Deconstructing Google's Megaton Climate Bet in Brazilian Rice Fields

SAN FRANCISCO & SAO PAULO – September 16, 2026 — As generative artificial intelligence drives an unprecedented expansion in global data center infrastructure, the tech industry's greenhouse gas emissions are surging. Between 2019 and 2024, Google saw its emissions spike by over 40%, laying bare the limitations of traditional, paper-based carbon offsets. In response, the search giant is engineering a physical, industrial-scale intervention that completely reimagines how corporate balance sheets interact with the global carbon cycle.

Today, Google and climate technology developer Terradot announced a megaton-scale environmental initiative spanning more than 200,000 hectares of commercial rice operations in Rio Grande do Sul, Brazil. The transaction, which stands as Google's largest carbon removal purchase to date, introduces a highly structured, dual-timeline framework: Google will procure 1 million tonnes of near-term methane abatement by 2030, paired with 1 million tonnes of permanent carbon dioxide removal (CDR) by 2040.

By stacking two distinct climate technologies—Alternate Wetting and Drying (AWD) and Enhanced Rock Weathering (ERW)—on the exact same acreage, the partnership is attempting to solve the fundamental friction of modern climate finance: the agonizing choice between cooling the planet today or permanently sequestering the carbon of yesterday.

The Dual-Timeline Architecture

Climate change operates on two distinct biological and chemical clocks. Methane, a superpollutant, is responsible for roughly one-third of human-caused warming to date. While it boasts a global warming potential over 80 times that of carbon dioxide over a twenty-year period, it clears the atmosphere in roughly a decade. Eliminating methane immediately depresses near-term peak temperatures. Conversely, carbon dioxide accumulates and persists for centuries, requiring durable, long-term removal strategies.

"For years, climate action has been framed as a choice, either move fast on the warming happening now or remove carbon permanently. Our project with Google eliminates that trade-off," said James Kanoff, CEO of Terradot. "We are pairing two proven technologies and deploying them together at a massive scale. With this new model, every tonne of removal meets the same rigorous standard, and delivers climate impact faster than ever."

In practice, this means Terradot leverages the fast-acting methane reductions to bridge the gap while the slower, permanent mineral carbon removal phases in. For corporate buyers like Google, this bridged impact profile is highly attractive.

"We built this agreement to be a template, not a one-off,” said Randy Spock, Head of Carbon Removal at Google. “It's the largest carbon removal purchase Google has ever made, and by incorporating action on superpollutants, we realize that impact by 2030. We think it provides a replicable path to making a real dent in climate change in both the near and long term, and we’ve structured it so any buyer, supplier, or government can follow this same approach."

The Biogeochemistry of Rice and Rock

The physical execution of this megaton bet relies on the unique agronomic landscape of Southern Brazil. Rio Grande do Sul cultivates approximately 1 million hectares of flooded rice, accounting for over 70% of Brazil's national output. Continuously flooded rice paddies create anoxic, oxygen-deprived conditions where methanogenic archaea thrive, digesting organic matter and releasing vast quantities of methane.

Terradot's first intervention, AWD, involves periodically draining these flooded paddies. Dropping the water level introduces oxygen into the topsoil, effectively starving the methanogenic archaea and cutting methane fluxes by 30% to 70%. For local farmers, many of whom are still rebuilding infrastructure following the catastrophic regional floods of May 2024, AWD offers immediate operational relief. Rice irrigation relies on heavy diesel and electric water pumps that run continuously for months; reducing water consumption by up to 30% translates directly into vital fuel and electricity savings.

Simultaneously, Terradot will deploy massive volumes of crushed volcanic basalt across these drained fields. When atmospheric CO2 dissolves in rainwater, it forms a dilute carbonic acid. As this acid reacts with the crushed silicate minerals, it dissolves them, capturing the carbon as dissolved inorganic bicarbonate. This bicarbonate eventually washes into drainage systems and the open ocean, safely locking away the carbon for tens of thousands of years.

Beyond carbon sequestration, the crushed rock acts as a potent agromineral. It releases secondary nutrients like calcium and magnesium, allowing growers to substitute expensive, imported synthetic fertilizers. Crucially, the basalt also releases plant-available silica, which thickens the epidermal cells of the rice plants, enhancing their natural resistance to fungal blast and preventing stem lodging during severe weather.

Mitigating the Nitrous Oxide Trade-Off

While the integrated model is elegant on paper, altering soil hydrology at a landscape scale introduces severe forensic accounting risks. The most glaring red flag in any AWD project is the nitrous oxide trade-off. Introducing oxygen into previously waterlogged soils triggers complex nitrification cascades that can provoke pulses of nitrous oxide—a greenhouse gas with roughly 273 times the warming potential of CO2. If farmers drain fields carelessly, these invisible chemical flares can easily erase the project's net climate benefits.

To proactively address this, Google is providing a targeted R&D grant to both Terradot and the Stanford Doerr School’s Soil & Environmental Biogeochemistry Lab. The mandate is to establish high-frequency soil flux monitoring across the region, utilizing automated static flux chambers to quantify the exact balance between methane suppression and nitrous oxide emissions.

“Google and Terradot are weaving together technologies that allow for rigorous validation and scaling to bend the warming curve in our lifetimes,” said Scott Fendorf, PhD, Chief Scientific Advisor and Co-founder of Terradot, whose Stanford lab will spearhead the open-source data collection.

This ground-truth data is fused with synthetic aperture radar imagery from Sentinel-1 satellites, which penetrates cloud cover to verify that the 200,000 hectares are actually being drained according to schedule.

For the rock weathering component, Terradot relies on intellectual property secured during its February 2026 acquisition of Eion. By measuring immobile trace elements like titanium and zirconium that are naturally present in the basalt, Terradot's geochemical tracing mathematically isolates the carbon removal signal from the noisy background of soil variability, meeting the strict verification protocols of registries like Isometric and Puro.earth.

Scaling a Replicable Blueprint

Backed by $58 million in funding from heavyweights like Microsoft's Climate Innovation Fund, Gigascale Capital, and John Doerr, Terradot is aggressively positioning itself to dominate the global ERW market. The company is targeting 1 million tonnes of annual CO2 removal by 2035.

By stacking two revenue streams—methane avoidance and mineral carbon removal—on the exact same hectare, Terradot radically improves the unit economics of farmer engagement, heavy logistics, and field verification. Securing a megaton forward commitment from a tier-one buyer like Google allows the company to build dedicated supply-chain infrastructure, driving the historically high cost of ERW (often $250 to $400 per tonne) down toward the industry's holy grail of sub-$100 per tonne.

However, exporting this blueprint from the sprawling, mechanized plains of Brazil to the world's largest rice basins in Asia presents a formidable logistical hurdle. Southern Brazil features massive commercial operations, allowing for rapid deployment of bulk-spreading machinery over thousands of contiguous hectares. In contrast, rice cultivation in the Mekong Delta or the Indo-Gangetic Plain is dominated by smallholder farms averaging less than two hectares. Aggregating, contracting, and verifying interventions across millions of fragmented Asian plots will require an entirely different operational architecture.