Drones, AI and Carbon Credits: How Re.green Is Restoring Brazil’s Rainforests

Drones, AI and Carbon Credits: How Re.green Is Restoring Brazil’s Rainforests

Brazil harbors some of the world’s largest tropical forests, but they face intense pressure from deforestation and land use change. The Atlantic Forest along Brazil’s eastern coast once covered roughly 350 million acres; today only about 12% remains, according to the World Wildlife Fund. The Amazon — largely in Brazil and one of the planet’s most biodiverse regions — has lost almost 20% of its forest cover in the past 50 years.

Technology-driven restoration

Re.green, a Brazilian ecological restoration startup, combines artificial intelligence, satellite imagery and drones to identify degraded land across the Amazon and Atlantic Forests and design tailored restoration plans. The company won the Earthshot Prize in the "protect and restore nature" category for its "transformative" use of technology to revive tropical forests.

How it works

The process begins with AI algorithms that analyze scientific datasets and high-resolution satellite imagery to pinpoint candidate parcels — usually degraded lands with low agricultural productivity. Re.green secures these areas either by purchasing them or arranging long-term leases, often for 50 years or more, with farmers and ranchers.

The same AI system selects from a library of more than a dozen restoration models and prescribes site-specific approaches: intensive, mechanized planting where infrastructure permits, or passive natural regeneration where conditions are favourable. In remote locations with limited access, drones perform much of the fieldwork, including aerial seed-planting.

Native species and nursery capacity

All projects prioritize diverse native species. Many seedlings are produced at Bioflora, a tree nursery re.green acquired in 2021 that can grow about 2 million seedlings per year. The company says the ultimate goal is to restore fully functioning tropical forests that approximate the original primary forest structure.

Financing restoration

Re.green’s algorithms also build financial models for each tract using land acquisition costs and estimates of carbon sequestration over time. Those models inform monetization strategies: sustainable forestry and verified carbon removal credits.

In some mechanisable, well-connected areas the company plants a narrower mix of five to ten native species intended for high-value timber harvest in about 20 years, generating returns that help fund restoration at scale. In other sites, restoring forest generates carbon removal credits that companies can buy to offset emissions. Re.green says its credits demonstrate clear additionality and are verified by independent auditors before issuance.

Partnerships and scale

This year re.green agreed to sell nearly 3.5 million tonnes of carbon removal credits to Microsoft tied to restoring 33,000 hectares, supplementing a 2024 agreement for 3 million tonnes. It has also partnered with Nestlé to restore 2,000 hectares of Atlantic Forest in southern Bahia.

To date, re.green has planted more than 6 million seedlings across 30,000 hectares in four Brazilian states. The company aims to raise $60 million by the end of 2025 to scale its AI and satellite capabilities, and to plant 65 million seedlings by 2032. The Earthshot Prize is expected to increase visibility, attract partnerships and accelerate research and development.

"A tropical forest provides a lot of value for us in society," said Thiago Picolo, CEO of re.green. "It’s our job to find assets that can be monetized so we can scale faster and restore more land."

Impact and outlook: Re.green’s approach aims to make large-scale ecological restoration financially viable by blending cutting-edge technology, practical restoration techniques and market mechanisms. If it succeeds at scale, the model could be replicated in other tropical regions facing deforestation.