Yan Luhui, Founder and CEO of Carbonstop: AI Agent, Defining the Next Generation of Carbon Management | 36Kr 2025 AI Partner Conference for All Industries
On August 27th, the 2025 AI Partner Conference for All Industries, jointly hosted by 36Kr and China Europe International Business School, grandly kicked off at the Zhongguancun Software Park in Beijing. The theme of this conference is "Chinese Solutions," which is divided into two major chapters: "Chinese Solutions" and "Who Will Define the Next AI Era." Centering around four major topics, namely "The Golden Age of Chinese Innovation," "Can Super Intelligent Agents Become the Core Form of Next-Generation AI," "Chinese Solutions Reshape the Global Technological Competition Landscape," and "The Prosperous Scene of the Integrated Innovation of AI and All Industries," the conference comprehensively presented the latest breakthroughs and ecological systems of Chinese AI, shared the growth path and future prospects of Chinese-style AI, and explored the innovative models of Chinese solutions.
On the same day, Yan Luhui, the founder and CEO of Carbonstop, delivered a keynote speech titled "AI Agent: Defining the Next Generation of Carbon Management."
The following is the content of the speech, edited by 36Kr:
Dear friends,
Hello, everyone. I'm Yan Luhui from Carbonstop. Today, I'd like to share with you a very niche field - AI agents in the carbon neutrality field, as well as our exploration and latest achievements in this area. I hope it can inspire those who are interested in this topic.
The core reason why we've invested a great deal of effort and cost in developing AI agents is the extremely serious and challenging problem we need to address - global warming and climate change. Currently, the global average temperature has risen by 1.55 degrees Celsius compared to the pre-industrial era. Limiting the temperature increase to within 1.5 degrees Celsius by the end of this century is the core goal of global temperature control. This goal was exceeded in 2024, highlighting the increasing urgency of addressing climate change. Enterprises in all industries need to take actions in carbon emission reduction.
Globally, hundreds of thousands of enterprises have actively or passively participated in carbon emission reduction. Currently, there are four core platforms with direct requirements for carbon emission disclosure, reporting, and reduction: MSCI, the Carbon Disclosure Project (CDP), the Science Based Targets initiative (SBTi), and EcoVadis. Chinese enterprises' carbon disclosure, emission reduction, and achievement of carbon neutrality are all closely related to these four platforms. In the past five years (2020 - 2025), the number of enterprise users on these platforms has increased significantly. The number of Chinese enterprises setting Science Based Targets (SBTi) has increased from over 10 to over 1,000, and more than 20,000 Chinese enterprises have participated in the EcoVadis assessment. More and more entities are starting to pay attention to this key issue of climate change.
It's worth noting that many enterprises have shifted from being "compliance-driven" to "value-creation-driven," from passive response to proactive action. In China, 3,700 enterprises have been included in the carbon trading scope, covering industries such as power generation, steel, cement, and aluminum smelting. Leading enterprises in various industries, such as Baidu, Alibaba, Tencent, ByteDance, Hillhouse Capital, Dazheng Capital, and CATL, have also taken proactive measures in carbon management, promising and gradually achieving their carbon neutrality goals. It is predicted that in the next 30 - plus years, China will invest up to 140 trillion yuan to achieve carbon neutrality, indicating an extremely large market space.
However, enterprises still face many challenges in managing and reducing carbon emissions. Firstly, there is a high professional threshold. Let me conduct a small interaction. Please raise your hand if you've heard of "carbon management" or "carbon footprint." Please raise your hand if you know about "carbon neutrality." Please raise your hand if you've used "Ant Forest." (Observing the scene) As you can see, about two - thirds of the people didn't raise their hands for "carbon management/carbon footprint," which shows the high cognitive threshold in this field. Secondly, it's difficult to collect data. Collecting enterprise carbon data requires strong collaboration, involving the operation layer, various departments, and even the upstream and downstream of the supply chain. Especially for leading enterprises, their direct emissions account for a limited proportion, and over 90% of their emissions are "Scope 3 emissions" (from upstream suppliers, downstream consumers, etc.), making collaboration even more difficult.
To address these pain points, we launched China's first AI agent in the carbon neutrality field - Carbon AI Agent in July this year. This product is based on Carbonstop's 14 - year industry accumulation: we've served thousands of customer projects, calculated the carbon footprints of nearly ten thousand products, and accumulated a vast amount of enterprise carbon data. We hope that Carbon AI Agent can help enterprises break through efficiency bottlenecks and achieve unprecedented improvements in carbon management efficiency.
Considering that most of you are not practitioners in the carbon management field, I'll briefly introduce four core application scenarios:
The first is AI modeling. For example, calculating the carbon footprint of a computer was a daunting task in the pre - AI era. With Carbon AI Agent, you just need to input the keyword "calculate the carbon footprint of a computer," and it will automatically sort out the emission sources throughout the product's lifecycle - from the raw materials such as aluminum, plastic, and metal required for the computer to the gasoline, diesel, and electricity used in production. It will generate a complete emission list. The more detailed the input information, the more accurate the result. It can help enterprises quickly build an initial model with an accuracy of at least 70% - 80%, and it supports users to modify and adjust, solving the problem of the high professional threshold and the difficulty of getting started.
The second is data collection and processing. Collecting enterprise carbon data is often cumbersome and involves a large amount of data. Carbon AI Agent can import disordered data into the system, automatically sort it, generate standardized tables, clean up "dirty data," and output high - quality usable data, significantly reducing labor costs.
The third is CDP pre - scoring. The CDP (Carbon Disclosure Project) has a huge influence, representing trillions of dollars in investments. Investment banks and institutions require more than 20,000 enterprises globally to disclose carbon data through the CDP. Fortune 500 companies also allocate orders based on their suppliers' CDP ratings. In the past, enterprises wanted to improve their CDP scores but lacked a clear path. With Carbon AI Agent, by training high - quality scoring logic and data, it can import the enterprise's initial data and give a pre - score (e.g., currently at level C) and clearly indicate what improvements are needed to reach level B, helping enterprises attract more investments and orders in the capital market and customer market.
The fourth is ESG report generation. It's mandatory for listed companies to prepare ESG reports. With the AI + ESG agent, the time cost for enterprises to prepare ESG reports can be reduced by over 90%, and the cost can be reduced by over 70% compared to traditional consulting methods.
Our core product, "Carbon Cloud," has integrated Carbon AI Agent to enhance its capabilities. In the global competition of the carbon management industry, enterprises from the United States, France, etc., used to be in the lead. Now, we're leading globally in the comparison of key functional modules of AI agents and have won the recognition of many international customers. This also proves that Chinese enterprises have every opportunity to create leading examples in the field of large models and agents. In the carbon neutrality field, we hope that Carbon AI Agent can bring honor to our country and lead the improvement of global carbon emission management.
The core reason why we were able to launch the leading Carbon AI Agent in China is "data accumulation." AI is a tool, and it's difficult to achieve good results without high - quality data support. Since its establishment in 2011, Carbonstop started building China's carbon database, CCDB, in 2012. Currently, it has accumulated 510,000 high - quality core data. Based on these data, we continuously train and iterate the model, which is why Carbon AI Agent was successfully launched globally in July this year.
In the future, we'll continue to firmly promote the optimization of Carbon AI Agent. Combining our in - depth capabilities in the carbon field, we'll help more enterprises manage their carbon emissions better. Our company's mission is "to give every product a carbon footprint." We hope to start from accounting, help enterprises achieve climate transformation, and enable consumers to understand the environmental impact of their choices. That's all for my sharing. Thank you!