Google and McKinsey Back 320,000-Tonne Brazil Silvopasture Deal
Google and McKinsey have contracted 320,000 tonnes of carbon removal over ten years from a Brazilian silvopasture program that will keep cattle on the land while planting rows of native and commercial trees. Announced on 19 October 2026 through the Symbiosis Coalition, the agreements are a notable attempt to make long-duration carbon finance part of working-ranch economics rather than paying landowners simply to stop producing.
The distinction between a contract and a delivered removal matters. The 320,000 tonnes are future purchases, not credits already issued, and the companies disclosed neither the price nor the annual delivery schedule. Google separately said its agreement covers 270,000 tonnes. That leaves 50,000 tonnes for McKinsey by subtraction from the coalition total, although McKinsey did not publish a separate volume in the material reviewed by Agrilinkage.
At an even pace, the combined commitment would average 32,000 tonnes of carbon dioxide removal a year. Google represents 84.4% of the contracted volume and McKinsey 15.6%. Those calculations describe the purchasing commitments; they do not forecast how quickly trees will grow or when verified credits will be delivered.
An 8,700-hectare first phase inside a much larger plan
Symbiosis says the offtakes are intended to restore 8,700 hectares of degraded and semi-degraded pasture in the southeastern Amazon and central Cerrado. That area is 17.4% of Working Trees’ broader 50,000-hectare project. The developer’s website reports 2,167 hectares planted to date, involving about 450,000 trees and 18 partner farmers, while the coalition describes a pilot exceeding 2,000 hectares.
The contracted phase is therefore roughly four times the area Working Trees says it has already planted. Scaling from a pilot to thousands of additional hectares changes the operating challenge: planting material, local crews, landowner contracts, fire management, tree survival and long-term monitoring must all expand without weakening standards.
It would be misleading to divide 320,000 tonnes by 8,700 hectares and call the result a project removal rate. The crediting period, actual enrolled parcels, tree-harvest cycles and issuance schedule have not been fully disclosed, while the 50,000-hectare program is broader than the initial offtake area. A simple ratio can be a useful sense check, but not a substitute for the approved carbon methodology and verified monitoring reports.
Original contextual editorial illustration created for Agrilinkage. It does not depict the named Working Trees project or a specific ranch.
How the ranch model is meant to work
Silvopasture deliberately combines trees, forage and livestock. Working Trees plans rows of commercial timber and native species with grazing lanes between them, allowing cattle production to continue. The coalition identifies teak and African mahogany among the commercial species and says native species will account for at least 25% of planting, rising to 50% by 2030.
The commercial trees are intended to produce future timber revenue, giving landowners a reason to maintain the system after carbon payments decline. Native species are expected to add denser, longer-lived carbon stocks, habitat, forage or non-timber products. Working Trees says it covers most implementation costs and supplies technical support, which addresses one of the principal barriers to adoption: ranchers may benefit over time but face high upfront planting and management costs.
The claimed production benefits—more shade, better water retention, healthier soils, reduced heat stress and greater drought resilience—are plausible features of well-designed silvopasture, but they remain site-specific outcomes. They depend on rainfall, soil, tree density, forage management, cattle stocking, species choice and maintenance. The announcement does not publish controlled productivity data from the Brazilian pilot, so these benefits should be treated as objectives rather than verified project results.
The real asset is the long-term purchase contract
For the carbon market, the important innovation is financial. A ten-year buyer commitment can give a developer predictable demand before most removals exist, making it easier to finance planting and persuade ranchers to sign long-duration agreements. Symbiosis explicitly compares this structure with power-purchase agreements that helped renewable-energy projects secure capital.
That does not mean the contract value is known. Without a price per tonne, Agrilinkage cannot calculate how much capital the offtakes unlock, how much reaches landowners, or whether carbon revenue is sufficient to cover establishment and monitoring costs. Any estimate of dollars per hectare would be speculation.
The structure also transfers performance obligations into the future. Trees must be planted, survive, grow and remain accounted for. Carbon must be measured against a credible baseline, independently verified and ultimately issued under an identified standard. Buyers may have delivery protections, replacement obligations or buffer arrangements, but those contract terms were not disclosed.
Four integrity tests will decide whether the credits are valuable
First is additionality: the project must demonstrate that the trees and management changes would not have happened without carbon finance. Existing ranch investment, timber economics and public incentives all affect that counterfactual.
Second is permanence. Timber harvests do not automatically invalidate a project, but harvested carbon, wood-product storage, replanting and reversal risk must be treated consistently. Symbiosis says harvest cycles are included in ongoing carbon quantification, yet the announcement does not provide the full methodology, buffer pool or reversal provisions.
Third is leakage. Keeping cattle on enrolled land may reduce the risk that ranching is pushed into another area, but it does not prove that stocking pressure, herd expansion or related land-use change will not shift elsewhere. Landscape-level monitoring is still necessary.
Fourth is measurement. Working Trees says its system combines geotagged tree records, phone photographs, timestamps, satellite imagery, machine learning and field data. That can make monitoring cheaper and more auditable, but digital records are evidence inputs, not verification by themselves. The decisive documents will be the registry methodology, baseline, validation report, monitoring reports and third-party verification statements.
Why this matters beyond one carbon purchase
Brazil has vast areas of cattle pasture, including substantial degraded land, so a model that raises land productivity without displacing livestock could attract both climate capital and agricultural investment. The coalition says the design may eventually be suitable for 5 million to 10 million hectares, while Working Trees states a one-million-hectare deployment ambition by 2035. Those are expansion ambitions, not contracted or planted areas.
The current scale is materially smaller: 8,700 hectares tied to the offtakes and just over 2,000 hectares reported as planted. The gap between those figures is precisely why the transaction is worth watching. It tests whether corporate demand can finance the difficult middle stage between a pilot and a repeatable agricultural system.
The deal also arrives as technology companies seek more removals while data-centre and artificial-intelligence infrastructure raise electricity demand. Carbon removals can address residual emissions, but they are not a substitute for reducing emissions inside buyers’ own operations and supply chains. The quality of the purchase depends on both sides of that equation.
What Agrilinkage will watch next
The first hard indicators are hectares formally enrolled, hectares planted, tree survival after the first dry seasons, the share of native species and the number of participating ranchers. The second group is agricultural: forage performance, stocking rates, animal heat stress, drought losses and net ranch income compared with suitable local baselines.
The third group is carbon-market delivery: the named standard and methodology, independent validation, first monitoring period, issued credits, tonnes actually delivered to each buyer, reversal provisions and any public price information. Publication of land-tenure safeguards and benefit-sharing terms would also show whether scale is being built on durable local agreements.
Until those data appear, the transaction should be understood as a significant financing commitment and a test of a promising production model—not proof that 320,000 tonnes have already been removed.
Primary sources and methodology
Agrilinkage reviewed the 19 October 2026 announcements from Google and the Symbiosis Coalition, the Working Trees project information available on 20 October, and independent reporting for cross-checking. Calculations in this article use the disclosed 320,000-tonne collective commitment, Google’s 270,000-tonne volume, the ten-year term, the 8,700-hectare offtake area and the 50,000-hectare broader project.
Comments