Marché mondial du traitement et du raffinage de l'éthanol 2025-2033

Marché mondial de la transformation et du raffinage de l'éthanol Segmentation

• Distillation et rectification Distillation traditionnelle et avancée pour l'éthanol de qualité carburant et industrielle.
• Technologies de déshydratation Tamis moléculaires, déshydratation azéotropique et déshydratation par membranes pour l'éthanol de haute pureté.
• Filtration & Purification Charbon actif, échange d’ions et filtration céramique pour l’élimination des impuretés.
• Valorisation des sous-produits Le DDGS (drèches de distillerie séchées), la lignine et la gestion des autres coproduits.

• Éthanol à base de maïs La matière première la plus répandue en Amérique du Nord.
• Éthanol à base de canne à sucre Principal contributeur en Amérique latine et dans certaines régions d'Asie.
• Biomasse cellulosique et lignocellulosique ??? Comprend les résidus de culture, le bois et les herbes ; segment à croissance rapide.
• Mélasse et autres sous-produits de fermentation Niche mais précieux pour la production spécialisée d’éthanol.

• Mélange de carburants E10, E15, E85 et autres mélanges pour essence.
• Applications industrielles Solvants, peintures, revêtements et produits chimiques ménagers.
• Applications pharmaceutiques et de santé Désinfectants, antiseptiques et éthanol de haute pureté pour formulations pharmaceutiques.
• Applications de boissons Distilleries et brasseries utilisant de l'éthanol raffiné pour la production de spiritueux et de liqueurs.

• POET LLC
• Green Plains Inc.
• Valero Energy Corporation
• ADM (Archer Daniels Midland Company)
• Royal Dutch Shell
• BP Biocarburants
• Ra??zen S.A.
• Cosan Limited

• Moteur de prévision de la demande Pheonix module les expansions de capacité en fonction de l'adoption technologique, de la disponibilité des matières premières et des mandats réglementaires.
• Analyse automatisée des cinq forces de Porter suggère un pouvoir de fournisseur modéré, avec une concurrence croissante et une innovation continue dans les technologies de purification.
• Tendances en matière de durabilité inclure la récupération d'énergie, la réduction de la consommation d'eau et l'utilisation circulaire des sous-produits.

• Le traitement et le raffinage de l’éthanol sont essentiels pour fournir de l’éthanol de haute pureté dans les segments des carburants, de l’industrie et de la pharmacie.
• Les innovations technologiques améliorent les rendements et réduisent les coûts tout en préservant la qualité.
• L’intégration avec la production d’éthanol en amont et les marchés de matières premières garantit continuité et efficacité dans l’ensemble de la chaîne de valeur.
• Les pratiques de raffinage durables et la valorisation des coproduits renforcent la valeur économique et environnementale.

Table des matières

Synthèse de direction
1.1. Aperçu du marché (2025-2033)
1.2. Aperçu du marché et tendances clés
1.3. Recommandations des analystes
1.4. Importance stratégique de l'éthanol dans les applications énergétiques, industrielles et de bioéconomie à l'échelle mondiale
1.5. Pheonix AI Perspectives stratégiques

Aperçu du marché
2.1. Définition et portée du marché mondial de l'éthanol
2.2. Aperçu de la production et du traitement de l'éthanol (fermentation, distillation, déshydratation, purification, valorisation des sous-produits)
2.3. Dynamique du marché
2.3.1. Facteurs moteurs
2.3.2. Contraintes
2.3.3. Opportunités
2.3.4. Défis
2.4. Impact des objectifs de durabilité et des politiques de transition énergétique
2.5. Pheonix AI Insights : Paysage futur de la production et de la consommation

Écosystème de l’industrie de l’éthanol
3.1. Chaîne d'approvisionnement en matières premières (maïs, canne à sucre, mélasse, biomasse cellulosique, résidus lignocellulosiques, autres sous-produits)
3.2. Fournisseurs de technologies et fabricants d'équipements
3.3. Opérateurs d’usine de production
3.4. Partenaires de distribution et de logistique
3.5. Secteurs en aval consommateurs d'éthanol (carburant, industrie, agroalimentaire et boissons, produits pharmaceutiques, produits chimiques, bioénergie)

Segmentation du marché ??? par technologie de traitement
4.1. Distillation et rectification — distillation traditionnelle et avancée pour l’éthanol de qualité carburant et industrielle
4.2. Technologies de déshydratation — Tamis moléculaires, déshydratation azéotropique et par membrane pour l’éthanol de haute pureté
4.3. Filtration et purification — Charbon actif, échange d'ions et filtration céramique pour l'élimination des impuretés
4.4. Valorisation des sous-produits — Drêches de distillerie séchées (DDGS), lignine et gestion des autres coproduits

Segmentation du marché ??? Par type de matière première
5.1. Éthanol à base de maïs — la matière première la plus répandue en Amérique du Nord
5.2. Éthanol à base de canne à sucre — principal contributeur en Amérique latine et dans certaines parties de l’Asie
5.3. Biomasse cellulosique et lignocellulosique ??? Comprend les résidus de cultures, le bois et les graminées ; segment à croissance rapide
5.4. Mélasse et autres sous-produits de fermentation — un créneau, mais précieux pour la production spécialisée d’éthanol

Segmentation du marché ??? Par application d’utilisation finale
6.1. Mélange de carburants — E10, E15, E85 et autres mélanges d'essence
6.2. Applications industrielles — Solvants, peintures, revêtements et produits chimiques ménagers
6.3. Applications pharmaceutiques et de santé ??? Désinfectants, assainisseurs et éthanol de haute pureté pour formulations de médicaments
6.4. Applications dans les boissons — Distilleries et brasseries utilisant de l’éthanol raffiné pour les spiritueux et les liqueurs

Analyse du marché régional
7.1. Amérique du Nord
7.1.1. États-Unis
7.1.2. Canada
7.2. Amérique latine
7.2.1. Brésil
7.2.2. Argentine
7.2.3. Reste de l’Amérique latine
7.3. Europe
7.3.1. Allemagne
7.3.2. France
7.3.3. Royaume-Uni
7.3.4. Reste de l’Europe
7.4. Asie-Pacifique
7.4.1. Chine
7.4.2. Inde
7.4.3. Japon
7.4.4. Reste de l'Asie-Pacifique
7.5. Moyen-Orient et Afrique

Paysage concurrentiel
8.1. Analyse des parts de marché des principaux acteurs
8.2. Cartographie de la capacité de production et de la localisation des usines
8.3. Développements stratégiques (expansions, partenariats, fusions, acquisitions)
8.4. Profils des entreprises (jusqu'aux 15 principaux acteurs)
Société Archer Daniels Midland
POET, LLC
Green Plains Inc.
Raízen S.A.
??? Valero Renewable Fuels Company LLC
??? Ressources Flint Hills
D'autres

Prévisions et perspectives du marché (2025 ? 2033)
9.1. Perspectives du volume de production mondiale d'éthanol (en milliards de litres)
9.2. Perspectives de la valeur du marché mondial de l'éthanol (en milliards USD)
9.3. Projections de croissance régionale
9.4. Scénarios de modélisation prédictive de Pheonix AI

Cadre réglementaire et politique
10.1. Directives mondiales en matière d'énergies renouvelables
10.2. Incitations et mandats de production d'éthanol au niveau national
10.3. Réglementations commerciales et impacts des droits de douane

Tendances émergentes et innovations
11.1. Production d'éthanol de deuxième génération (2G) et avancé
11.2. Innovations en enzymes et bioprocédés
11.3. IA, IoT et optimisation des processus
11.4. Technologie de distillation, de déshydratation et de conversion des déchets en éthanol

Recommandations stratégiques
12.1. Stratégies d’entrée pour les nouveaux entrants sur le marché
12.2. Stratégies d’expansion de capacité pour les acteurs existants
12.3. Diversification vers les coproduits (DDGS, lignine, CO₂, biodiesel)
12.4. Pôles d'investissement par région et technologie

Annexe
13.1. Abréviations et glossaire
13.2. Méthodologie de recherche
13.3. Sources de données
13.4. Références de Pheonix AI Research

Paysage concurrentiel

The competitive landscape of this step is not the competitive landscape of ethanol, and confusing the two is why the category is usually misdescribed. Ethanol itself is a fragmented commodity industry: the largest single producer in the United States holds roughly 17 per cent of national capacity, no firm holds a double-digit share of the world market, and more than a hundred single-plant operators compete on a price they do not set.

The processing step is the opposite. It is supplied by a small engineering field into that fragmented buyer base, and the concentration sits in the equipment and the adsorbent rather than in the molecule.

Three layers that do not compete with each other

Tier one: full-scope process licensors. A handful of firms will design and guarantee an entire plant — fermentation through distillation, dehydration and stillage. Their number is small enough to count on two hands, and their share is regional rather than global: one firm dominates Indian molasses and grain work, another the European wheat and beet complex, and two or three the North American dry mill. None holds a majority of world licensed capacity, which is why this market is properly described as moderately consolidated rather than concentrated.

Tier two: unit-operation suppliers. Distillation internals, reboilers, centrifuges, evaporators, dryers, molecular sieve units, corn-oil recovery systems, and the control layer. These firms compete on performance per unit and on reference installations, and several are divisions of large process-equipment conglomerates that treat ethanol as one vertical among many. Their pricing power depends on whether the buyer is buying a train or a single exchanger.

Tier three: consumables and service. Molecular sieve beads, activated carbon, ion-exchange resin, membranes, caustic, gaskets, and the maintenance labour that keeps an energy-intensive step running. This is the only layer with an annuity, and it is where the durable revenue in the step actually sits.

The barrier is a reference plant, not a patent

No licensor wins a project without a reference on that feedstock. A firm with twenty corn plants and no cane reference will not be invited to bid in Brazil, and vice versa. That makes competition ferocious at the point of first entry and almost absent afterwards: once a plant is built around a licensor’s configuration, the sieve specification, the operating envelope and the maintenance schedule are all set by that decision for twenty years.

The practical consequence for a buyer is asymmetric. Capital equipment is bought in a competitive tender with several credible bidders and a hard specification. Consumables are bought from a qualified supplier list that a plant is reluctant to disturb, because a failed sieve batch means a plant shut for a week. The first is priced by competition; the second by switching cost.

Why project lumpiness defines the field

A licensor’s year is decided by two questions: is India building, and is Brazil retrofitting? Both are policy derivatives rather than demand derivatives. When a blending mandate is raised, the order book fills within two quarters and lead times stretch. When the mandate is met, the book empties just as fast. That is why the engineering firms in this step behave cyclically even when ethanol volumes are stable, and why their pricing power appears and disappears with the policy calendar rather than with the price of ethanol.

What the fragmented buyer base cannot do

Individual plant operators have almost no leverage over long-lead equipment, because they buy one train at a time into a queue. They have real leverage over commodity service and over the merchants who move their product. The cooperative and trade-association layer is the only place where the buyers in this market aggregate, and where they do, it is to negotiate on freight and on the credit — never on equipment.

Chaîne de valeur

The processing step’s cost stack is dominated by one input: heat. Distillation and rectification are a boiling and condensing problem, and the energy required to drive a litre of ethanol from about 10 per cent concentration in the ferment to the 95.6 per cent azeotrope, and then past it to anhydrous, exceeds every other cash input the step consumes. Everything else in the value chain is organised around reducing that thermal load or recovering value from what the heat leaves behind.

Where the cost actually sits

In a North American dry mill, the processing block’s cash cost divides roughly into energy, maintenance and labour, consumables, and the fixed cost of the step’s own equipment. Energy is the largest single line by a wide margin, and within energy the split between distillation and drying is the one that plant engineers argue about, because in a mill that dries its full stillage the dryer can consume more thermal energy than the distillation train.

That single fact reorganises the whole value chain. If drying costs more than the water removal is worth, the rational move is not to dry.

The wet-product radius is a value-chain boundary

Wet distillers grains can be sold locally without drying, and the price they command is set by the cost of the alternative feed in that locality. The distance they can travel economically is set by their water content. So the boundary of the wet-product market is not a geographic fact and not a regulatory fact: it is the cost of drying, and it moves whenever gas prices move. A plant with a dairy within reach and expensive gas sells wet, skips the largest single thermal load in its processing step, and converts what looks like a by-product problem into a cost advantage its competitor three hundred miles away cannot copy.

This is why by-product recovery is a fifth of this market’s value rather than a footnote. It is not waste disposal. It is where a plant’s conversion cost is decided.

The two structural models, and why they price differently

The Brazilian cane model burns bagasse in its own boilers. Its processing step is energy-self-sufficient, which makes the cash cost of the step low, but it requires capital in the boiler house and in the cogeneration train, and it ties the plant to the cane season. The North American dry mill model buys natural gas and runs year-round on corn. Its processing step is capital-light relative to the cane mill and cash-heavy, and it is therefore directly exposed to a fuel price it does not control.

The two models produce the same product and do not have the same cost structure. That is why the same unit operation costs roughly USD 0.07 a litre in one place and USD 0.13 in another.

Who captures the margin

No one in this chain earns a processing margin as such. The plant earns the spread between what it pays for feedstock and what the market pays for ethanol and co-products, and processing cost is subtracted from that spread. This is why the step matters out of proportion to its share of revenue: feedstock is a pass-through set by a farm economy, and the ethanol price is set by mandates and by an export arbitrage, so conversion cost is the only line the operator actually controls.

The upstream interface is where the step also decides its own yield. Gallons per bushel is a fermentation number, but it is a distillation and stillage number too, because losses at the bottom of the column and in the centrifuge are losses of product. A plant that improves its yield by three per cent has improved its processing economics more than one that renegotiated its gas contract.

Downstream, the step hands off a specification. Anhydrous fuel ethanol enters a blender’s terminal; a monograph grade enters a validated pharmaceutical supply chain with documentation attached to every batch. The same litre, sold into the second channel, carries a processing step three to four times more expensive and a margin that is not a commodity margin at all.

Activité d'investissement

Two capital flows run through this step at the same time, and almost every published view of the industry conflates them. The first is capacity capital: greenfield plants, which buy the step as a full equipment train. The second is retrofit capital: existing plants adding carbon capture, corn-oil recovery, higher-protein distillers grains, or heat integration. The first is a policy-cycle event with multi-year lead times. The second is a payback calculation measured in months.

In 2025 the second flow is the large one, and the reason is arithmetic rather than sentiment.

An industry that stopped building

US ethanol capacity stood at 18.541 billion gallons on 1 January 2025, having added 64 million gallons across a full year. That is a rounding error in a fleet of nearly two hundred plants, and it is the clearest evidence available that greenfield construction in the world’s largest ethanol market has effectively stopped. The equipment market did not stop with it. It moved from building trains to modifying them.

The credit is the capex driver

The carbon-intensity credit changed what a rational plant invests in. Because the credit is earned on the production pathway rather than on the volume, the return on a dollar of retrofit capital is now set by how many kilograms of carbon dioxide equivalent the retrofit removes per million British thermal units of fuel, not by how many extra litres the plant can push. That has redirected the entire capital programme toward three things: capture, co-product upgrading, and energy integration.

Capture attaches at the fermentation vent, which is the cheapest carbon dioxide stream in the industrial economy because it is already nearly pure and already separated. That is why so much of the recent investment in this step is, in engineering terms, a compression and dehydration problem bolted onto a plant that was not designed for it.

Co-product upgrading is the better-returning half of the same programme. Extracting corn oil from the stillage stream, or shifting a mill from conventional distillers grains to a higher-protein product, changes both revenue and carbon intensity at once, at a capital cost that a single plant can carry. It is the only category of investment in this market where the payback does not depend on a tax provision.

Where new capacity is actually being built

Three places, and each has a different reason. India built out capacity against a blending target, and the equipment cycle that accompanied it was the largest single order book the licensors have seen in a decade. Brazil’s corn belt is building dry mills in states with no cane and cheap grain, which are pure greenfield processing builds with full scope. Second-generation and alcohol-to-jet remain small in volume but are the only place in the market where genuinely new process configurations are being commissioned rather than copied.

The second-generation numbers should be read carefully. Brazilian second-generation ethanol reached 120.2 million litres, up 104 per cent, which is a real achievement and 0.36 per cent of the country’s ethanol. A market that is doubling off a third of a per cent is a technology story, not yet a capacity story.

What the investment direction is not

There is no consolidation wave in the step itself. The equipment and licensing field is made up of private firms and conglomerate divisions, and the ownership of process technology has been stable for a decade. The mergers happen in the ethanol production layer — assets change hands when a plant’s margin fails — not in the layer that supplies it. That asymmetry means the suppliers in this market keep their customers when the customers change owners.

Direction: rising. But rising for retrofit and geography rather than for capacity, and therefore more exposed to the credit calendar than to the ethanol price. The single largest risk to the investment line in this report is not a fall in ethanol demand. It is a provision expiring.

Technologie et innovation

The core technology of this step is nineteenth-century chemistry executed with twenty-first-century control systems, and that is not a criticism. Distillation works. The innovation that matters commercially is not in the separation principle but in how little energy can be used to achieve it, and in what can be recovered from what is left over.

What earns money today

Heat integration. Multi-effect and multi-pressure distillation, mechanical vapour recompression, and heat recovery from the stillage stream are the highest-return engineering available to an existing plant. The principle is simple — reuse the vapour instead of condensing it against cooling water — and the execution is plant-specific. A well-integrated train can cut the thermal demand of distillation substantially against a conventional two-column design, and the payback is measured against a gas bill rather than against a subsidy. This is the least fashionable and most reliable technology in the market.

Not drying. The cheapest thermal improvement in a dry mill is to avoid evaporating water it does not need to evaporate. Selling wet distillers grains, or partially dried product, removes load from the dryer. It is not a technology purchase so much as a logistics decision, and it is often the single largest available energy saving on a site.

Co-product upgrading. Corn-oil extraction by centrifuge and the shift to higher-protein distillers grains are genuine process innovations with real margins. They change the revenue line and the carbon intensity of the plant at the same time, which is a rare combination, and they are licensed by a small number of firms whose process designs are protected in practice by know-how rather than by patents.

Capture attachment. The fermentation vent stream is nearly pure carbon dioxide. Capturing it is a compression, dehydration and pipeline problem rather than a separation problem, which is why it is the cheapest capture in heavy industry per tonne abated — and why the equipment that does it now sits inside this step’s scope.

What is proven and does not yet pay

Cellulosic ethanol has been demonstrated at commercial scale and remains constrained by pretreatment capital and enzyme cost per litre. The engineering works. The economics do not clear without support, which is why second-generation volume remains a fraction of one per cent of the market. That is a statement about cost, not about credibility.

Alcohol-to-jet is a different case and deserves separating from it. Converting ethanol to sustainable aviation fuel adds a new chemical step downstream — dehydration to ethylene, oligomerisation, hydrogenation — and its first commercial run in late 2025 was a genuine milestone. It is not an improvement to ethanol refining; it is a new customer for refined ethanol, and if it scales it raises the value of the specification work this report sizes.

What is mostly marketing

Digital twins, artificial intelligence process optimisation, and autonomous plant claims are largely advanced process control with new labels. The underlying techniques are sound and the gains are real but bounded, typically low single-digit percentage improvements in energy or yield that are worth having and do not constitute a step change. Buyers should treat a vendor claiming double-digit savings from software with the same scepticism they would apply to a vendor claiming double-digit savings from a pump.

The consumable annuity

Molecular sieve beads are the most interesting recurring revenue in the step. A pressure-swing adsorption unit’s beads degrade through hydrothermal ageing and are replaced on a multi-year cycle, on a schedule the plant does not fully control. That is an annuity with a switching cost attached, and it is the closest thing to a moat in this market. Membrane dehydration is the long-run competitor, and it has taken share slowly at small scale where the economics favour a lower-capital, lower-purity route.

Where the frontier actually is

Not in the separation. The frontier is in three places: the energy balance, where the remaining gains are real but incremental; the co-product slate, where chemistry can still create value; and the carbon accounting, where measurement and verification have become engineering disciplines in their own right. Technology maturity: mature core, with growth confined to the co-product and capture edges. Patent activity: low. Innovation intensity: moderate. A market that describes itself as high-technology is usually selling something other than chemistry.

Risque de marché

Six risks matter to this market, and they are not equally weighted. One of them is dominant, in the sense that if it materialises the other five become secondary.

Ranked

One: the credit calendar. The single largest determinant of investment in this step is a provision with an end date. The carbon-intensity credit applies to fuel produced after 31 December 2024 and sold by 31 December 2029, and requires a pathway at or below 50 kilograms of carbon dioxide equivalent per million British thermal units. Any retrofit that only pays because of that credit has a five-year window in which to do so. If the provision expires rather than being extended, the retrofit pipeline stops within about a year, and the processing market loses its fastest-moving segment. This is the dominant risk in the report.

Two: the crush margin. Demand for this step is derived, not primary. A plant with a negative crush margin idles, and an idled plant processes no litres. The 2025 US average net margin of USD 0.21 per gallon against a long-run average near USD 0.13 is a good year, and good years invite the acreage response that ends them. A return to mid-cycle margins would not change the step’s cost per litre; it would change the number of litres that get processed.

Three: energy price divergence. The step is a thermal process, so it imports whatever the regional gas market charges. A European wheat plant paying roughly USD 0.13 a litre for a step that a US dry mill runs at USD 0.105 and a Brazilian mill at USD 0.07 is not suffering a temporary disadvantage. That gap has already closed European capacity, and it will close more of it if the differential persists.

Four: substitution, at two levels. In the demand layer, electrification of the light-duty fleet reduces the gasoline pool that ethanol is blended into, which would eventually reduce the litres this step processes. In the step itself, membrane dehydration competes with molecular sieve adsorption, which matters to the consumable annuity rather than to the total market size. Substitution risk is high, but it is slow at the first level and structural at the second.

Five: scale and closure risk at the plant level. The plants most exposed to a margin squeeze are the ones with no by-product recovery, no capture, and no route to a premium grade. Those are precisely the plants whose closures would remove processing volume permanently. Closure is not evenly distributed across the fleet; it is concentrated in the smallest and least integrated units.

Six: permitting and water. Air permits for dryers and land-application limits for stillage are the two binding constraints on expanding an existing site. Neither is fatal, and both can add two years to a project that a customer expected to commission in one.

What the scenarios are worth

The conservative path puts this market at USD 16.00bn in 2033 against a baseline of USD 18.46bn, so the downside is worth USD 2.46bn of terminal value. The optimistic path puts it at USD 20.24bn, an upside of USD 1.78bn. The asymmetry is deliberate and it points the wrong way for the market: the downside is larger than the upside, because a credit expiry removes capital that cannot be spent later, while a credit extension only accelerates spending that was already planned.

Overall risk: moderate. Geopolitical exposure: moderate. Substitution risk: high. The distinction matters. This is not a politically fragile market in the way a tariff-exposed commodity is — ethanol crosses borders, but the processing step is domestic by construction, since equipment is installed where the plant is. It is a policy-dependent market, which is a different and more specific exposure.

The risk nobody prices

Very few operators model the step separately from the plant. That is rational at the site level, where the step is a cost line, and dangerous at the portfolio level, because it means capital is allocated without a view of which unit operation is earning its cost of capital. A plant that knows its cost per litre but not its cost per unit operation cannot tell whether to spend on heat integration, on sieves, or on a dryer it could switch off.

Paysage réglementaire

Regulation reaches this step through three separate doors, and they are administered by different agencies with different incentives. Product specification decides what a litre must be. Environmental permitting decides whether the equipment that makes it can run. Credit eligibility decides whether the investment that improves it pays. A supplier selling into this market has to satisfy all three, and the third has become the commercial one.

Door one: what the litre must be

Fuel ethanol is a specified product, not a commodity with a quality description. In the United States the controlling specification is the ASTM standard for denatured fuel ethanol, which fixes ethanol content, water, acidity, gum, sulfur, and the denaturant. In Europe the corresponding standard for ethanol as a petrol blending component performs the same function. Neither is optional, and both are enforced at the terminal rather than at the plant, which is why a specification failure is discovered by a customer rather than by an inspector.

The denaturant is where regulation and economics intersect. Denatured alcohol rules prescribe what may be added and in what proportion, and the denaturant is part of the volume the plant sells. Every basis point of denaturant is a basis point of ethanol the plant did not make, so the rules quietly set a small but real share of the plant’s yield.

For pharmaceutical grades the bar moves from specification to validation. A monograph grade must meet the pharmacopoeial standard and be made under good manufacturing practice, with batch documentation, change control and supplier qualification. Ethanol is classified as a low-toxicity residual solvent under the international guidance on residual solvents, which is helpful, but the process documentation burden is what makes this grade expensive, not the chemistry.

Door two: whether the equipment can run

The two environmental pressures on this step are air and water, and both are local.

Air. Distillers grain dryers are the largest point-source emission in a dry mill — particulate and volatile organic compounds both. In the United States they are the reason a capacity increase can trigger a major new-source review even when the increase is small, which is the practical reason plants debottleneck in increments. A dryer is also the piece of equipment most often idled for economic reasons, which means the emission profile of a plant changes with the gas price.

Water. Stillage and vinasse disposal is the binding constraint in cane country. Land application is regulated at state level — in São Paulo, by a technical standard that sets application rates against soil characteristics — and in India by consent conditions that have periodically shut distilleries rather than let them discharge. Where vinasse can be returned to the cane field at the right rate, it is a fertiliser substitute and a cost advantage. Where it cannot, it is the constraint that limits plant size.

Door three: whether the investment pays

This is the newest door and now the most consequential. The carbon-intensity credit in the United States rewards a pathway, not a product, and its threshold and its sale window both bite. Alongside it, the California low-carbon fuel standard and the European renewable energy framework create three separate accounting systems for the same litre, each with its own methodology for land-use change, co-product allocation and energy credit.

The European framework is in motion. The Commission’s proposal of 16 December 2025 would soften the 2035 target for renewable energy in transport from 100 per cent to 90 per cent, with the residual compensable through sustainable renewable fuel credits. For a plant deciding whether to spend on heat integration, that single change is worth more than any feedstock forecast in this report.

What is not applicable. The European carbon border adjustment mechanism does not cover ethanol. It covers cement, iron and steel, aluminium, fertilisers, electricity and hydrogen. Buyers routinely assume otherwise, and a supplier that plans for a carbon border cost on ethanol is planning for a cost that does not exist.

The compliance burden, measured properly

The cash cost of compliance in this step is modest — a fraction of processing value. The time cost is not. A monograph-grade qualification can take longer than the plant upgrade it depends on, and a capture project’s permitting can outlast the credit window it was designed to use. The binding constraint in this market is calendar, not cash, and it is why the suppliers who win are the ones who can sequence a project against a regulatory deadline.