In 2024, a police raid on Brazil’s largest producer of carbon credits marked the lowest point yet in the long reputation crisis of REDD projects. Major companies that had bought credits from those forest conservation efforts suddenly saw themselves linked to a criminal organization.

REDD, which stands for reducing emissions from deforestation and forest degradation, is a methodology based on a simple concept: protecting strategic swaths of rainforest helps avoid greenhouse gas emissions. The parties doing the protecting, therefore, can sell credits — amounting to the emissions they’re estimated to have avoided — to companies trying to offset their carbon footprint.

However, the organization targeted by Brazilian authorities, led by Ricardo Stoppe, was also making money from illegal logging and land grabbing in southern Amazonas state. The scandal was revealed first by Mongabay. Yet the case was far a one-off: according to a study reported by U.K. newspaper The Guardian, 90% of rainforest carbon offsets by the world’s leading certifier were worthless.

In other investigations, Brazilian journalists and NGOs have unmasked companies illegally occupying public lands or profiting from traditional communities’ territories without their approval. In a December 2024 report, Mongabay showed that partnerships between REDD project developers and loggers with a history of environmental crimes are common in the Brazilian Amazon.

“We have seen a series of liabilities, whether due to the quality of the projects, the low level of integrity, failures in due diligence, or methodological problems,” Gustavo Pinheiro, from the climate solutions consultancy Grupo Trie, told Mongabay. “It’s a long list of liabilities that have led to this loss of confidence [in REDD].”

As carbon credit prices have gone into free fall, a different type of operation is gaining ground as a finance solution for conservation: Restorating large swaths of the Amazon and selling the carbon credits they absorb. And Brazil, with 109 million hectares (269 million acres) of degraded pastures, surged as the restoration golden goose.

Different from REDD, where credits are estimated from deforestation that didn’t happen, these initiatives consist of replanting the forest in degraded areas. Plants absorb carbon as they grow, and carbon credits are calculated based on this absorption.

“You have an area of pasture, then you start planting the forest and it’s growing,” Beto Mesquita, a forest engineer and director of forestry and public policy at BVRio, an NGO working in market solutions to socioenvironmental problems, told Mongabay. “You can very easily show how much that forest has grown in five years, 10 years. It’s much clearer [than REDD].”

The possibility of combining large profits with a positive agenda has attracted several players. Biomas, for example, is one of the leading Brazilian restoration companies, established by the banks Itaú, Santander and Rabobank, mining giant Vale, pulp and paper company Suzano, and meatpacker Marfrig. The company re.green is backed by some of Brazil’s most prominent investors, including former central bank president Armínio Fraga and the Moreira Salles family, owners of Itaú.

Another market leader is Mombak, founded in 2021 by two young entrepreneurs from the digital bank Nubank and the Brazilian transport app 99. It has four restoration areas in Pará state in different implementation stages, covering a combined 20,000 hectares (nearly 50,000 acres). According to the company, it is the world’s largest restoration initiative.

With an aggressive marketing strategy and a strong presence in international and national media, the startup has already raised 1.25 billion reais ($218 million) from top global investors like CPP (Canada Pension Plan Investment Partners), Bain Capital Partnership Strategies, and the Rockefeller Foundation and from financial institutions like the World Bank and BNDES, Brazil’s national development bank.

Despite not yet issuing a single credit, Mombak has already signed contracts with Google and Microsoft. “Financial resources are available and growing, but for the size of the market, you will need much more,” Mombak’s science director, Renato Crouzeilles, told Mongabay.

Since Brazil’s regulated carbon market isn’t working yet, restoration credits are sold in the so-called voluntary market, where companies buy credits not because they’re obliged to by law, but because they want to show green credentials to clients and investors. Since there’s no legal framework in place, the rules are established by private certifiers like Verra.

Mombak’s projects are still being analyzed by Verra under a methodology specifically designed for so-called afforestation, reforestation, and revegetation (ARR) projects. However, the company says there’s still little information on how to quantify the carbon absorption in restoration projects.

“The market is completely new and in its infancy. We’re doing something from scratch,” Crouzeilles said. “I have to say how much carbon my farm will sequester over the next 50 years. This information doesn’t exist in the literature. We do the best modeling to get farm-by-farm estimates in the Amazon, but my project would be much less risky if this information existed.”

Mesquita from BVRio expressed concerns about the financial sustainability of projects that rely entirely on carbon credits. Brazil has just approved its regulated market, but its full implementation is only expected by 2030. At last year’s COP29 climate summit, in Baku, Azerbaijan, international negotiators made progress on creating a centralized carbon trading mechanism, but were unable to carry it through to completion. “I think it’s a good bet, but it’s a gamble,” Mesquita said. “Nobody knows exactly what the demand for carbon credits will be in 10 years and how much the market will be paying for it.”

Despite the uncertainty, Brazil’s public sector has also dived right into the restoration carbon market. The federal government has committed to restoring 12 million hectares (30 million acres) of native vegetation by 2030, an area roughly the size of North Korea. The Brazilian National Plan for the Recovery of Native Vegetation (Planaveg) includes awarding concession over public lands, such as conservation areas, to private companies that commit to restoring the vegetation. In exchange, they can sell the carbon credits generated in those areas.

The rollout should begin in Bom Futuro National Forest, in Rondônia state, which had 14,000 hectares (34,600 acres) deforested. By 2026, the Lula administration plans to grant a total of 350,000 hectares (865,000 acres) of public forests to private companies, according to Brazilian newspaper Folha de S.Paulo. In March, the federal government launched a bid to select restoration projects focused on land reform settlements. A few weeks later, BNDES announced a partnership with Brazil’s state oil company Petrobras to support restoration projects in the Amazon.

Pará, the Brazilian Amazon’s top state for deforestation and the host of this year’s COP30 climate summit, is also getting involved. The state government opened a public bid for companies interested in restoring 10,000 hectares (nearly 25,000 acres) inside Triúnfo do Xingu Protected Area (APA do Xingu), in Altamira municipality. Systemica, a company partially owned by Brazilian bank BTG Pactual, won. “The Triunfo do Xingu Recovery Unit is a pilot project,” said Governor Helder Barbalho at the launch of the bidding, promising the same model will be replicated throughout Pará.

Restoration in practice

Scientists have already proved the rainforests’ remarkable ability to begin regenerating itself as soon as humans withdraw their cattle and chainsaws. But the task is not always so simple, especially when it comes to complex ecosystems with a delicate balance between thousands of animals, insects, plants and microorganisms. “There is nothing simple in forest restoration,” Mesquita said.

A successful restoration project means that the vegetation not only looks like a rainforest but also provides all its ecological services, like carbon storage, water resource protection, and biodiversity corridors. For this to happen, it’s not enough to withdraw from an area. “Forest restoration without having to plant a seedling or a seed is possible, desirable and necessary. But it requires a minimum of monitoring,” Mesquita said.

Restoration managers must ensure no fire or cattle enter the area, and no dominant species preventing the growth of native plants, he said. “It may seem that the vegetation has recovered, but if you don’t have a minimum monitoring process you can’t say that the area is being restored,” he added.

More intervention is needed depending on the condition of the land and the surrounding landscape. Soils that have been eroded and compressed over years of pastoral and agricultural activity, for example, are hardly suitable for natural restoration. It’s the same if no healthy forest is nearby to provide the airborne seeds to the restoration area.

Given this, organizations working in the Brazilian Amazon use different strategies. One of them is muvuca, a technique developed by Indigenous people and researchers in the Xingu River Basin in Mato Grosso state, and now practiced in other parts of the country. Instead of growing seedlings and then planting them, seeds from different native species are mixed up and planted directly in the ground. Behind this apparent carelessness, there’s a careful choice of seeds so that some specific plants grow earlier, preparing the soil and providing protection for those that come later.

Besides being cheaper, the technique generates income for traditional communities who collect forest seeds, like members of the Xingu Seed Network, supported by Instituto Socioambiental (ISA), a non-profit working to protect traditional communities’ rights. “It’s a model that is starting to take off in Brazil, with absurdly positive results,” Pinheiro, from Grupo Trie, told Mongabay.

Another option is nucleation, where small species groups are planted in different nuclei that will later connect. Finally, there’s the so-called total planting, where many seedlings are grown in the same area. This is Mombak’s primary technique, with the company growing 1,700 seedlings per hectare (about 700 per acre). “It’s a much more expensive activity, with much more demand for people, inputs such as limestone and seedlings,” said Mombak’s Crouzeilles, without detailing the cost. “But I sequester carbon faster because I know which species I’m planting.”

Rioterra, an NGO with 16 years of experience restoring small properties in Rondônia, uses a combination of these methods in its projects. However, the first part of the work is to help the soil recover from years of use as pasture. “We usually restore areas of brachiaria [a type of grass planted for cattle], which is very aggressive. Unfortunately, the only thing that stops it is a herbicide,” Alexis Bastos, project manager at Rioterra, told Mongabay. “And after planting, you have to avoid competition between the grass, which grows much faster, and the native plants. If I have too much grass, it covers the plants and prevents them from doing photosynthesis.”

Even when native plants take root and start growing, the work is far from over. For many years, it will be necessary to keep fighting pasture and invasive species, controlling ants, maintaining fences to prevent the entry of cattle, constructing firebreaks, and eventually replanting some species that die. According to Bastos, Rioterra spends from $5,000 to $8,000 per hectare ($2,000-$3,200 per acre) over the decades to keep them safe.

“Many financiers want to cut costs and say they’re planting forests, but they’re actually planting seedlings,” he said. “But how am I going to look after this seedling for 30 years if you pay me a pittance? If you want to plant a forest, you have to be aware of the maintenance costs involved.”

In times of climate change, the risks are even higher. In 2024, for example, the Amazon registered a severe drought and its worst fire season in 14 years, and the projections for 2025 look gloomy. In the face of such uncertainty, it’s hard to ensure the permanence of the restoration projects for the next 100 years, as foreseen in Mombak contracts, for example.

“The risks of permanence are exacerbated by the climate crisis, the process of savannization of the Amazon, the change in the hydrological cycle and the increase in fires,” Pinheiro said. “Just as a conserved forest can catch fire, a restored forest can catch fire, and then you can’t guarantee the permanence of that biomass stored in the trees, which literally goes up in smoke.”

New model, old problems

Mombak’s main project is Turmalina, a former cattle ranch located in Pará’s Mãe do Rio municipality, where 2,300 hectares (5,600 acres) are being reforested with native species. Besides acquiring rural properties, Mombak also partners with farmers interested in diversifying their income. Either way, the company is susceptible to the same risks as other REDD projects when it comes to land tenure in the Amazon.

“Buying land or entering into a partnership with a rural landowner in the Amazon is like looking for a needle in a haystack because the legal and land insecurity are enormous,” said BVRio’s Mesquita.

“This is one of the biggest difficulties of our project because land titling in the Amazon is extremely problematic,” said Crouzeilles, adding that Mombak conducts a detailed due diligence to ensure the land has proper documentation.

Companies like Mombak work in large areas using high technology and expensive machinery to reduce personnel. The goal is to improve efficiency and accelerate both restoration and the generation of carbon credits. “Today, our exclusive source of income is carbon credits,” Crouzeilles said.

Mombak says it employs around 80 people at its four projects and that it tries to engage with the surrounding communities by allowing them to enter the properties to collect fruits like açaí. Even so, this model attracts criticism from other groups working in the Brazilian Amazon, who argue that restoration has to be done in partnership with local people. At the same time, these critics say acquisition of large areas intensifies land speculation.

“When these companies start buying up a lot of land, this space will be used for a purpose that isn’t necessarily for collective use,” Rioterra’s Bastos said. “This creates future pressure on these areas, which can be dangerous because they will be conflict zones.”

Rioterra started working on restoration in Rondônia in 2009, recovering only 30 hectares (74 acres) in a permanent preservation area. “And we hardly got any people to restore these 30 hectares. Nobody was interested in that at that time,” Bastos said. Since then, the organization has restored 7,000 hectares (17,300 acres) to date, spread over hundreds of small rural properties, besides projects in Indigenous territories and conservation units in partnership with public bodies.

Bastos said the communities themselves consistently plant the seedlings, and carbon credits are used in specific cases as extra income. “We work with restoration to generate income and employment, produce food, and socially include women and young people,” he said. “Today, many companies work with restoration because they want carbon credits, and restoration is the end in itself. We don’t do that.”

Belterra, a company focused on agroforestry projects and food production, follows a similar approach. “Carbon doesn’t sustain our model, it can help unlock it. What sustains the model economically are the products of agroforestry,” Valmir Ortega, Belterra’s founder, told Mongabay. In 2023, tech giant Amazon invested 90 million reais ($15.7 million) in Belterra, in a project that involved cacao crops and carbon credits.

The company runs initiatives in five Brazilian states in partnership with more than 370 rural producers. It manages all the steps of the business, from growing to selling forest products, primarily cacao and açaí, but also cassava and banana, which take less time to start harvesting. “That’s the beauty of the agroforestry system,” Ortega said. “You have a succession of products over the years that the producer can manage and choose from.”

According to the company, its agroforestry systems generate gains per hectare 10 times higher than soybean cultivation and up to 40 times higher than with cattle, on top of their positive environmental impact. “And it brings something that ecological restoration doesn’t, which is a very significant social impact in terms of generating income, jobs and food,” Ortega said.

Banner image: Restoration projects have been developed in large Amazon properties with high investments and intensive use of machinery. Image courtesy of Mombak/Raimundo Paccó. 

Top brands buy Amazon carbon credits from suspected timber laundering scam

FEEDBACK:Use this form to send a message to the author of this post. If you want to post a public comment, you can do that at the bottom of the page.