Ongoing research log · global materials · revised as new data lands

Every year the world pulls 106 billion tonnes out of the Earth. This much of it comes back around:

6.9%

Ten years ago the figure was 9.1%. The circular economy is not stalling — measured as a share of everything we consume, it is going backwards. Not because recycling collapsed, but because extraction is winning the race.

Global circularity metric, by report edition

Source: Circle Economy, Circularity Gap Report editions 2018–2025. Editions measure the share of secondary materials cycled back into the global economy.

Reading 01 · the direction of travel

Recycling grew. Consumption grew faster.

Absolute recycling volumes are at an all-time high in almost every stream. The circularity rate still fell — because the denominator exploded. In the six years to 2024 the world consumed more material than in the entire 20th century.

106 Gt

Material extracted and consumed globally each year — tripled in five decades, on course for roughly 160 Gt by 2060 without intervention.

6.9%

Share cycled back. Down from 9.1% a decade ago. Virgin extraction rose ~3× faster than secondary supply.

>55%

Share of global greenhouse-gas emissions attributable to resource extraction and processing — plus ~40% of health-related air-pollution impacts.

11.8%

EU circular material use rate (2023) against a 23.4% target for 2030. The world's most ambitious regulator is running at half pace.

The gap isn’t widening because recycling failed. It’s widening because we never stopped digging.
Reading 02 · the material ledger

Eight streams, eight different stories.

“Circularity” is not one market. It is eight economies with wildly different physics, margins and clock speeds. Averaging them hides everything that matters. Read each on its own terms: the past decade, the position now, and the decade to 2035.

Fe

Steel & iron

Quietly winning
2015 – 2024

The most successful circular material on Earth, largely unnoticed. Around 1.89 Gt of crude steel a year, with roughly 630–680 Mt of scrap consumed annually. Construction steel recovery exceeds 85% in mature markets.

Now

Electric arc furnaces run near 29% of global output — but only ~10% in China, which makes over half the world's steel. China's 2000s building boom is only now becoming scrap.

2026 – 2035

Scrap becomes both the constraint and the prize. The hidden ceiling is copper contamination from shredded vehicles and appliances. Investment case: sorting and de-tramping, not new furnaces.

Al

Aluminium

Strong, supply-limited
2015 – 2024

Recycled metal settled at roughly a third of global supply. Remelting uses about 5% of the energy of primary smelting — the single best energy arbitrage in materials.

Now

Demand forecast to climb around 40% by 2030 on packaging, EVs and grid build-out. Scrap simply cannot keep pace; hundreds of billions of beverage cans are still landfilled each year.

2026 – 2035

The unglamorous high-ROI play: alloy-level sorting (LIBS/XRF), closed-loop automotive scrap contracts, and can-to-can systems. Physical assets, commodity-backed, 20-year lives.

Cu·Li·Ni

Copper & critical minerals

Overbuilt now, essential later
2015 – 2024

Secondary sources supply roughly a third of total copper use. Battery recycling attracted enormous capital on the assumption that EV feedstock would arrive by 2025.

Now

It didn't. The EV fleet is too young; most packs are still on the road. Announced li-ion recycling capacity vastly exceeds available black mass, and over 80% of it sits in China.

2026 – 2035

The IEA finds recycling could cut primary supply requirements for copper, lithium, nickel and cobalt by roughly a quarter to 40% by 2050 — but secondary supply only becomes material after 2030. Survive the valley, own the decade.

PET·PP

Plastics & packaging

Falling behind
2015 – 2024

Roughly 9% of all plastic ever produced has been recycled. Waste reached 353 Mt in 2019 and is on track to nearly double by 2040. Recycled content in new plastic sits near 6% globally.

Now

Voluntary targets have visibly broken. The Global Commitment's 2025 goals will be missed; reuse is stuck near 1–2% of packaging. Two rounds of UN treaty talks — Busan 2024, Geneva 2025 — closed without agreement.

2026 – 2035

Regulation replaces pledges. The EU packaging regulation applies from August 2026 with binding recycled-content minima to 2030. Europe becomes the de facto global plastics rulebook; compliant PET capacity gets a demand floor.

TEX

Textiles & fibre

Structurally broken
2015 – 2024

Global fibre production hit roughly 124 Mt in 2023, heading for ~160 Mt by 2030. Less than 1% of clothing is recycled fibre-to-fibre. The ~8% “recycled” share is almost entirely bottle-derived polyester — a downcycle, not a loop.

Now

Circulose (Renewcell) filed for bankruptcy in early 2024 with a working plant and no long-term offtake. The clearest signal in the sector: supply arrived years before brand demand.

2026 – 2035

Separate textile collection became mandatory across the EU in 2025; ecodesign rules and Digital Product Passports land from 2027. Sorting infrastructure — not chemistry — is the binding constraint.

WEEE

Electronics & e-waste

Going backwards
2015 – 2024

62 Mt generated in 2022, up 82% since 2010 — five times faster than documented recycling. Only 22.3% is formally collected and treated.

Now

Around $91bn of recoverable metals sits inside the annual stream. On current trends volumes reach 82 Mt by 2030 while the collection rate falls toward 20%.

2026 – 2035

The EU right-to-repair directive must be in national law by July 2026, and repairability scores are spreading. Repair and refurbishment — not shredding — is where the value and the health case both sit.

CDW

Cement, concrete & sand

The biggest blind spot
2015 – 2024

The built environment consumes 30–40% of all materials. Sand and gravel — around 50 Gt a year — are the second most-extracted resource on the planet after water.

Now

Headline recovery rates look high because demolition rubble is crushed into road base. Almost none of it re-enters structural concrete, and reusable steel sections are still routinely melted.

2026 – 2035

Largest tonnage, lowest technology risk, least capital. Material passports, design-for-disassembly, structural reuse markets and calcined-clay cements. This is where a 2030 target is actually won or lost.

BIO

Food, nutrients & biomass

Huge, under-capitalised
2015 – 2024

1.05 billion tonnes of food wasted in 2022 — about 19% of everything reaching consumers, and 8–10% of global emissions. Households account for the majority.

Now

Nutrient loops are barely closed at all: phosphorus is a finite, geopolitically concentrated resource that we mine, eat and flush.

2026 – 2035

Anaerobic digestion with contracted feedstock, nutrient recovery from wastewater, and organics collection in fast-urbanising cities. Modest returns, extraordinary system value.

Reading 03 · the capital map

Where the money works, where it burned, where it never went.

A decade of circular investing has produced a clean pattern. Assets with contracted offtake and a regulated demand floor perform. Assets built on a sustainability narrative — however good the technology — do not.

Lane A

Working, with healthy returns

  • Metals recovery & advanced sortingCommodity-backed cash flows, existing offtake, 15–30 year assets. Aluminium and steel scrap upgrading is the most boring, most reliable trade in the sector.
  • Recommerce & resaleAsset-light with real unit economics — Vinted, Back Market, ThredUp, brand-run programmes like Levi's SecondHand and Patagonia Worn Wear. Margin comes from logistics and trust, not technology.
  • Repair & remanufacturingCaterpillar Reman, Michelin retreading, Renault's Refactory at Flins. Reman lines routinely beat new-build margins on the same part.
  • Regulated packaging recycling in the EUExtended producer responsibility plus binding recycled-content minima create a price floor that private contracts never could.
  • Waste-to-value infrastructureVeolia's growth plan targets a ~30% increase in plastics recycling capacity; hazardous waste and water reuse carry the returns.

Common factor: someone is legally or contractually obliged to buy the output.

Lane B

Capital ran ahead of demand

  • Textile-to-textile chemistryRenewcell built the plant, proved the pulp, and went insolvent in February 2024. Brands praised it publicly and bought at pilot volumes.
  • Battery recycling, 2021 vintageLi-Cycle halted its flagship Rochester hub and later sought creditor protection; Umicore paused its Nysa battery-materials investment. The feedstock curve is a 2030 story, funded on a 2025 timetable.
  • Advanced / chemical recycling of mixed plasticsRepeated schedule slips, yields below prospectus, and unsettled mass-balance accounting rules.
  • Consumer reuse platformsLoop and similar refill ventures retrenched — returns logistics are brutal without policy support.

The lesson is not that the technology was wrong. It is that voluntary brand commitments are not offtake.

Lane C

Missing the attention entirely

  • Reuse and refill systemsStill under ~2% of packaging worldwide. Needs standardised containers, shared washing infrastructure and EPR fee modulation that rewards reuse. Almost no capital.
  • Structural reuse in constructionSteel section reuse, material passports, deconstruction over demolition. Biggest tonnage on Earth; a rounding error in climate-tech funding.
  • The informal sectorRoughly 20 million waste pickers collect much of what actually gets recycled globally — and receive a negligible share of investment or formal recognition.
  • Nutrient & organics recovery in the Global SouthWhere waste volumes are growing fastest and infrastructure is thinnest.
  • Product data infrastructureDigital Product Passports are legally due from 2027. Very few supply chains can currently produce the underlying data at all.
  • Repairability in medical, industrial & agricultural equipmentSoftware locks and parts monopolies quietly destroy more recoverable value than any consumer category.

Unfashionable, unbranded, high-tonnage. This is where the next decade's alpha sits.

Reading 04 · the 2030 stake

Circularity is a public-health project wearing an industrial coat.

The reason to close loops before 2030 is not tidiness. Extraction, processing and uncontrolled disposal are among the largest measurable causes of human illness on the planet — and they are the one lever that acts on climate, biodiversity, pollution and health at the same time.

~40%

Share of health-related particulate-matter impacts driven by resource extraction and processing (UNEP International Resource Panel, 2024).

7 million

Premature deaths a year attributed to ambient and household air pollution (WHO).

18 million

Children estimated by WHO to work in informal waste processing, including e-waste sites — exposed to lead, mercury and dioxins.

$249 bn

Estimated US health costs in a single year from just three plastic-associated chemicals (Minderoo-Monaco Commission on Plastics and Human Health).

−39%

Potential cut in global greenhouse-gas emissions under Circle Economy's modelled circular scenario, alongside a ~30% reduction in material use.

−28%

Reduction in virgin material extraction needed simply to bring consumption back within a safe planetary operating space.

Every tonne we don’t dig up is a tonne of dust that never reaches a lung.
Reading 05 · what 2026–2035 costs

The bill, and who should be holding it.

Published cost estimates are patchy and inconsistent — but the ones that exist agree on the shape: the circular pathway is cheaper than the linear one over any horizon longer than a decade. The barrier is not total cost, it is who pays first.

  • $600 bn

    Cumulative capital investment to 2040 for the plastics system-change scenario modelled by Pew and SYSTEMIQ — against roughly $2.5tn for business as usual, with lower whole-system costs.

  • 2× by 2030

    What the EU must do to its circular material use rate — 11.8% today, 23.4% target. Meeting it implies a step-change in construction and textile recovery, not more kerbside bins.

  • $91 bn / yr

    Value of metals already flowing through the e-waste stream. Capturing even half changes the economics of critical-minerals supply outright.

  • $4.5 tn

    Widely cited estimate of circular-economy economic opportunity to 2030. Treat it as a direction of travel, not a forecast — but note that nothing close to it has been deployed.

Where a top brand should actually put its weight

1. Sign offtake, not pledges. A ten-year purchase agreement at a floor price does more for a recycler than a decade of joint press releases. Renewcell and Li-Cycle failed for the want of exactly this.

2. Design for the second life at the first drawing. Mono-material packaging, no adhesives on fasteners, published disassembly instructions, spare parts at reasonable prices for a decade.

3. Build the data now. Digital Product Passports are legally due from 2027. Material composition data is the hardest, slowest part — start three years early.

4. Fund the parts of the chain you don't own. Sorting plants, reverse logistics, and the informal collectors who already deliver your recycled feedstock.

5. Stop revising targets quietly. Publish the miss, publish the reason. Credibility, once spent on a rollback, is not recoverable.

Reading 06 · brands in motion

Who is actually doing something — and what it teaches.

Notable corporate action, including the retreats. The rollbacks of 2024–2025 are as instructive as the wins: they mark exactly where voluntary ambition met an infrastructure that did not exist.

Consumer electronics

Apple

Reports well over half of the material in its products as recycled or renewable, with commitments to 100% recycled cobalt in Apple-designed batteries, recycled rare earths in magnets and recycled tin, gold and aluminium — backed by its own disassembly robotics.

InsightVertical integration lets one company create its own recycled-material market. Almost nobody else can. The industry-wide lever is repairability, not robots.

Automotive

Renault Group

The Refactory at Flins — Europe's first plant dedicated to vehicle circularity — plus The Future Is NEUTRAL, a dedicated closed-loop materials business targeting substantial circular-economy revenue by 2030.

InsightCircularity structured as a profit-and-loss with its own balance sheet outperforms circularity structured as a sustainability programme.

Beverages

The Coca-Cola Company

In late 2024 restated its packaging goals: a collection ambition of roughly 70–75% for bottles and cans by 2035 and 35–40% recycled material, replacing earlier 100%-by-2030 language.

InsightThe most honest data point in the sector. Collection cannot be bought by one brand; it requires deposit systems and EPR that companies must actively lobby for.

FMCG

Unilever

Cut virgin plastic use materially against its 2019 baseline, then in 2024 rebased its targets — a lower virgin-plastic reduction, on a longer timeline, with reusable/recyclable design goals pushed toward 2030.

InsightSachet economies in emerging markets remain unsolved. Redesigning the format beats optimising the polymer.

Home & furniture

IKEA (Inter IKEA)

Targeting a fully circular business by 2030: buy-back and resale, free spare parts, and a majority of materials already renewable or recycled across the range.

InsightSpare parts given away for free are the cheapest lifetime-extension programme in retail — and the least copied.

Building materials

Holcim

Recycles millions of tonnes of construction and demolition material annually through its ECOCycle technology, with recycled aggregates in ready-mix concrete lines.

InsightThe only heavy-industry player treating rubble as feedstock at scale. If cement's circular share moves, the global number moves.

Mobility & materials

Michelin

Roughly a third of tyre material is already recycled or bio-sourced, with a stated route to 100% sustainable materials by 2050 and a long-standing retreading business for truck fleets.

InsightRetreading is circularity that was profitable before it was fashionable. Look for these legacy loops in every industry.

Lighting & health tech

Signify · Philips

Both report circular revenues as a headline financial metric — light-as-a-service, refurbished medical systems, take-back and parts harvesting on imaging equipment.

InsightOnce circularity is a reported revenue line, it survives budget cycles. Turning it into a KPI the CFO owns is the real structural change.

Batteries

CATL · Brunp

Operates among the world's largest battery-recycling systems, reporting very high recovery rates for nickel, cobalt and manganese and steadily improving lithium recovery.

InsightChina built the recycling capacity next to the cell plants. Western projects built it next to the venture capital. Geography of feedstock decides the winner.

Environmental services

Veolia

Its current strategic plan puts billions into growth areas including a substantial expansion of plastics recycling capacity, hazardous waste treatment and water reuse.

InsightThe utility model — long contracts, regulated volumes — is the only proven way to finance sorting infrastructure at national scale.

Flooring

Interface

Carbon-negative carpet tile lines and a long-running take-back programme, with recycled and bio-based content designed in from the yarn upward.

InsightTwo decades of consistency beat a decade of ambitious targets. Circularity compounds only if the strategy outlives the CEO.

Resale platforms

Vinted · Back Market

Second-hand fashion and refurbished electronics marketplaces operating at genuine scale and, in Vinted's case, profitability — without owning a single factory.

InsightThe fastest-growing circular businesses of the past decade were not recyclers. They were marketplaces that made the second life easy.

Reading 07 · sources & further research

The reference shelf.

Everything above traces back to these. Grouped for onward research rather than listed in order of appearance — this log is meant to be built on.

How to read this log. Figures are quoted from the primary sources listed above and carry their own vintages — circularity metrics lag by one to three years, and company disclosures are self-reported. Where an estimate is contested or a range, it is described as such in the text rather than rounded into false confidence. Corrections, better data and counter-evidence are welcome: this is a working document, not a verdict.