Alex Hall, CEO of Concrete.ai: “We haven’t seen technological advancements in the world of concrete design and manufacturing in almost 50 years. »
Concrete.ai
Fver the past decade, researchers at UCLA’s Institute for Carbon Management have been working on how to use data to reduce the environmental damage caused by concrete. Today, the startup based on their work, Concrete.ai, said field tests using its AI-based software reduced emissions by 30%, while reducing costs by more than $5 per cubic meter.
This is a big problem because cement, the key ingredient in concrete, is responsible for 8% of global emissions of carbon dioxide, the gas that is catastrophically warming the planet. Yet concrete is ubiquitous – used in buildings, roads and other structures around the world – because of its durability and low cost. Cement, whose main ingredient is usually crushed limestone, is a major producer of greenhouse gases, both because of the chemical reaction that creates it and the fossil fuels needed to heat the kilns where it is produced . If you can use less cement in your concrete, while still maintaining enough strength for the job, this results in a significant reduction in carbon emissions.
“From an impact perspective, you’re talking about three times more emissions than aviation,” said Alex Hall, CEO of Concrete.ai. Forbes. “We haven’t seen technological advancements in the world of concrete design and manufacturing in almost 50 years.”
Although we may think of concrete as a commodity product, it actually encompasses millions of possible formulations with varying structural differences. Different types of concrete use different amounts of cement, depending on the strength needed. For example, concrete used to build columns generally requires more cement than basic concrete slabs. Los Angeles-based Concrete.ai uses generative AI to optimize different concrete mixes, asking concrete makers to swap fly ash or slag for cement, for example, or modify rocks or aggregates that are combined with it in order to use less cement. Its goal is to reduce the amount of cement needed while still creating concrete strong enough for what it is intended to do: reduce costs and at the same time reduce the environmental damage of the material.
“Some of the patented modeling allows us to run 3 to 4 million different iterations on a specific recipe,” Hall said. “Depending on what you are looking for, you select the optimal recipe. It’s mass calculation.”
The startup plans to announce today at the World of Concrete event in Las Vegas that its technology, which it calls Concrete Copilot, is now available for commercial use. To date, the early-stage startup has identified three commercial clients and plans to sign a fourth soon. Each customer represents several concrete plants, and Hall said he plans to have 80 by the end of the year. He expects revenue to reach $1.5 million in 2024, up from just $250,000 last year. Hall said Concrete.ai’s ultimate hope is to reduce the annual global carbon footprint by some 500 million tons by optimizing concrete mixes.
“We have been working for over 10 years trying to understand how to use AI and machine learning to reinvent old traditional materials like concrete.”
Solving the cement and concrete problem is attracting more and more attention from entrepreneurs and investors. Prometheus Materials, a spin-off from the University of Colorado, has developed a process to turn algae into cement using a process similar to the natural formation of corals and shells. Earth
MOON
New federal and state regulations, including government purchasing rules in the Inflation Reduction Act and a recently enacted “Buy Clean Concrete” program in New York State, are providing incentives also encourages developers to consider reducing their carbon footprint.
Concrete.ai researchers are also studying other complex materials that could benefit from its AI-based formulations to reduce costs and reduce environmental damage. Hall said the company’s technology could also be used to help validate so-called “green cement” (a term for low-carbon alternatives, such as those created by Prometheus and Brimstone). and other new materials.
In 2016, Hall, a longtime executive at Holcim, the Swiss building materials giant, saw an article with a photo of UCLA Pritzker sustainability professor Gaurav Sant holding a vial of concrete and explaining how to transform CO2 in sustainable concrete. “I thought, ‘This is the future, I’ve got to get involved with these guys,’” he said. In 2017, while working at Suffolk Construction, Hall, 52, became an advisory board member for CarbonBuilt, a company separate from UCLA that integrates CO2 emissions into ultra-low-carbon concrete.
Meanwhile, Mathieu Bauchy, a 38-year-old computational materials scientist and associate professor at UCLA, was working on the model that underlies Concrete.ai. “We’ve been working for over 10 years trying to understand how to use AI and machine learning to reinvent old traditional materials like concrete,” he said. “It’s just a data problem.”
In 2021, Bauchy (who is also the company’s chief technology officer) and Sant expanded this research to Concrete.ai. Hall joined the company as CEO in September. The company has raised a total of $3 million and is looking to raise an additional $2 million to expand its research and development.
For the past three years, Concrete.ai has been testing its models in collaboration with concrete producers across the United States. He optimized the mixes used in more than 2 million cubic meters of concrete, enough to fill 681 Olympic swimming pools. The company said it saved an average of $5.04 per cubic meter, while achieving an average carbon reduction of 30%, by optimizing mixes to reduce the amount of cement needed. Hall said he was “100 percent shocked” by the results.
Chris Rapp, vice president and general manager of VCNA Prairie Materials, a subsidiary of Brazilian cement giant Votorantim Cimentos with operations in Illinois, Indiana and Michigan, began working with Concrete.ai for its first tests in the field. “Over the last four or five years, we’ve seen the market become more aware of the carbon footprint of these buildings,” he said. “As they [developers] become more aware of their carbon footprint, they put more pressure on us to solve this problem.
In the beginning, Rapp said, the company did its own R&D and researched alternative products. Then, a few years ago, Prairie, which has 25 factories in the Midwest, connected to Concrete.ai and began testing the tool. Given its size, he said, it was a better solution than other solutions. “We are such big consumers of materials that we need something that is scalable,” he said.
Today, Prairie has deployed Concrete.ai in about half of its plants. It is now being used to optimize materials for a large industrial warehouse project that Rapp estimates will require 20,000 to 30,000 cubic meters of concrete. For comparison, a concrete truck only carries eight cubic meters at a time.
“Right now, no other material can reach the scale of concrete: that’s like a ton of concrete per person that we produce every year,” Bauchy said. “At some point we have to decide: Do we want to build new things and repair the infrastructure we have? And if we want that, we need to keep working with cement and figure out how to use it more efficiently. »
