How AI could help Brazil grow sustainable aviation fuel, quite literally, on trees
Unwanted Fruit
One of the most promising answers to aviation’s carbon problem may begin with a fruit that many Brazilian farmers have spent years trying to get rid of.
Macaúba is a native palm found across large parts of Brazil. It is thorny, its fruit falls onto pastures, and on many farms it was something to remove, work around or simply ignore. Now it is attracting billions of dollars.
Acelen Renováveis, backed by Abu Dhabi-based Mubadala Capital, is trying to turn macaúba into the principal feedstock for a new generation of sustainable aviation fuel, or SAF.
The company plans to build a chain stretching from degraded pasture in Minas Gerais and Bahia to a biorefinery beside the Mataripe refinery. The scale alone makes it an ambitious bet, with much depending on whether a largely wild native plant can become a reliable commercial crop fast enough to supply an industry that barely exists today.
Aviation is one of the hardest sectors to decarbonize for a simple reason: aircraft need enormous amounts of energy without enormous amounts of weight. Batteries may work for cars and perhaps short regional flights, but they remain impractical for most long-distance aviation. Liquid fuel is difficult to replace at 35,000 feet.
That leaves SAF as one of the few near-term alternatives capable of working in existing aircraft and fuel infrastructure. Governments are now creating demand through regulation. Brazil’s recently regulated ProBioQAV program requires domestic airlines to begin reducing emissions through SAF in 2027, starting at 1% and rising progressively to 10% from 2037.
Mandates, however, can create a market without creating an economical product. SAF remains at least three times more expensive than conventional jet fuel. The International Energy Agency expects global consumption to increase from around one billion litres in 2024 to nine billion litres by 2030, yet even then it would meet only about 2% of aviation fuel demand.
Much of the attention naturally falls on building refineries, although the availability and cost of feedstock may prove every bit as important. Most SAF produced today relies on used cooking oil, animal fats or vegetable oils. These sources can get the market moving, but supplies are limited and many already have competing uses.
Reaching beyond mandates and premium buyers will require crops that produce more oil per hectare, reducing both land requirements and transport costs.
Smart Seed
Acelen estimates that the palm can produce seven to ten times more oil per hectare than soybeans. It is native to Brazil, adapted to tropical and semi-arid conditions, and the project is designed around degraded pasture rather than replacing food crops or native vegetation. Researchers are also examining integration with livestock and higher-value uses for other parts of the fruit.
On paper, it is close to the perfect biofuel crop. The slight inconvenience is that it is not yet really a crop.
Brazil’s success with soybeans, sugarcane and eucalyptus reflects decades of genetic improvement, agronomic research, mechanization, infrastructure and farmer experience. Macaúba is only beginning that process.
Its seeds germinate poorly in nature, and individual palms vary considerably in productivity and oil content. Harvesting has also yet to be proven economically across large commercial areas. A plant that survives in the wild is not automatically one that performs predictably in a farm budget.
To accelerate the work, Acelen created the Agripark research centre in Montes Claros, Minas Gerais, together with Embrapa, universities and other research institutions.
AI-assisted research developed with the State University of Montes Claros has reportedly helped lift germination above 80%, compared with only 3% to 5% in nature.
Automated seed-processing systems can handle up to 1.7 million seeds a month, while the centre is designed to produce 10.5 million seedlings a year. Researchers are also selecting higher-performing genetic material, testing planting arrangements and developing cloning techniques.
AI’s role here is fairly unglamorous, focused on identifying promising plants, improving germination protocols and learning from field data across different soils and climates.
Yet this could be its most useful contribution to SAF, helping reduce the biological variation that makes a wild plant difficult to use as an industrial input. Even with faster analysis, however, the palms still need years to mature, and results from the nursery must be repeated through drought, pests and variable farm management.
Long Haul
Acelen plans to invest around US$3 billion in the integrated project, including approximately US$1.5 billion in a biorefinery in São Francisco do Conde, Bahia. Scheduled to begin operating in 2029, the facility will have capacity to produce one billion litres of SAF and renewable diesel annually.
The industrial timetable is running ahead of the agricultural one. Macaúba will not be available at sufficient scale when the refinery opens, so Acelen has secured used cooking oil from Trafigura and certified soybean oil from Bunge as transitional feedstocks.
These contracts allow the refinery and orchards to develop on different timetables, while also showing how much easier it is to schedule an industrial project than an agricultural one. Construction can follow milestones; a new crop must be tested across soils, seasons and farming systems.
Acelen ultimately expects to develop 144,000 hectares of macaúba across several production hubs in Bahia and northern Minas Gerais. So far, it has acquired roughly 3,000 hectares. Around 20% of production is expected to come through partnerships with small and medium-sized growers, using a model similar to the forestry industry.
Farmers could gain a new source of income from degraded or underused pasture, potentially alongside cattle. Acelen has discussed supplying seedlings, technical assistance and guaranteed purchasing to participating families, but adoption will depend on the terms behind that support.
Who finances establishment before the palms produce? Who carries the risk if yields fall short? How will prices be calculated? Can harvesting and local crushing be done economically? Will carbon income reach the landowner? The answers will shape the business model and determine how much of the early risk sits with growers.
The project has advantages that most new crops lack: a deep-pocketed sponsor, a biorefinery under development, transitional feedstock and Brazil’s new SAF mandate. Even so, the gap between a few thousand acquired hectares and 144,000 productive hectares is enormous.
Few countries are better placed to try. Brazil already knows how to build large biofuel industries, adapt crops to tropical conditions and process agricultural commodities at scale. A successful macaúba chain would create value in genetics, production systems, machinery, carbon measurement and refining, with far more of it remaining in Brazil.
Acelen has already attracted the capital, research partners and buyers needed to get started. We will only know whether macaúba works as a commercial crop after several seasons of consistent yields across much larger areas. Farmers will also need to make money after covering planting, early maintenance, harvesting and transport costs.
In the end, the refinery can make the fuel, but the farmer will decide if it flies.
Thanks for reading.
KFG
Kieran Finbar Gartlan is an Irish native with more than 30 years’ experience living and working in Brazil. He is Managing Partner at The Yield Lab Latam, a leading venture capital firm investing in AgriFood and Climate Tech startups in Latin America.