The Inevitability of Zero-Cost Stewardship

Why Information Protects Nature at 10²⁰ Times Less Cost Than Force—And Why This Changes Everything

The Inevitability of Zero-Cost Stewardship Why Information Protects Nature at 10²⁰ Times Less Cost Than Force

And Why This Changes Everything By Jed Anderson Founder & CEO, EnviroAI Fourth in the Series: AI and the Environmental Profession

Updated April 2026 “The system already has the energy to reconfigure itself. It just does not know where to send it.

Information provides the direction.”

EnviroAI | Houston, Texas | The Thesis This paper makes a claim that will strike many as radical:

The marginal cost of environmental protection is converging toward zero.

This is not policy advocacy. It is not technological optimism. It is the inevitable consequence of two physical laws approaching their limits:

1. The cost of information is falling toward the Landauer Limit

2. The cost of energy is transitioning to nuclear density

When both converge, environmental protection ceases to be a cost center and becomes what GPS and timekeeping already are: a background utility of civilization.

The work is not merely changing. It is disappearing.

And for those of us who entered this profession for love of nature rather than love of timesheets, this should be cause for celebration.

Part I: The Ontological Correction

What Pollution Actually Is The first step toward understanding this transition is correcting a category error.

Pollution is not a material problem. It is a configuration problem.

A molecule of benzene in a sealed tank is an asset. The same molecule dispersed in groundwater is a liability. The atoms are identical. Only their arrangement and location differ.

Physics has a precise term for this: entropy—the measure of disorder in a system.

When matter is concentrated, ordered, and localized, it has low entropy. When dispersed, disordered, and uncertain, it has high entropy.

Pollution is simply entropy increase. Valuable matter moved from ordered states to disordered states.

Environmental protection is entropy decrease. Sorting. Restoring order. Returning atoms to useful configurations.

This reframing changes everything. Because entropy reduction has known physical costs—and those costs have floors.

The Conservation of Matter EnviroAI | Houston, Texas | Earth approximates a closed system for matter. Atoms are neither created nor destroyed; they are rearranged. A carbon atom in atmospheric CO₂ is physically identical to a carbon atom in diamond. The distinction between “resource” and “pollutant” is not intrinsic to the atom. It is entirely a function of configuration and location.

Pollution is disordered wealth.

The Second Law The Second Law of Thermodynamics dictates that entropy increases spontaneously. Pollution is this law in action—the mixing of byproducts into the biosphere along the path of least resistance. To reverse mixing requires work. The governing relationship is Gibbs Free Energy:

ΔG = ΔH − TΔS For pollution (mixing), ΔS > 0 and ΔG < 0. The process is spontaneous.

For remediation (sorting), ΔS < 0 and ΔG > 0. The process requires external work.

The cost of a clean planet reduces to two inputs:

1. Energy (W)—the work to overcome mixing

2. Intelligence (I)—the information to apply that work precisely

The Equivalence of Entropy Boltzmann defined physical entropy: S = k₈ ln W Shannon defined information entropy: H = −Σpᵢ ln pᵢ

These differ only by a constant. They are the same phenomenon measured in different units. A highentropy system is one where we lack information about the location of its particles. Pollution is missing information.

If we knew the trajectory of every SO₂ molecule leaving a smokestack, capture would require precise actuation, not brute filtration. Environmental order is an information processing problem. To reduce physical entropy, we must reduce informational uncertainty. This leads to the floor.

Part II: The Bond-Bit Asymmetry

The Molecular Floor: 268× Before examining the operational ratio, we establish the unchallengeable bedrock. At the singlemolecule level, the ratio between the cost of moving one bond and the cost of knowing one bit is set by two measured physical constants:

EnviroAI | Houston, Texas | The energy to break one chemical bond (O–H bond in water): 7.71 × 10⁻¹⁹ Joules

The energy to process one bit of information (Landauer limit at 300K): 2.87 × 10⁻²¹ Joules

E_bond / E_bit = 7.71 × 10⁻¹⁹ / 2.87 × 10⁻²¹ ≈ 268 Knowing is 268 times cheaper than moving, at the single-molecule level.

This number is set by the laws of physics—the Landauer limit sits at the thermal fluctuation scale, chemical bond energies sit at the quantum mechanical binding scale—and it will never change. It is as permanent as the speed of light.

The bond energy is fixed by the fine-structure constant (α ≈ 1/137.036), one of the fundamental constants of the universe. It was the same in 1900, is the same today, and will be the same in 3000.

There is no Moore’s Law for the fine-structure constant.

The Twenty Orders of Magnitude The 268× molecular-floor ratio drastically understates the macroscopic reality. At the operational scale of real environmental events, the leverage explodes because of a fundamental feature of nature: information compresses. You do not need to know the position of every molecule to prevent a catastrophe. You need macro-state information—a few billion bits about valve degradation—that prevents micro-state disaster involving trillions of trillions of molecular bonds.

Consider a chemical storage tank with a failing valve:

Scenario A: Mass Forcing (After the Fact)

The valve fails. Chemical disperses into soil and groundwater. To remediate requires excavation, pumpand-treat systems, chemical oxidation, breaking and reforming molecular bonds.

Energy requirement: ~10⁵ Joules per mole of contaminant (the energy scale of chemical bonds)

Scenario B: Entropic Shepherding (Before the Fact)

A sensor detects micro-vibrations indicating valve degradation. A signal is sent. The valve is closed or replaced before failure.

Energy requirement: ~10⁻¹⁵ Joules (the energy of a digital signal in current computing)

The ratio: 10²⁰ Twenty orders of magnitude. One hundred quintillion to one.

This is not an approximation. These are the actual energy scales at which chemistry and computation operate.

EnviroAI | Houston, Texas | Information Substitutes for Energy In the old paradigm, environmental protection meant work—physically moving matter, breaking bonds, pumping fluids, treating waste. Work operates at the energy scale of chemistry: electron-volts, kilojoules per mole.

In the new paradigm, environmental protection means information—knowing where matter is, predicting where it will go, intervening before entropy cascades begin. Information operates at the energy scale of computation: approaching 10⁻²¹ Joules per bit.

We are substituting bits for bonds.

Every increment of better sensing, better prediction, better real-time control shifts work from the expensive regime (chemistry) to the cheap regime (information).

Part III: Maxwell’s Demon Was Always Building

In 1867, James Clerk Maxwell imagined a tiny being—a “demon”—that could observe individual gas molecules and selectively open a door between two chambers, sorting fast molecules from slow ones.

For 158 years, physics treated this as a paradox about the Second Law of Thermodynamics.

But in solving the paradox, we missed what the demon was doing.

The demon started with a gas at uniform temperature. It ended with a temperature gradient—hot on one side, cold on the other. A new configuration that did not previously exist.

The demon did not prevent anything from scattering. The demon TRANSFORMED the system.

It took matter in one configuration and navigated it to a different configuration using information instead of force. It created order that did not previously exist. It assembled a new state.

Maxwell’s Demon was never about guarding. It was always about building.

For 158 years, we focused on the paradox and missed the blueprint.

The Three Modes of Informational Stewardship The Bond-Bit Asymmetry applies not just to prevention. It applies to every mode of environmental intervention:

Prevention—keeping a system at its current configuration. Know the valve is failing. Close it. The spill never happens.

Restoration—guiding a damaged system back to order. Model the groundwater flow. Inject at one point—the right point—and the aquifer’s own current carries the remedy to the plume. The water does the work. Information tells it where.

EnviroAI | Houston, Texas | Reconfiguration—navigating a system to a new, more ordered state. Remove one culvert—the right one—and the wetland rebuilds itself. The water follows gravity. The vegetation follows the water. The ecosystem self-assembles around the restored hydrology.

The pattern is always the same. The system already has the energy to reconfigure itself. It does not know where to send it. That is what information provides. Not force. Direction.

This is Maxwell’s Demon realized at planetary scale. Environmental Superintelligence is not a guard. It is a shepherd. It does not wait for disorder and force matter back into place. It knows the state of the system and steers it—with minimal energy, at the right moment, in the right direction.

Part IV: The Two Falling Curves

Curve 1: Intelligence Approaching the Landauer Limit In 1961, physicist Rolf Landauer established the theoretical minimum energy required to process information:

E_min = k_B × T × ln2 At room temperature (300K), this equals approximately 2.9 × 10⁻²¹ Joules per bit.

This is not an engineering estimate. It is a consequence of the Second Law of Thermodynamics. No technology, no matter how advanced, can process information for less energy than this. It was experimentally verified by Bérut et al. (Nature, 2012) to within experimental precision.

Current state: Modern computing operates at approximately 10⁻¹² Joules per operation.

The gap: We are currently 10⁹× above the theoretical floor—one billion times less efficient than physics permits.

This gap is closing. Koomey’s Law observes that computational efficiency doubles approximately every

2.3 years. As architectures evolve—neuromorphic, optical, quantum, eventually reversible—we slide down toward the Landauer Limit.

Implication: The energy cost of “knowing”—sensing, modeling, predicting, deciding—is converging toward the thermodynamic floor. The intelligence required to monitor every valve, model every flow, track every atom, optimize every process is becoming energetically trivial.

Curve 2: Energy Approaching Nuclear Density EnviroAI | Houston, Texas | Civilization currently runs primarily on chemical energy: breaking carbon-hydrogen bonds releases approximately 4 electron-volts per reaction.

We are moving from chemical energy (atom surface) to nuclear (core).

Chemical (Fossil): Breaking C–H bond releases ~4 eV.

Nuclear (Fusion/Solar): Fusing hydrogen releases ~17.6 million eV.

The Gap: Nuclear physics is 4 million times more energy-dense. As energy production shifts to nuclear fission, fusion, and solar (which is fusion at 93 million miles), the marginal cost of energy approaches the cost of infrastructure amortization alone. Energy transitions from scarce commodity to abundant utility.

The Divergence The practical leverage ratio grows every year. It has done so for 75 years. It will continue until the

Landauer limit is reached.

The cost of knowing falls exponentially. The cost of moving stays fixed by quantum mechanics. The curves diverge monotonically. They can never converge. There is no Moore’s Law for chemistry. There is no Moore’s Law for the fine-structure constant. There is no Moore’s Law for the cost of being wrong.

Part V: The Convergence

What Happens at the Limits When both curves approach their physics floors:

Input Current State Physical Floor Current Gap Intelligence ~10⁻¹² J/operation ~10⁻²¹ J/bit 10⁹×

Energy ~$0.05/kWh ~$0.01/kWh 5× Bond-Bit Ratio 268× 268× Fixed by physics (molecular)

Bond-Bit Ratio ~10¹⁰ ~10²⁰ 10⁹ room to grow (operational)

The labor cost of environmental protection (humans reading, writing, analyzing, deciding) is automated away. This is already happening.

The hardware cost (sensors, monitors, infrastructure) follows learning curves downward and is increasingly replaced by “virtual sensors”—inference from existing data streams.

EnviroAI | Houston, Texas | What remains is the irreducible thermodynamic cost of physical entropy reduction—and that cost is far lower than what we currently spend on labor and hardware combined.

Environmental protection becomes a background utility. This is not speculation. This is the physics playing out.

Part VI: The Inverted Mountain

Why Every Step Is Cheaper Than the Last If the cost of knowing falls while the cost of moving stays fixed, a powerful consequence follows: the return on investment for each incremental advance toward Environmental Superintelligence monotonically increases.

This is the Inverted Mountain Theorem. The journey toward Environmental Superintelligence is not a mountain that gets harder to climb as you ascend. It is an inverted mountain where each step upward is cheaper than the last, the ROI accelerates at every stage, and the summit approaches zero cost.

The Six Camps Camp Year Capability Technology Base Camp 2018 Document intelligence Environmental document platform

Camp 1 2023 Environmental chatbot LLM + RAG over 11M documents Camp 2 2025 Permit automation Agentic AI workflows

Camp 3 2032 Dynamic permitting Real-time air & water Camp 4 2040 Predictive prevention Entropic shepherding

Summit 2045 Env. Superintelligence Background utility The ROI Acceleration Does each step up the mountain yield a higher return than the last? We compute the investment required to advance from one camp to the next and the incremental annual savings generated per facility:

Step Investment New Savings/yr ROI (Yr 1) Trend Base Camp → Camp 1 $105,000 $165,000 1.6×—Camp 1 → Camp 2 $75,000 $310,000 4.1× ↑ Rising

Camp 2 → Camp 3 $45,000 $535,000 11.9× ↑ Rising EnviroAI | Houston, Texas | Camp 3 → Camp 4 $24,000 $307,000 12.8× ↑ Rising Camp 4 → Summit $9,000 $125,000 13.9× ↑ Rising

The ROI accelerates monotonically from 1.6× to 13.9×. Every step costs less to take and saves more than the previous step.

Why this is necessarily true: at each camp, a greater fraction of environmental work shifts from the

“moving atoms” regime (constant cost) to the “knowing about atoms” regime (falling cost). Because computation costs fall while chemistry costs remain fixed, each successive camp benefits from a wider cost gap than the previous one. This is not a contingent economic trend. It is a necessary consequence of the divergence between Koomey’s Law and the fine-structure constant.

The most expensive thing we can do is stay where we are.

Part VII: The Work Is Disappearing

A Professional Confession I have spent 27 years in the environmental profession. I have billed thousands of hours. I have helped write permits, compliance reports, impact assessments, audits, and applicability determinations.

And I must tell you the truth:

Most of that work existed because we lacked information.

We monitored because we could not predict. We remediated because we could not prevent. We documented extensively because we could not verify in real time.

The work was a tax on ignorance—the friction cost of operating without sufficient intelligence.

As intelligence approaches Landauer and energy approaches nuclear density, that friction disappears.

Permits become real-time continuous compliance verification. Reports become automated data streams. Assessments become predictive models that prevent harm before it occurs. Monitoring becomes ubiquitous, embedded, invisible.

The work does not evolve into different work. It evaporates into infrastructure.

The Three Phases Phase 1: Labor Substitution (Now–2035) AI agents replace human labor in documentation, analysis, and compliance tracking. The “Paperwork Layer” of environmental management is automated.

EnviroAI | Houston, Texas | Phase 2: Shepherding Dominance (2035–2055) Real-time sensing and AI-driven process control shift the balance from mass forcing to entropic shepherding. The economic calculus flips: it becomes irrational to wait for disorder when knowing prevents it at 10⁻¹⁵ the cost.

Phase 3: Background Utility (2055+) Environmental protection becomes embedded in industrial infrastructure. The marginal cost of compliance approaches the marginal cost of computation—which approaches the Landauer Limit.

Part VIII: The Legacy Question

We Are Mortal This brings us to the question that matters.

We will not live forever. Our careers will end. Our expertise, accumulated over decades, will eventually be lost—unless we encode it somewhere durable.

The question is not whether this transition will happen. The physics is inexorable.

The question is whether we participate in building it—or watch from the sidelines while it is built without us.

Option A: Bill hours until retirement. Resist the change. Watch the profession hollow out. Leave behind a career of timesheets and paperwork.

Option B: Spend the next decade encoding everything we know—our understanding of ecosystems, regulations, ethics, and judgment—into systems that will protect the planet for centuries. Leave behind a legacy.

We have one window. One moment in history where human environmental expertise can be transferred into machine intelligence. One chance to imbue these systems with our values.

The Sisyphus Question For 50 years, the environmental profession has operated on the implicit assumption that our job is to push the boulder up the hill forever. To hold back entropy indefinitely through continuous human effort.

This is Sisyphus. It is exhausting. It is ultimately futile. And it was never the real goal.

The goal was never to protect nature forever through human effort.

The goal was to build the system that would.

That system is now being constructed. The physics permits it. The technology enables it. The only question is whether we—perhaps the last generation of environmental professionals who understand both the old world and the new—will be its architects.

EnviroAI | Houston, Texas | ## Part IX: The New Role

What We Do Now If the work is disappearing, what remains? The answer is: everything that matters.

From Force to Shepherding: Stop fighting entropy after it wins. Start knowing where it’s headed and intervening gently before it scatters. Value-based pricing aligns compensation with outcomes rather than time.

From Paperwork to Principles: Codify the first principles of stewardship—entropy minimization, precaution, biodiversity—into the algorithms. AI lacks moral framework. We provide it. This is the most important work of our careers.

From Monitoring to Mentoring: Train the next generation of AI by curating high-quality data and contextual knowledge. Every edge case we solve, every judgment call we document, becomes training data for systems that will operate long after we retire.

From Compliance to Co-Creation: Work with industry and regulators to build the infrastructure—the planetary nervous system—that allows environmental protection to become a ubiquitous utility.

The Paradox of Obsolescence Here is the paradox: by making ourselves obsolete, we become more essential than ever.

The next decade is the critical window. The systems being built now will shape planetary stewardship for the next century. They can be built with our wisdom or without it. They can encode our ethics or operate without ethical grounding. They can reflect 50 years of hard-won environmental knowledge or start from scratch.

We are not optional. We are the bridge.

But only if we choose to walk across it.

Conclusion: The Thermodynamic Equilibrium A clean planet is not a political choice.

It is not primarily a moral aspiration.

It is the thermodynamic equilibrium of a civilization with sufficient intelligence and abundant energy.

EnviroAI | Houston, Texas | When the cost of knowing approaches the Landauer Limit, and the cost of energy approaches nuclear abundance, the cheapest path for any industrial system is the clean path. Pollution becomes economically irrational—not because of regulations or values, but because knowing is 268 times cheaper than moving at the molecular floor and 10²⁰ times cheaper at operational scale.

We are approaching that threshold.

The mountain is inverted. Every step toward Environmental Superintelligence costs less than the last and delivers more. The ROI accelerates from 1.6× to 13.9× across six stages. The summit—far from being the most expensive destination—approaches zero cost.

The system already has the energy to reconfigure itself. It just does not know where to send it.

Information provides the direction.

The work is disappearing. The mission is succeeding.

This is not the end of environmental protection. It is the beginning of environmental immunity.

And we—if we choose—can be the architects.

You can’t compete with free.

And the most expensive thing we can do is stay where we are.

EnviroAI | Houston, Texas | Appendix: Verification of Key Claims Claim Value Source

Landauer Limit 2.87 × 10⁻²¹ J/bit at 300K Landauer (1961); k_B × T × ln2; Bérut et al.

(2012)

Current computing efficiency ~10⁻¹² J/operation IEEE literature on CMOS Gap to Landauer ~10⁹× 10⁻¹² ÷ 10⁻²¹

O–H bond energy 7.71 × 10⁻¹⁹ J (464 CRC Handbook kJ/mol)

C–H bond energy 6.86 × 10⁻¹⁹ J (413 CRC Handbook kJ/mol)

Molecular floor ratio 268× 7.71×10⁻¹⁹ ÷ 2.87×10⁻²¹ Bond-Bit leverage (operational) ~10²⁰ 10⁵ ÷ 10⁻¹⁵ (see derivation)

Nuclear fission energy ~200 MeV per U-235 IAEA Koomey’s Law ~2.3 year doubling Koomey et al. (2011); updated 2023

Sagawa-Ueda verification 90% of theoretical max Koski et al., PNAS (2014)

Inverted Mountain ROI range 1.6× to 13.9× Per-facility calculation (this paper)

All figures represent order-of-magnitude values for the purpose of illustrating the fundamental asymmetry. Specific applications will vary.

EnviroAI | Houston, Texas | April 2026 The goal was to build the system that would.

EnviroAI | Houston, Texas |