How an AI-powered waste system generates biochar carbon credits
A system that uses artificial intelligence to sort municipal waste could soon be producing a new revenue stream for its owners: hundreds of thousands of carbon credits.
Colorado-based recycling company AMP removes organics and recyclables from trash using AI-powered sorting technology. Rather than being sent to landfill, where it would release the potent greenhouse gas methane, the organic material is processed into biochar, a form of carbon that stays stable for hundreds of years. Following a successful pilot, AMP said last month it had signed on to provide waste-processing services to 1.2 million residents served by Southeastern Public Service Authority of Virginia.
Interest in biochar has surged in recent years, with Microsoft and other prominent buyers purchasing credits from companies around the world. The substance is created by heating organic matter — often residues from agriculture or timber production — in a low-oxygen environment. At around $150 per ton of carbon dioxide equivalent, the credits are relatively cheap compared to other methods for “durable” carbon removal, which is often defined as storing carbon for a century or more.
The world’s biggest supplier of biochar carbon credits is Exomad, a Bolivian company that processes waste from sustainable forestry residues and has contracted for 1.7 million credits, according to CDR.fyi, a provider of data on carbon removal markets. Biomass from municipal waste is a new form of input, however. Few municipalities require residents to separate organic waste; most is incinerated or sent to landfill.
Thousands of items a minute
At the Virginia facilities, AMP’s technology scans conveyor belts of unsorted waste to identify recyclables, organics and landfill items such as plastic bags. As the items fall off the end of the belt, the AI fires compressed air jets that direct individual items to belts dedicated to specific waste types.
AMP’s sorting units, which process thousands of items a minute, are typically set to create waste streams that are 90 percent pure. The system can achieve 95 percent or even 99 percent purity, added Matanya Horowitz, the company’s founder and chief technology officer, but that requires more time and is generally not necessary for biochar production. Once separated, the organic material will be transported to a nearby facility in Portsmouth, Virginia, for processing into biochar.
“The waste goes from being something really carbon intensive to something that’s actually pretty close to carbon neutral,” said Horowitz. “And we are showing it can fit into a lot of existing waste flows. You don’t need to sort of set up a new collection route and green bins.”
Price expected to drop below $100
AMP is currently considering which registry standard to use when it issues credits for the carbon that remains locked in the biochar. Horowitz said the project will process 540,000 tons of waste annually, a number that could grow to 700,000 tons of waste over time and generate hundreds of thousands of tons of carbon credits. The credits are expected to cost between $120 and $140 initially, with the price falling below $100 over the long term as future projects allow the operation scale, said Horowitz.
Some carbon credit projects have been criticized for failing to prove that the credit revenue was essential for implementation. If prevailing economics mean a wind turbine is going to be built, for example, the project should not generate credits. Horowitz said that the Virginia work passed this “additionality” test because diverting the organic waste from landfill is not required by regulation or supported by other revenue sources.
The first use of AMP’s biochar will be as “daily cover” — a layer of material placed on top of a landfill to mop up odors, limit methane release and deter birds from feeding on the trash. Longer term, Horowitz hopes to add other uses, including as an ingredient in concrete.
The post How an AI-powered waste system generates biochar carbon credits appeared first on Trellis.
