Matter does not leave.

In 1789 Antoine Lavoisier established that matter is neither created nor destroyed in a chemical reaction: what goes in comes out, in the same quantity, rearranged. Einstein’s 1905 result extended the principle rather than replacing it — mass and energy are interchangeable, and the total is conserved.

Applied to an economy, that has one unavoidable consequence. Every ton of material a process takes in leaves it again, as product, as emission, as effluent, as heat, as something buried. There is no “away”. A second law of thermodynamics argument completes it: the energy that did the work is degraded in doing it, so putting the material back where it came from always costs more than letting it out.

Neither law was hidden. Both were known long before the industrial economy was designed, and the economy was designed as though neither applied.

This was said to economists in 1969, in their own journal. Robert Ayres and Allen Kneese built Production, Consumption, and Externalities on what they called the fundamental law of conservation of mass, and drew the conclusion that follows from it: externalities are not occasional failures at the edges of otherwise efficient markets. They are normal, pervasive and inherent in production, because the mass balance requires them. A price that omits them is not a slightly imperfect price. It is the wrong number.

That is the false price the practice exists because of. For two centuries we designed, engineered and financed the physical world on it: materials, buildings, process heat, transport, agriculture, the cities themselves. Every one of those decisions was optimized against a cost base that left out what the physics guaranteed would happen.

In 1974 James Lovelock and Lynn Margulis published the Gaia hypothesis in Tellus: the Earth’s atmosphere is not a passive backdrop but a system held far from chemical equilibrium by the biosphere itself. Oxygen at 21%, methane co-existing with it, surface temperature stable across billions of years of rising solar output — none of that is what a dead planet’s chemistry produces.

The hypothesis attracted a serious objection, and it is worth stating rather than skipping: critics argued that it implied purpose, since natural selection acts on organisms rather than on planets. Lovelock’s reply was a model, Daisyworld, showing that regulation can emerge from ordinary competition with no foresight at all.

What survived is the part that matters here. The idea that the Earth is a coupled system of biological, chemical and physical processes with feedbacks — that it can be measured, modeled and pushed — became Earth system science. Planetary boundaries are its descendant, which is why section 05 can state how many of the nine have been crossed rather than argue about whether limits exist.

For an allocator the consequence is narrow and practical. A regulated system has operating ranges, and processes that hold it there can weaken. Those feedbacks are the infrastructure under every asset in a portfolio, and they are the part no balance sheet records.

This section exists because a bench specialist read the draft and pushed back, and the push-back was right.

The substantive point

For most of industrial history, a material’s properties came from what it was made of. Steel behaves like steel because of its chemistry. That constraint is what made the redesign in section 03 look like a chemistry problem.

Architected materials — metamaterials — break that assumption. Their behavior comes from geometry and topology at a scale below the part, which is why the field describes them as deriving function from structure rather than composition. The consequences are real and already demonstrated: negative Poisson’s ratio, materials that contract when heated, mechanical cloaking, programmable stiffness that can be switched electrically, and lattices tuned to trap or steer energy.

A composite built from two ordinary positive-expansion materials can be architected to have negative or near-zero thermal expansion — a property neither constituent possesses. That is a genuine expansion of what can be built, and it matters for exactly the things this practice assesses: structures that do not move with temperature, batteries and enclosures that survive cycling, isolation that is directional.

The correction that has to travel with it

The paper behind the claim describes mechanical metamaterials that go beyond Newton’s third law of motion — the rule that every action has an equal and opposite reaction — not the third law of thermodynamics, which concerns entropy as temperature approaches absolute zero. The two are unrelated, and the source’s own title says Newton.

It matters for a second reason. In those systems the non-reciprocity is produced by active feedback control: a controller sits over each mass, measures its neighbors, and actuates. The physics literature is explicit that a material violating this kind of reciprocity must be dissipative, driven or active — it needs an energy source or a momentum sink. Active metamaterials work because they carry local reservoirs of energy, which is precisely why they can do what passive, energy-conserving materials cannot.

So nothing here repeals conservation of mass or the laws of thermodynamics. The energy is still paid for. What changes is the menu of behaviors available to a designer who is willing to pay for it, and that menu is now much longer than chemistry alone allowed.

Why this is an opportunity, and for whom

• A new category exists that is genuinely early. Most of this work sits between laboratory demonstration and manufacturable part. That is the position where a technically literate allocator can take a real position before the category is priced — and it is the exact stage where our assessment is built to operate.
• It widens the redesign in section 03. If waste is a design failure, architected materials give designers a second lever: change the geometry rather than the chemistry. Zero-expansion structures, self-isolating assemblies, lattices that recover energy, and components made from fewer material classes are all easier to keep in a loop.
• The constraints are manufacturing, durability and power. A lattice that performs in a lab at centimeter scale may not survive fabrication, fatigue or twenty years of weather, and an active material stops behaving when its power does. Those are the questions, and they are ordinary technical readiness questions.

The diligence rule it produces

A claim that a material “beats” a physical limit is the oldest pitch in the sector, and it is now easier to make because the vocabulary is unfamiliar. Three questions separate the real thing from the pitch, and they will be added to the technical readiness scoring notes:

1. Which law is actually being invoked? Newton’s third law, Maxwell-Betti reciprocity, and the laws of thermodynamics are different claims. A team that cannot say which one is describing a result it does not understand — and we caught exactly that confusion in a secondary source while writing this section.
2. Where does the energy come from? If the behavior requires activity, there is a power budget, a controller, a failure mode and an operating cost. If a team claims the behavior is passive and free, that is a finding.
3. What survives scale-up? The property has to persist through manufacturing tolerances, fatigue and field conditions, not only in the demonstration cell.

Credit where it belongs. This section exists because Lawrence Culliford read a draft and said the argument as written was true for naturally occurring chemistry and incomplete for engineered composites. That is what the bench is for, and it is why the review happens before publication rather than after.

The false price has a second half. The first is that waste was treated as free. The second is that nature left intact was treated as worth nothing — a stock with no return until it is cut, drilled or cleared.

The economics already exists

John Krutilla broke that assumption in 1967, in Conservation Reconsidered, published in the American Economic Review one year before the Wild and Scenic Rivers Act. His argument was that development comes at the opportunity cost of preservation, and that cost-benefit analysis as then practiced was structurally biased toward development because it counted only the first. He introduced what became existence value — that people hold value for a thing that continues to exist, whether or not they ever use it — and he made the decisive point about irreversibility: a dam can be removed, a canyon cannot be rebuilt, and the technology that makes the development valuable will probably improve while the asset destroyed will not be replaced. He then applied it with Charles Cicchetti and Anthony Fisher to the case for preserving Hells Canyon on the Snake River.

Three ideas descend from it, and all three are underwriting ideas rather than environmental ones:

• Option value — what it is worth to keep a choice open when the future is uncertain (Weisbrod, 1964).
• Quasi-option value — the value of delaying an irreversible decision until better information arrives (Arrow and Fisher, 1974).
• Irreversibility — an asymmetry that ordinary discounting handles badly, because the loss is permanent while the gain is not.

Any allocator recognizes these. They are the language of real options, and they were applied to landscapes twenty years before they were applied to venture portfolios.

The standing-forest comparison

The same argument was made empirically in 1989, when Charles Peters, Alwyn Gentry and Robert Mendelsohn valued one hectare of Amazon forest at Rio Nanay in Peru three ways and published it in Nature. Sustainable harvest of fruit, latex and timber returned a net present value several times that of clearing the same hectare for cattle.

The 1989 study has been argued with for thirty-five years — over market saturation, transport costs, and whether one accessible hectare represents a basin — and the argument is healthy. What has not been overturned is the structure of the finding. Later work has extended rather than reversed it:

Why this belongs in a diligence practice

Four things follow, and each is something we assess rather than admire.

• The comparison is a discount-rate argument. Extraction pays now; a standing forest pays a smaller amount annually and indefinitely. Which wins depends entirely on the discount rate and the time horizon — the same variables that decide whether a climate position is early or wrong.
• Irreversibility belongs on the risk register. A converted forest, a drained aquifer, a collapsed fishery: the option is not recoverable at any price. That is not a moral claim, it is an asymmetry a risk register can carry.
• The people who live there are part of the economics, not a footnote. Food security, tenure and Indigenous knowledge are inputs to whether a standing-forest model produces anything at all, and a business case built without them is a business case that fails in the field.
• This is the same question as the rest of the practice. Who is compelled to pay for the forest standing, on what budget line, on what date? Where nobody is, it is a timing bet, and it should be sized as one — however good the argument is.

The oldest argument against this whole agenda is that it is a tax on prosperity: cleaner means dearer, and regulation costs jobs. It has been tested.

It was contested from inside the labor movement. Environmentalists for Full Employment, founded in 1975 and directed by Richard Grossman, existed to challenge the claim that environmental protection destroyed jobs, publishing Jobs & Energy in 1977. The argument was completed in Fear at Work: Job Blackmail, Labor and the Environment (Kazis and Grossman, 1982), which documented how the threat of job loss was used to block protections rather than predict outcomes.

It was contested from inside business strategy. Michael Porter and Claas van der Linde argued in 1995 that well-designed environmental standards can trigger innovation that partially or fully offsets compliance costs — the Porter hypothesis. It remains argued over. What is not argued over is that the opposite claim, that standards straightforwardly destroy competitiveness, has never been established either.

And the cost curves moved. Lazard’s 2025 analysis puts utility-scale solar’s levelized cost down about 84% since 2009, and finds unsubsidized wind and solar the cheapest new-build generation in the United States for the tenth consecutive year, while gas-fired generation reached a ten-year high. IRENA reports utility-scale solar and onshore wind at roughly $40 per MWh globally in 2025, less than half the cost of new combined-cycle gas, and firm solar-plus-storage at $54–82 per MWh, down from over $100 in 2020.

What this means for the work

We do not underwrite a green premium, and we do not underwrite a green discount. Every assessment asks whether a cost case survives without a subsidy that may not be there in three years, because a company whose economics depend on the trade-off argument being resolved in its favor is carrying a policy bet it has not named. What the fifteen-year cost record does establish is narrower and more useful: the direction of travel is not a matter of belief, and a technology that is expensive today is not thereby expensive in 2032. That is a timing judgment, and timing is what we assess.

Three strands of one technique converge on this practice.

Option pricing. Black and Scholes gave finance a way to value a right rather than a thing. What matters here is the logic: value depends on volatility, time and irreversibility, not only on expected cash flows.

Preservation as an option. Krutilla’s existence value, Weisbrod’s option value and Arrow and Fisher’s quasi-option value (section 06) apply that logic to a landscape. Keeping a forest standing is holding an option. Cutting it exercises one, permanently, at a strike price nobody computed.

Business-value modeling. The founding principal co-founded the Microsoft Business Value team, which applied off-the-shelf financial instruments — total cost of ownership, net present value, internal rate of return and Black-Scholes option analysis — to enterprise technology decisions, and built the methodology behind multi-billion dollar Enterprise Agreement revenue. That work answered a question with the same shape as this one: what is a capability worth when its benefits are diffuse, deferred and not yet on anyone’s ledger?

Where this goes next. The practice’s forward agenda is an instrument of the same kind for closed-loop and circular businesses: a model in which the costs a company currently externalizes — because carrying them would hit reported shareholder value — are priced on the timetable in section 05, and in which an avoided cost is stated with the rigor of a revenue line. Nobody should claim that model exists yet. The materials for it do: mass balance, option value, ecosystem accounting, disclosure timetables, and the habit of writing down what would change our view.

Not for this practice — for the field it works in.

• The missing middle gets funded. The distance between a technology that works and a first commercial deployment is where good companies die correct and early. It is fundable when timing is assessed rather than hoped for.
• Timing is priced. A position where nobody is yet compelled is a timing bet. Timing bets are legitimate; they should be sized as timing bets, by name, in the investment memo.
• The method spreads. It is published rather than protected, because a method that is argued with is worth more than one that is merely owned. If a competitor copies it and a better company gets funded, the thesis wins.
• The phrase becomes redundant. When the costs are carried, “the business case for nature” is just the business case.

We serve the allocators who move first. Family offices, funds, corporate venture arms, LPs and developers who are early on the adoption curve and prepared to hold a longer clock than a fund life. They carry the risk of being right and early, which is the risk this practice was built to assess.

The work is delivered to the founders. In most engagements the allocator pays and the company receives the work. Founders are who it is for, not who pays for it.

The field benefits either way. A company that should not be funded and is not funded is a good outcome: the capital goes to one that should be, and the category does not learn the wrong lesson from an avoidable failure. A category that keeps mistaking timing failures for technology failures gets less capital every cycle.

Stated because the argument above attracts three misreadings, and each would cost us the audience that matters.

• This is not a quantum argument. Quantum mechanics governs the atomic scale; it does not prescribe how to design a materials loop, a building or a supply chain. The case rests on conservation of mass, the second law, and industrial ecology, all of which a sceptical engineer can check.
• This is not an ESG score. We do not rate companies against a sustainability taxonomy. We assess whether a specific company reaches a compelled buyer before its capital runs out, and unpriced costs enter that assessment only where they change the timing or the cost case.
• This is not valuation, audit or assurance. We do not value natural capital, certify an inventory or issue an opinion a lender can rely on. We assess what the arrival of those costs does to one company’s economics, and we say what would change our view.