxy composite part becomes a material that earns a carmaker nothing toward the recycled-content targets, drags on a mass-based recyclability test, and arrives with a disposal bill attached. Nothing was banned. The accounting was simply made honest, and honest accounting is unkind to composites. It is also a notably sharper standard than the one applied to metal, where genuine closed-loop steel recycling already exists and no target has been set for it yet.
The Energy Case Nobody Is Arguing About
Here is what makes all of this frustrating rather than merely interesting. The environmental argument for recovering carbon fiber is not close.
DOE’s bandwidth study on carbon fiber reinforced polymer composites puts current typical carbon fiber production at 228,800 Btu per pound, or about 458 million Btu per ton of fiber. The companion study on advanced high-strength steel measures blast-furnace ironmaking, its hungriest single step, at 11.72 million Btu per ton, with coking, hot rolling, cold rolling and electric-arc steelmaking adding roughly a dozen more between them. The two studies do not share identical boundaries and were never written to be read side by side. No reasonable adjustment closes a gap that size. Per unit of mass, carbon fiber costs something on the order of twenty times the energy of steel to produce.
Recovery avoids nearly all of it. The OSTI work cites pyrolysis recovery at 15 to 20 percent of the energy needed to make virgin fiber. The Washington prepreg program reported embodied-energy reductions over 90 percent for the recycled fraction, alongside a feasible 15 percent cut in part manufacturing cost. Oak Ridge National Laboratory and NREL have both spent years attacking the problem from different directions, the latter with bio-derivable epoxies that can be reshaped and reused at least three times.
So: a material that is brutally expensive to make, cheap to recover, mechanically intact after recovery, and functionally identical to virgin fiber in the applications it actually ends up in. And the entire global business built on that is forecast to be worth $427 million in 2031, set against a virgin industry producing more than half a million tonnes in the same year.
That is not a technology failure. It is a demand failure, and it happened for an unglamorous reason. Nobody was required to care, and chopped fiber only has so many places to go.
What to Actually Take From This
The forecast is not wrong. Recycled carbon fiber is a real business, it is growing, and it will keep taking share in exactly the parts where short fiber is the correct engineering answer.
But growth is not closure, and the language around composites has gotten sloppy on that distinction. A circular material comes back as itself. Steel does. Aluminum does. Even lithium is getting there, with recovery rates that would have sounded absurd a decade ago. Carbon fiber comes back as a shorter version of itself, drops one rung down the value ladder, and does it once.
A loop that only turns one way is not a loop. It is a chute.
Whether that matters is decided long before anything reaches a recycler. It is decided by how much carbon fiber gets made in the first place, and that number is compounding at 18 percent a year while the cleanup compounds at 14. Recycling capacity is the wrong thing to watch. Specification is the right one.
For owners, the consequence shows up somewhere less abstract than a recycling plant. It shows up at the body shop, where bonded and composite structures are already rewriting the math on what counts as a repairable car. Materials chosen for the factory floor get paid for twice: once in the showroom, and again the day something goes wrong.
Here is the question worth kicking around. If two otherwise identical cars were sitting in front of you, and one was built from materials that could genuinely come back as cars again, would that be worth real money to you at signing — or is recyclability the last thing on anyone’s mind at the dealership? Make the case either way.
Would you trust recycled carbon fiber as much as new in a car you own? Sound off in the comments.