Thursday marked the third Demo Day for the Google for Startups Accelerator: Climate Change program, where startups in the program presented the status of their startup, capping off 10 weeks of programming and mentorship from Google’s robust network of in-house experts, training, and credits to use Google technology.

This year, the 12 companies mostly fell into three broad categories: Artificial intelligence, electric vehicle infrastructure, and providing companies with better data to decarbonize their operations. There are a couple exceptions: For example, Sesame Solar is decarbonizing disaster response, and Bodhi is improving the customer experience for home solar installations.

Google’s startup accelerator programs are all focused on using artificial intelligence, and some have industry themes like gaming or the cloud economy, particular geographies like India or Brazil, or underrepresented founders like Black founders or Latino founders. All the programs are equity free, meaning Google does not take a stake in the companies for participating, and so far 1,100 startups have participated since the programs launched in 2016.

For this latest cohort, all of the participants had to be somewhere between their seed and series A rounds of investment, already generating revenue or with an established user base, with five employees or more, and with the potential to benefit from Google’s Cloud, artificial intelligence and machine learning capabilities.

Matt Ridenour, Head of Startup Ecosystem at Google in the U.S., told CNBC he derives a sense of meaning in supporting climate change startups.

“I care about climate tech for many reasons, but most personally, having three young children, I often think about the world that they are inheriting. When I read the headlines about the dangers of the climate crisis, I feel a personal obligation to be a part of supporting innovative climate solutions to scale,” Ridenour told CNBC. “This is one of the greatest gifts I believe I can offer to my children and future generations.”

The programs are also good for Google business because they get early stage companies using the company’s technology, giving it an early edge over competitors like Amazon, Microsoft and Apple.

“Google sees value in supporting the best startups and founders around the world. As they work with our people, products and tools, we mutually benefit. And supporting early stage companies sparks further innovation in the ecosystem, providing further opportunities for developers to build their business on Google products — like Cloud and Android for example,” Ridenour told CNBC.

Google has hosted three climate change startup accelerators for North American companies in the last three years, and all 33 of the participants are all still operating, a spokesperson for Google told CNBC.

Alphabet-owned Google is itself in the midst of a company-wide push to focus on improving its product offerings with artificial intelligence. Many of the companies in the latest climate change accelerator employ AI and machine learning to help with various tasks such as agricultural soil monitoring, decarbonization of commercial buildings, and improving the process of recycling textiles.

“Teams are leaning deeper into developing AI and ML models to address climate change,” Ridenour told CNBC. “By partnering with emerging technologies like these, startups can have an outsized positive impact, developing solutions and innovations faster and more accurately than ever before.”

Agrology helps farmers adapt to climate change by providing field-level data on smoke, drought, irrigation optimization, microclimate weather forecasts from extreme weather, pest and disease outbreaks. Also, Agrology has a system to monitor the carbon content in soil to help farmers quantify carbon sequestration they achieve with regenerative farming practices and, if they are interested, participate in the carbon credit markets.

During the Google accelerator, Agrology made its product more accurate.

“Through mentorship they received in the accelerator, Agrology was able to build a new, more efficient API that uses integrated Google Machine Learning products, increasing their training and testing dataset by over 400%, and reducing their error rate by 4x,” Ridenour told CNBC. “This will help them deliver more accurate data to farmers so they can grow better and more sustainably.”

Another startup within the cohort, Cambio, is using AI to help companies decarbonize large commercial buildings.

“Once companies have set their climate pledges, they find that data tracking and decarbonization across any real estate, whether it’s owned or occupied, is the hardest part of their sustainability journey. Implementation remains a blackbox,” Stephanie Grayson, a co-founder of Cambio, said on Thursday during the demo day.

Cambio provides a baseline carbon footprint for a building, and then uses AI based on previous building projects and recommendations from leading building scientists and data scientists to provide the customer with a path on how to get that building to net-zero. “The bottom line is we’re democratizing best in class building science across the industry at large,” Grayson said.

“During the accelerator, Cambio was able to connect with Google’s real estate team to get direct product feedback and discuss the topic of decarbonizing buildings,” Ridenour told CNBC. “Armed with Cambio’s ML models, managers can plot an entire real estate portfolio’s path to net zero, a near-term requirement for publicly-traded companies as part of the SEC’s latest carbon emissions transparency proposal.”

Another example is Refirberd, which is using spectroscopy and artificial intelligence to sort recycled textiles, remove buttons and zippers, and send processed textiles to the recycler that can best manage that particular batch of textiles.

Eugenie.AI uses artificial intelligence to help heavy manufacturers track their emissions, report that data for any relevant compliance standards and reduce those emissions with recommendations on how to solve a particular problem.

“As cars become more and more electrified, a variety of startups are tackling the massive EV industry opportunity in creative ways,” Ridenour told CNBC. Indeed, 14% of new cars sold in 2022 that were electric, up from 9% in 2021 and less than 5% in 2020, according to the International Energy Agency.

Batt Genie, one of the startups Google picked for its most recent climate change cohort, was spun out of Venkat Subramanian’s labs at the University of Washington and uses software to improve the function and efficiency of lithium ion batteries, which are used in consumer electronics, electric vehicles and grid storage battery applications.

The battery management system, or BMS, in a lithium ion battery monitors how much charge is left and regulates charging. Batt Genie’s software aims to makes the BMS system more efficient and productive. If a traditional electric vehicle battery lasts for about six years, the same battery can last for 12 years with Batt Genie’s improved BMS, CEO Manan Pathak said on Thursday.

Another startup within the cohort, ElectricFish Energy, is making an energy storage system that both charges electric vehicles quickly which have smart chargers that store cheap, clean power from the grid when it is available.

“The current state of electric grid is fundamentally broken,” Anurag Kamal, CEO ElectricFish, said on Thursday. “We are the only ones who understands that EV charging is incredibly connected to feeding energy back to the grid itself,” meaning that the ElectricFish device can serve as a source of backup power.

Another company working to improve EV infrastructure is Voltpost, which converts lampposts into electric vehicle chargers. Voltpost has partnered with the New York City Department of Transportation to pilot its lamp posts into EV chargers. And Voltpost is also conducting a pilot at the Detroit Smart Parking Lab in Michigan. During the accelerator, Voltpost connected with the Google Maps team to discuss whether electric vehicle charging locations could be added to Google Maps or Android Auto.

For example, Cleartrace provides auditable emissions data for companies.

“The issue is data around the electricity space, the energy space, and the environmental reporting space, is very hard to come by, very siloed, very error prone,” CEO Lincoln Payton said on Thursday. Before starting Cleartrace, Payton was the head of investment banking for BNP Paribas Americas. “I retired from that to address the biggest issue I saw, which is the quality data available in the transfer to the renewable energy world.”

Cleartrace is particularly focused on measurement techniques for Scope 3 emissions — emissions associated with a company’s entire supply chain or value chain, which can be fiendishly difficult to track. It’s also looking at helping companies certify how green their operations are, particularly for processes like direct air capture of CO2 emissions and hydrogen production.

Another data-focused company is Finch, which puts sustainability scores on products to help consumers make more climate-conscious shopping decisions. Finch has a browser extension that works on Amazon and Target websites and gives products a sustainability rating between zero and ten, then suggests a more sustainable alternative if applicable.

“For most of the population who believes in climate change and wants to do something about it, but doesn’t necessarily have more than seven minutes to research it online, this is a perfect solution,” Lizzie Horvitz, the founder and CEO of Finch said on Thursday.

Finch sells the data it gathers from consumer behavior to clients, including manufacturers and investors, Horvitz said.

“We are able to see who is buying what and why — that women, for instance, between the ages of 35 and 40 are twice as likely to buy aluminum-free deodorant as men of the same age and location,” said Horvitz.

This kind of data closes what Horvitz calls the “say and do gap,” meaning the difference between what consumers say they will do in a focus group, and what they actually do at checkout.