对话哥伦比亚大学可持续投资中心主任Lisa Sachs：谁在解决真正的绿钢生产难题？需警惕纸面...

来源：市场资讯

　　当前，环境权益证书正成为众多企业青睐的减排工具。但这种交易模式，究竟能否真正推动行业实现实质性减排？围绕这一核心问题，绿色金融60人论坛（GF60）联合新浪财经，与哥伦比亚大学可持续投资中心主任Lisa Sachs教授发起1.5℃ Talk对话。在Sachs教授看来，环境权益证书恰恰折射出当前全球推动难减排行业转型时一个普遍存在的深层误区：我们正在用方法论层面的精巧设计，替代对物理世界真实改变的追求。

　　Lisa Sachs 教授是哥伦比亚大学气候学院可持续投资中心主任，哥伦比亚大学气候金融硕士项目主任。自2008年加入哥大可持续投资中心以来，Lisa Sachs教授创建并领导了该中心的跨学科研究和咨询工作，致力于推动投资法律、政策和实践与可持续发展目标的一致性。她在全球范围内享有盛誉，专注于研究法律、政策和商业实践如何影响国际投资流动，并促进可持续发展。她与各国政府、地区和国际发展机构、金融机构、企业、民间组织以及学术机构密切合作，力求深入理解投资流动与可持续发展的关系，并推动投资政策和实践的改善，以实现SDGs和应对气候变化的巴黎协定目标。

Lisa Sachs 教授

　　以下为Sachs教授的观点全文：

　　今年早些时候，微软宣布达成一项交易：购买瑞典生产商Stegra所生产绿钢对应的环境属性。此举被广泛称赞为“有创新性”，“能够塑造市场”。但实际情况是，这笔交易转移的是“宣称权（claims）”，并没有改变生产本身；它恰恰折射出当下推动绿钢市场规模化时一个更普遍的问题。

　　微软并没有采购绿钢用于在生产地市场中实际使用。它购买的是一类环境属性，从而使其能够将自己在其他地区采购或使用的钢材宣称为“绿钢”，尤其是在那些绿钢生产尚不具备可行性的市场，比如中国（全球很大部分钢铁产量都来自这里）。亚洲钢铁市场的实体系统并没有因此发生变化。没有新增绿钢产能，没有降低电价风险，也没有缓解融资约束。这一采购行为或许让资产负债表变得好看，但并没有改变其映射的底层工业现实。

　　这种情形其实并不陌生。从结构上看，它和碳抵消是同一种逻辑：用一种方法论层面的安排，让企业可以宣称取得进展，却无需改变其所处系统的排放轨迹。就像碳抵消一样，人们对这类做法的赞誉暴露出一个更深的问题：我们已经习惯于奖励那些在方法论上“说得通、算得清”的进展，而不是那些真正改变物理世界的变化。

　　过去几年，围绕绿钢的讨论越来越集中在标准、信息披露、薄记申明（book-and-claim）体系以及买方承诺上。这些工具之所以受欢迎，是因为它们容易辨识，能够整齐地嵌入企业净零目标与可持续发展框架叙事，在单个企业层面核算，也便于企业对外宣布“取得进展”。但问题在于：它们绕开了更艰难、也更关键的问题，即在钢铁生产活动发生的地方，钢铁生产本身有没有变得更清洁、更便宜、更容易融资。

　　简而言之，这些工具并没有触及真正的约束。

　　如果绿钢之所以难以规模化，原因只是买方缺乏传递需求信号的渠道，或者市场对绿钢的定义不够清晰，那么市场本应早已形成。事实并非如此。我们看到的是，尤其在亚洲，绿钢仍主要停留在试点阶段，几乎完全依赖企业资产负债表融资，而项目融资基本缺席。

　　这一区别非常关键。工业市场的扩张，本质上是通过融资实现的。一个项目如果无法以独立项目的方式获得融资，并不意味着行业缺乏雄心，而是因为项目风险和现金流达不到金融机构的承保/授信门槛。

　　薄记申明（book-and-claim）体系对这一点没有任何实质改变。它确实把环境属性与实体供给分离开来，但它并不会降低绿钢生产成本，不会稳定电价，也不会降低氢能投资风险。它做的只是事后重新分配属性，而不是让下一个项目更容易融资、或更低成本建设。

　　钢铁生产并不像太阳能光伏那样，能够沿着成本递减的S曲线快速下行。尽管人们常常用光伏来类比绿钢。太阳能之所以能够规模化，是因为每新增一块组件的安装，都会直接扩大制造规模、触发学习效应、降低单位成本，并在部署发生的同一市场提升竞争力。购买太阳能板，会改变未来生产太阳能板的经济性。

　　绿钢不是这样运作的。它的成本主要受电力市场（价格及其波动性）、资本密集度以及工艺路线配置所主导。从瑞典某项目购买环境属性，并不会降低中国氢基炼铁面临的电价波动风险，也不会改变电网接入规则、调度约束或融资条件。这里不存在一种能够跨市场传播的“靠宣称产生学习效应”。

　　买方承诺也存在类似的误判。人们常把它描述成“缺失的需求信号”，但在实践中，它的作用比这有限得多。在供给侧（生产端）风险基本解决之后，买方承诺可以帮助稳定项目利用率；但它无法让成本结构本身不可行的项目变得可行，也无法说服贷款人去为那些暴露在电价波动和运营不确定性之下的资产提供融资。

　　这并不是纸上谈兵的批评。无论是绿钢还是其他难减排行业，短期限、附条件、与投产时间表脱节的承诺，已经一再被证明无法撬动融资。买方承诺这类工具真正发挥作用的，是在项目经济性已经接近市场出清时，而不是更早之前。

　　那么，为什么我们总是反复将希望寄托到这些工具上来？

　　部分原因在于，我们深陷于“企业个体视角”的思维框架里。标准、目标、披露、转型计划和承诺，问的都是“单个企业靠自己能做什么”。它们与企业问责体系天然契合，却把电力市场、基础设施和产业政策中的政治与制度复杂性，当作外生条件来看待。于是，人们便可以在缺乏系统性变革的情况下，仍然宣称“取得了进展”。

　　但绿钢不是一个企业层面的挑战，而是一个系统性挑战。系统性问题，不会被个体行动者设计的工具所解决。

　　正是受到微软这笔交易的触发，我和同事们决心去识别中国绿钢规模化面临的真正约束。我们梳理了诸多相互交织的风险、约束与激励因素，最终发现其中有一个问题最为关键：氢基炼铁在20至30年的资产寿命周期内，需要巨大且持续的电力投入。即便可再生能源平均电价已经很低，中国工业用户仍然面临价格波动、限电（弃电/限供）风险、拥塞和接入不确定性。真正具有决定性的风险，不是电价“高不高”，而是电价“稳不稳”。

　　当然，波动性风险虽然最关键，却并非唯一风险。它还与技术风险、资本密集度、政策风险等叠加，而这些风险目前大多都要由钢铁企业自己吸收。风险过度集中，最终使绿色炼铁只能停留在谨慎试点阶段。

　　尽管我们完全可以把这些真实的项目风险分离，但现实中我们很少这么做。我们常常笼统地把无法融资的项目称为“高风险”，却不进一步说明：究竟是哪些风险阻断了融资、哪些风险可以通过设定边界或风险共担池来处理、哪些风险则即便获得支持也依然不可行。结果就是，各类粗放式干预不断出现，如泛化的混合融资、普遍性补贴、象征性的需求聚合，它们改善了“观感”，却没有改变融资现实，也没有改变生产结果。

　　这些问题正是我们最近在上海与恒隆地产和GF60共同召集的一场研讨会所聚焦的主题。会议的目标非常明确：推动亚洲绿钢市场发展。讨论的重点并不是标准或承诺，而是经济性。

　　我们讨论了一系列具体方案，核心是如何为绿色炼铁的风险设定边界、通过共担池集中降低风险，从而压降绿钢生产成本。当成本下降、经济性改善、与传统钢材的成本差距逐步收窄之后，可信买方的承购承诺才能真正发挥作用，去解决剩余的利用率风险，进而形成真实的市场需求。

　　同样的问题也出现在其他难减排行业，以及那些尚未真正启动转型的行业中。迄今为止，产业界及其合作方投入了大量精力设计“可以宣称进展”的方法论工具，却没有把重点放在改变物理现实的系统性约束上。

　　这些转型，即便困难重重，并非不可实现。前提是，我们不要再把“可以宣称”误当成“已经改变”。真正的起点应当是经济性、风险与融资。我们手里并不缺处理这些问题的工具，而且仍有很大的创新空间。只要这些核心问题没有被解决，我们就会继续奖励那些看起来很创新的交易，而底层系统本身却几乎原封未动。

　　英文原文如下：

　　Lisa Sachs, Feb 5, 2026

　　Earlier this year, Microsoft announced a deal to purchase the environmental attributes associated with green steel produced by Stegra in Sweden. The transaction was widely praised as innovative and market-shaping. In reality, it shifted claims without changing production, and was emblematic of a broader failure in how we are trying to scale green steel.

　　Microsoft did not buy green steel for use in the market where it was produced. It bought attributes that allowed it to claim "green steel" for steel procured or used elsewhere, specifically in markets where green steel production is not yet viable, like in China, where most of the world's steel is produced. The physical system in Asian steel markets did not change. No new green steel capacity was built. No power price risks were reduced. No financing constraints were eased. The claim improved a balance sheet narrative, not the underlying industrial reality.

　　This should feel familiar. It is structurally the same as offsets: a methodological solution that allows a company to claim progress without changing the emissions trajectory of the system in which it operates. And like offsets, the celebration reveals a deeper problem. We have become accustomed to rewarding what can be claimed methodologically rather than what changes in the physical world.

　　Over the past several years, the green steel conversation has increasingly focused on standards, disclosure, book-and-claim systems, and buyer commitments. These tools are legible. They fit neatly into corporate net zero and sustainability frameworks. They operate at the level of individual firms. And they allow progress to be announced without confronting the harder question of whether steel production itself is becoming cleaner, cheaper, and more financeable in the places where it actually occurs.

　　But these tools are not addressing the binding constraint.

　　If green steel were failing to scale because buyers lacked a way to signal demand, or because definitions were unclear, markets would already be forming. They are not. What we see instead, across Asia in particular, is that green steel remains limited to pilots, almost entirely financed on corporate balance sheets, with project finance largely absent.

　　That distinction matters. Industrial markets scale through finance. When projects cannot be financed on a standalone basis, it is not because ambition is missing. It is because risks and cash flows do not clear underwriting thresholds.

　　Book-and-claim does nothing to change that. It decouples environmental attributes from physical supply, but it does not reduce the cost of producing green steel. It does not stabilize electricity prices. It does not reduce hydrogen exposure. It reallocates attributes after production has occurred. It does not make the next project easier to finance or cheaper to build.

　　Steel production does not resemble the cost-declining S-curve of solar photovoltaics, which is often invoked by analogy. Solar scaled because each incremental panel installed directly expanded manufacturing volume, drove learning effects, lowered unit costs, and improved competitiveness in the same markets where deployment occurred. Purchasing solar panels changed the economics of future panels.

　　Green steel does not work that way. Its costs are dominated by power markets (prices and volatility), capital intensity, and process configuration. Buying environmental attributes from a project in Sweden does not reduce electricity price volatility for hydrogen-based ironmaking in China. It does not change grid access rules, dispatch constraints, or financing conditions. There is no learning-by-claiming effect that propagates across markets.

　　Buyer commitments suffer from a similar miscasting. They are often described as the missing demand signal, but in practice their role is far more limited. Commitments can help stabilize utilization once supply-side risks are largely resolved. They cannot rescue projects with unviable cost structures. They cannot persuade lenders to finance assets exposed to volatile power prices and uncertain operating conditions.

　　This is not a theoretical critique. Across green steel and other hard-to-abate sectors, commitments that are short-tenor, conditional, or disconnected from commissioning timelines have repeatedly failed to unlock finance. Where commitments have mattered, it has been after the economics were already close to market-clearing, not before.

　　So why do we keep returning to these tools?

　　Part of the answer is that we are deeply embedded in entity-level thinking. Standards, targets, disclosure, transition plans, and commitments all ask what an individual company can do on its own. They align with corporate accountability frameworks. They treat the political and institutional complexity of power markets, infrastructure, and industrial policy as exogenous dependencies. They allow progress to be claimed without systems change.

　　But green steel is not an entity-level challenge. It is a systems challenge. And systems challenges do not yield to tools designed for individual actors.

　　Prompted by the Microsoft deal, my colleagues and I were determined to identify the real constraints to scaling green steel in China. We sorted through the many interrelated risks, constraints, and incentives – but one emerged as paramount: hydrogen-based ironmaking requires enormous, continuous electricity inputs over asset lives of 20 to 30 years. Even where renewable generation is cheap on average, industrial users in China face volatile prices, curtailment risk, congestion, and uncertain access. It is the volatility - not the price of electricity- that is the decisive risk.

　　But while the volatility risk is the decisive risk, it is not the only one. It combines with technological risks, capital intensity, policy risks, and others that steel companies must currently absorb. That risk concentration limits green iron production to cautious pilots.

　　Although we can disaggregate the real risks, we rarely do. We often describe unfinanceable projects as "high risk," without specifying which risks block financing, which could be bounded or pooled, and which make a project infeasible regardless of support. The result is a proliferation of blunt interventions (generic blended finance, broad subsidies, symbolic demand aggregation) that improve optics without changing the financing realities or production outcomes.

　　These issues were the focus of a recent workshop we co-convened in Shanghai with Hang Lung and GF60, aimed explicitly at advancing the green steel market in Asia. The discussion did not center on standards or pledges. It centered on economics.

　　We discussed concrete proposals to bound, pool and reduce risks for green ironmaking in order to bring down the costs of green steel production. When the costs come down, the economics improve, closing the cost differential with conventional steel; at that point, credible buyers' offtake commitments can address remaining utilization risk, creating market demand.

　　The same challenge shows up across other hard-to-abate sectors and sectors yet to transition. So far, industry and its partners have focused on designing methodologies that allow progress to be claimed rather than addressing system constraints that change physical reality.

　　These transitions - even the hard ones - are feasible, if we stop confusing claims with change. The starting point has to be economics, risk, and finance. We have many tools to address those and room for further innovation. Until those are addressed, we will continue to reward transactions that look innovative while leaving the underlying system exactly as it was.