Three materials sit inside almost everything we touch: the bottle, the window, the shirt, the blade of a wind turbine. All three are made by heat, and heat is where their carbon lives. This is a working benchmark of what is genuinely on track, and what still needs calling out — with every number traceable back to its source. 三种材料几乎存在于我们触碰的一切之中：瓶子、窗户、衣物、风机叶片。它们都由高温制成，而碳排放正藏在这份高温里。本文是一份工作型基准研究：哪些已在轨道上，哪些仍需被点名——每一个数字都可回溯到出处。
Plate 01 · Working reference 工作底稿
Plastic, glass and fibre are usually studied apart — a chemicals report, a packaging report, a textiles report. Read apart, each looks like a slow-moving special case. Read together, they share the same three bottlenecks: high-temperature heat, feedstock that begins as fossil carbon, and a collection system that decides whether the material ever comes back.
塑料、玻璃与纤维通常被分开研究——化工报告、包装报告、纺织报告。分开看，每一个都像行动迟缓的特例；合起来看，三者共享同样的三个瓶颈：高温热源、源自化石碳的原料，以及决定材料能否回流的回收体系。
That is the whole reason to build one roadmap instead of three. A decision about clean electricity in 2028 changes a glass furnace, a steam cracker and a polyester spinning line at the same time. The lever is shared; only the vocabulary differs.
这正是要做“一张路线图”而非“三张”的理由。2028 年关于清洁电力的一个决策，会同时改变玻璃窑炉、蒸汽裂解装置与聚酯纺丝线。杠杆是共用的，不同的只是各自的行话。
The carbon is not in the material. It is in the heat, the molecule, and the missing return trip.碳不在材料本身，而在于热、分子，以及那趟缺失的“回程”。
Order-of-magnitude figures, each traceable to a public source listed in section 09. Ranges are given where methodologies disagree — that disagreement is itself part of the finding.
以下为数量级数据，均可回溯至第 09 节公开来源。方法学存在分歧之处以区间表示——分歧本身也是研究结论的一部分。
OECD puts plastics at roughly 3.4% of global greenhouse gases in 2019, on a path toward ~4.3 Gt by 2060 if nothing changes. About 90% of that carbon is bound up in production and conversion, not disposal.
OECD 测算 2019 年塑料约占全球温室气体的 3.4%，若无改变，2060 年或达约 43 亿吨。其中约九成来自生产与加工环节，而非废弃处置。Ref 01, 02
Smaller in absolute terms — and that is exactly why it is tractable. Melting runs at 1,500–1,600 °C; every extra 10% of cullet (recycled glass) in the batch cuts furnace energy by about 2.5% and CO₂ by about 5%.
总量较小，也正因如此更可攻克。熔制温度达 1,500–1,600 ℃；配合料中每增加 10% 碎玻璃，窑炉能耗约降 2.5%，CO₂ 约降 5%。Ref 05, 06
Synthetic fibre is now ~2/3 of global fibre volume, and polyester alone dominates growth. Recycled polyester carries roughly 30–45% lower cradle-to-gate emissions than virgin — but almost all of it comes from bottles, not from textiles.
合成纤维已占全球纤维产量约三分之二，增量主要来自聚酯。再生聚酯的摇篮到大门排放较原生低约 30–45%——但其原料几乎全部来自瓶片，而非旧衣。Ref 07, 08
Glass furnaces are rebuilt roughly every 12–15 years. Steam crackers and polyester lines run longer. That arithmetic is the single most important fact on this page: between 2026 and 2035 most assets get one rebuild decision, and that decision locks in emissions until the 2040s.
玻璃窑炉大致每 12–15 年冷修一次，裂解装置与聚酯生产线周期更长。这一算术是本页最重要的事实：2026–2035 年间，多数资产只会遇到一次重建决策，而这次决策将把排放锁定到 2040 年代。
Batch and cullet optimisation, oxy-fuel and hybrid furnaces, waste-heat recovery, lightweighting. Lowest capex, fastest payback, and still not fully harvested anywhere.
配合料与碎玻璃优化、全氧/混合燃烧窑炉、余热回收、轻量化。投资最低、回收最快，但在任何地区都尚未被充分挖掘。
Design-for-recycling as a specification, not a pledge: mono-material packaging, removal of hard-to-sort additives, deposit-return build-out, textile-to-textile collection pilots at city scale.
把“可回收设计”写进技术规格而非承诺：单一材质包装、去除难分选添加剂、押金回收体系建设、城市级衣物到衣物回收试点。
All-electric and hybrid melting where the grid allows; electrified process heat and heat pumps for drying, dyeing and finishing. Viability tracks the power-to-gas price spread and the carbon price, not the technology.
在电网条件允许处推进全电熔与混合熔制；干燥、染整环节采用电加热与热泵。可行性取决于电–气价差与碳价，而非技术本身。
Verified product carbon footprints, digital product passports, chain-of-custody for recycled content. Cheap, unglamorous, and the precondition for every premium that follows.
可核验的产品碳足迹、数字产品护照、再生成分的监管链。成本低、不起眼，却是后续所有绿色溢价的前提。
First commercial-scale enzymatic PET recycling, hydrogen and 100% electric furnace trials, bio-based and CO₂-derived monomers moving from pilot to FID.
首批商业规模酶法 PET 解聚、氢能与全电窑炉试验、生物基及 CO₂ 基单体从中试走向投资决策。
The molecule itself changes: waste-, bio- and CO₂-derived carbon replaces a material share of virgin naphtha. This is the step that cannot be retrofitted — it needs new plants on new sites.
分子本身发生改变：废弃碳、生物碳与 CO₂ 基碳替代相当比例的原生石脑油。这一步无法“加装”，需要新厂址上的新装置。
Electric crackers and near-fully-electric melting at scale, running on grids that are majority clean. Asset-cycle bound: mostly the second rebuild, not the first.
电裂解与规模化近全电熔制，依托以清洁电力为主的电网。受资产周期约束：多发生在第二次冷修，而非第一次。
Carbon capture on the concentrated, unavoidable streams — glass batch carbonates, cracker off-gas — plus permanent storage or utilisation.
对浓度高、不可避免的排放流（玻璃配合料碳酸盐分解、裂解尾气）实施碳捕集，并配合永久封存或利用。
Closed loops become the default commercial model: material services, take-back contracts, and secondary material priced as a strategic input rather than a waste discount.
闭环成为默认商业模式：材料即服务、回收合约，二次材料被定价为战略投入品，而非“废料折扣”。
Regions do not differ in physics. They differ in what electricity costs, what carbon costs, and who is legally obliged to take the material back.
各区域的物理规律相同，差异在于：电价多少、碳价多少，以及法律上由谁负责把材料收回来。
Indicative ranges assembled from public industry and agency material (Ref 01–14). Treat them as a shape, not a quote — actual returns swing with power price, carbon price and offtake certainty. Where a return depends on a carbon price or a mandate, that dependency is named.
以下区间综合自公开行业与机构资料（参考 01–14），仅代表趋势形状而非报价——实际回报随电价、碳价与承购确定性大幅波动。凡回报依赖碳价或强制要求者，均已注明。
Three groups matter, and they matter in this order: whoever buys the material sets the direction, whoever makes it sets the pace, and whoever funds it sets the ceiling. Capital follows a signed offtake — it very rarely leads one.
三类角色都重要，且顺序如下：买材料的人决定方向，造材料的人决定速度，出钱的人决定上限。资本跟随已签署的承购协议，极少先行。
Circular-economy and climate funds — Closed Loop Partners, Circulate Capital, Breakthrough Energy, the Alliance to End Plastic Waste, the Fashion Climate Fund, and large transition vehicles from Brookfield, TPG and Temasek — are the visible pools. What distinguishes the good ones is not ticket size but tolerance for 8–12 year horizons.
循环经济与气候基金是主要资金池。真正优秀者的区别不在于单笔金额，而在于能否容忍 8–12 年的回报周期。Ref 15–20
Saint-Gobain, Owens Corning, Ardagh, Şişecam, AGC and NSG are the names to watch in glass and fibre, because each is deciding what a furnace looks like on its next rebuild. In polymers, the consumer-goods buyers — the packaged-goods and apparel groups with public recycled-content commitments — are what pulls capacity into existence.
玻璃与纤维领域应关注即将冷修的制造商；聚合物领域，真正拉动产能落地的是公开承诺再生成分比例的消费品与服装采购方。Ref 21–26
SBTi and CDP decide what a target means. Textile Exchange, FEVE and the packaging regulations decide what recycled content means. The UN plastics treaty negotiations decide whether any of it becomes global. Attention follows these bodies more reliably than it follows press releases.
目标由谁定义、再生成分由谁认定、规则能否全球化——这些机构比新闻稿更能预示资金与注意力的流向。Ref 27–31
Everything above compresses into five. They are ordered by leverage per dollar, not by ambition.
以上内容可压缩为五项。排序依据是“每一元投入的杠杆效应”，而非雄心大小。
Cullet, flake and reclaimed fibre. The cheapest tonne of CO₂ in all three industries, limited by collection and sorting rather than by technology or capital.
碎玻璃、瓶片与再生纤维。三个行业中最便宜的减碳吨位，瓶颈在回收与分选，而非技术或资金。
Electrification, hybrid and hydrogen melting, heat pumps for wet processing. Governed by the asset rebuild cycle — this lever has a closing door, not an open one.
电气化、混合与氢能熔制、湿加工热泵。受资产冷修周期支配——这扇门正在关闭，而非敞开。
Bio-based, waste-derived and CO₂-derived carbon; enzymatic depolymerisation. The genuine 2035–2050 structural shift, and the one most often over-promised for 2030.
生物基、废弃碳与 CO₂ 基碳，以及酶法解聚。这是真正属于 2035–2050 的结构性变革，却最常被提前承诺到 2030。
Product carbon footprints, digital passports, chain-of-custody. Almost free relative to plant capex, and the single most consistent blocker on green premiums today.
产品碳足迹、数字护照、监管链。相对于工厂投资几乎零成本，却是当前绿色溢价最一致的堵点。
Binding recycled-content rules, carbon pricing, public procurement and long-term offtake. Without this fifth lever, the first four stay pilots forever.
强制再生成分比例、碳定价、公共采购与长期承购。没有第五根杠杆，前四根将永远停留在试点阶段。
Four of the five are engineering. The fifth is a promise to buy — and it is the one that keeps failing.五者之中有四项是工程问题，第五项是一句“我会买”的承诺——而恰恰是它一再落空。
The point of a benchmark is to be uncomfortable in public. Here is where the evidence currently sits.
基准研究的意义在于公开地令人不适。以下是当前证据所指向的位置。
Is the reduction claimed at the plant gate, or across the product's life? A 90% cut inside the fence can be a 12% cut in the world.
减排是在厂界内计算，还是覆盖产品全生命周期？厂界内减 90%，在真实世界里可能只有 12%。
Under mass balance, can the same green molecule be sold twice? Ask for the chain-of-custody rule by name.
在质量平衡法下，同一个“绿色分子”是否被售卖两次？请对方点名说明所采用的监管链规则。
Does the business case survive a grid that is 60% clean, not 100%? Most electrification models quietly assume the latter.
当电网清洁度是 60% 而非 100% 时，商业模型是否依然成立？多数电气化测算默认了后者。
Is this tied to a scheduled rebuild, or bolted on between cycles? Bolt-ons rarely reach final investment decision.
该方案是否与既定冷修/大修周期绑定，还是在周期之间临时加装？临时加装很少走到最终投资决策。
When the grant or pilot funding ends, who buys the output at a price that clears? If the answer is unnamed, it is a pilot, not a plan.
当补贴或试点资金结束后，谁以可覆盖成本的价格采购产出？若答案无名有姓不上，那就仍是试点，而非计划。
Every link below is a primary or institutional source. Where this page gives a range, the range comes from disagreement between these documents — go to the source before you act on any figure.
以下均为一手或机构来源。本页给出的区间，正源自这些文件之间的分歧——在依据任何数字行动前，请回到原始出处。
Reading note 阅读提示 — Figures on this page are order-of-magnitude and drawn from public documents current to the time of writing; payback and IRR ranges are indicative shapes, not offers or forecasts. Open the source before quoting a number. 本页数据为数量级估算，取自撰写时的公开文件；回收期与内部收益率区间仅为趋势示意，不构成报价或预测。引用前请打开原始来源。
The question was never only plastic, glass and fibre. It is whether a material industry can move faster than its own replacement cycle — and the honest answer today is that it moves exactly as fast as someone is willing to buy the result.
问题从来不只关乎塑料、玻璃与纤维，而在于一个材料产业能否跑赢自身的更新周期。今天诚实的答案是：它的速度，恰好等于有人愿意为结果买单的速度。
Next in this line of work: trace one furnace rebuild, one recycled-content contract and one product passport end to end, and see where the tonne is actually gained or lost.
这条研究线的下一步：完整追踪一次窑炉冷修、一份再生成分合约、一张产品护照，看清那一吨碳究竟在哪里被赢得或失去。
Open the raw sources above · 打开上方原始资料