Sustainability Under Pressure: AI, Geopolitics, and the Path Forward

- Welcome to another episode of "Supply Chain Frontiers," the MIT CTL podcast where we explore the trends, technologies, and innovations shaping the future of supply chain management. I'm your host, Mackenzie Berry. Today, we're taking you to a live recording during Climate Week New York City 2026, where the MIT Sustainable Supply Chain Lab hosted a panel of researchers and industry leaders to discuss the latest findings from the annual "State of Supply Chain Sustainability Report." Joining us are Dr. Sreedevi Rajagopalan, director of the Sustainable Supply Chain Lab, along with Rachel Schwalbach of C.H. Robinson, Philippine de T'Serclaes of Dassault Systemes, Severine Perrin of CO2 AI, and Heather Ewing of Breakthrough. Together, they discuss how companies are navigating sustainability amid geopolitical uncertainty, the rise of AI, and continued efforts to decarbonize supply chains.

- Good morning, everyone. Welcome to the launch of the 2026 "State of Supply Chain Sustainability Report" from the MIT Sustainable Supply Chain Lab. My name is Sreedevi. I am a research scientist at the MIT Center for Transportation Logistics and director of the MIT Sustainable Supply Chain Lab. I will take the next few minutes to walk you through some of the key findings of the report this year, and then we'll move on to the panel discussion with our esteemed panelists, and we will see what those findings actually mean in practice. So the title of the report this year is "Sustainability Under Pressure: AI, Geopolitics, and the Path Forward." And if you look at the title, it pretty much drives the central tension that we see in this year's report, which is sustainability commitments remain strong, but companies are finding it really difficult to deliver on those commitments because of the difficult environment they have been operating in that could be attributed to the changing geopolitical landscape, the rising advances in AI, and the ever-growing challenges of decarbonizing the physical supply chain. This research wouldn't have been possible without the generous support of our sponsors, our corporate sponsor, C.H. Robinson Foundation, our associate sponsors, Breakthrough, CO2 AI, with Dassault Systemes. We are very grateful to you for your support of this research. All right, a brief overview of the study this year: we collected 1,810 valid responses from professionals across 91 countries working in sustainability, supply chain, logistics, and procurement and other related functions. And the survey was available in five different languages. We have four broad themes for this year's report. The first is geopolitics and governance. We wanted to understand how tariffs and the uncertainty around tariffs basically impact companies' ability to deliver on their sustainability commitments, and also to see if they have their internal governance structures in place to be able to follow through their commitments. The second theme is on sustainability in the supply chain, like, where does sustainability actually live in the supply chain? How is sustainability integrated into the operational decisions, and what are the challenges companies are facing, both in terms of Scope three and overall supply chain as well? And the last theme revolves around freight transportation and warehouse automation. We wanted to understand how freight decarbonization, use of alternative fuels, warehouse automation, and AI are shaping the overall landscape, and what is the implications of these on energy use, emissions, and waste. The first theme, which is focused on governance and geopolitics, what we see is companies' commitment to sustainability has increased. In fact, the question that we asked was, "Have your commitment to sustainability changed as a result of policy changes?" Last year, what company said is a majority of them said that there is no change in their sustainability commitments, and a few person did said that they have increased their commitments. And interestingly, what we see this year to the same question is, like, a significant chunk of them actually have increased their commitments across North America, Europe, and Asia. But the downside is when we ask companies what has been the impact of tariff uncertainty and the overall tariff shock as itself on their commitment, and what we find is, overall, globally, 68% of the companies claim that they had to pause or scale back some of their sustainability initiatives because of the shocks as a result of the uncertainty around tariff. And even worse, what we see is one in three companies actually report increase in their operational emissions as a result of tariff shocks. And more than eight in 10 companies across North America, Asia, and Europe claim that geopolitical conflicts reduce the impact of their climate actions that they are taking, which is really sad. Moving on to the internal governance part, what we see is 83% of the respondents say that sustainability is important to very important to their long-term success. However, if you look at how these sustainability decisions are integrated into their day-to-day operational aspect, we see companies that have formal cross-functional team, 53% of them actually integrate sustainability into their day-to-day operational decisions. These are the companies that have only a single sustainability team, which is not cross-functional, which has maybe one or two people, only 13% of them actually claim that they integrate sustainability into their operational decisions. Another interesting thing is when we looked at companies that have publicly stated the sustainability goals, only 51% of them have a formal cross-functional team structure. And in terms of the board oversight, what we find is only 33% of them have a say at the board level. And this basically shows how the internal governance structure plays a very important role for companies to be able to invest in sustainability, also integrate sustainability into their day-to-day decisions. Moving on to the next theme, which is AI and sustainability, we asked whether companies find AI's benefit to be outweighing the environmental costs. And interestingly, what we find is a significant chunk of them say that they would continue to invest in AI while taking into account the environmental costs of AI. And a small percentage actually say that they have not been investing in AI because they're very cautious about the environmental downside. And overall, 37% basically say that AI's benefit clearly outweighs its environmental cost, and they have been investing in AI to a large extent. We see that, of course, in terms of AI use, a lot of companies are using AI for emissions measurement. They're more than decision making or scenario planning. And even in terms of emissions measurement, a lot of companies are using AI to get data where supplier data is missing, aggregate data from different sources, and also reporting. And when we ask companies what are the different sustainability goals that AI has been able to help them improve, 57% of them claimed operational efficiency, and about 47.7% of them said that AI is basically helping them to reduce emissions. I'm not sure if companies are there yet in terms of finding out the net environmental benefit of AI on overall emissions, but what we see is they at least perceive that AI has basically helped them to measure emissions and also reduce emissions. We also focused on the AI use in supply chain. Where is industrial AI and agentic AI is being deployed, and what are the benefits of these? Only 18.5% of the companies say that they have integrated AI widely into their organizational functions, and about 50% of them say that they broadly use it in multiple functions, but they haven't integrated it fully. And we asked, "What is the implication of using AI in supply chain functions on emissions or on sustainability goals?" And what we see is that those who have integrated AI well into their supply chain, 85.8% of them say that AI actually helps them reduce waste. And also, 84% of them claim that AI is helping them identify emissions hotspot, and 81% actually say AI is helping them reduce their energy use or fuel. What we also find from the report is close to 63% of the firms say that energy consumption is one of the top challenges in their supply chain. I don't think companies are there yet in terms of finalizing the net environmental benefit of AI. Going to the last theme, which is freight decarbonization and warehouse automation, what we find is a lot of companies are trying to decarbonize their freight, but more companies are actually working on operational efficiency, increasing operational efficiencies, rather than actually changing the fuel in the tank. If you see, 43% of the companies still use diesel primarily as a fuel, and also, in terms of regulatory and policy changes, that has a lot of impact, particularly for North America. Close to 43% of the firms say that the regulatory changes have actually made them rethink about their investments, and some have actually paused their investments in sustainability initiatives because they do not know where the policy is clearly headed. In terms of warehouse automation, what we see is 40% of the firms that we surveyed are either moderately or fully automated, advanced or fully automated, and they have also been using AI extensively in their decision making. But when we ask them what is the impact of automation on energy consumption and waste, what we find is, as the level of automation goes up, close to 80% of the respondents say that actually their energy consumption and waste have gone up as a result of automation. And so some of the key takeaways from this year's report is, of course, commitment is strong, but execution has become increasingly harder. And in terms of governance, I think it's important for companies not just to have goals in place, but also have the internal governance structures in place to be able to follow through those commitments. From a supply chain point of view, AI is being increasingly embedded in operational decisions, but not in strategic decisions. So companies are using AI to measure, but I think it's high time they use it to manage, and overall they look at more collaboration for managing their Scope three emissions. There is a clear operational efficiency benefits of AI use, but it's important that we move away from efficiency to net environmental benefit. And in terms of freight decarbonization, more than focusing on incremental gains, it's important for companies to go into deeper transitions and use alternative fuels and get into more capital investment. So the path forward is not just more goals. It's better governance, better measurement, smarter use of AI, and deeper transition in the physical supply chain. Let's move on to the panel discussion. I would want to welcome our panelists. We have Rachel, who's vice president of ESG from C.H. Robinson, Philippine, chief sustainability officer from Dassault Systemes, Severine, head of customer impact, CO2 AI, and Heather, vice president, enterprise solutions, Breakthrough. I would begin with a more broad question for each of you, which is, over the past 12 months, what is the most change you have seen in your sustainability conversation within your organization or within your customers? Is it AI? Is it economics? Is it policy uncertainty? Or is it customer? Maybe we can start with Rachel, you.

- Do I have to choose from one of those listed? I mean, all of the above. Two comments I would make about the last 12 to 18 months: in this room and the people that we talk to all the time, we all know the importance of sustainability. We know how it ties to business case, but there are plenty of people that don't go to New York Climate Week that hear a very different narrative and probably had very different thoughts about what was going to be happening over the 12 to 18 months, especially in the United States. What I do think has shifted slightly, and I don't know if it's necessarily internal to Robinson, but what we're seeing some of is maybe a focusing in on the things that are most important to a company's business and to their stakeholders, which is how we've always talked about sustainability in the past, but this is kind of pressure-testing it right now with so many other things happening externally, geopolitically and economically. Really, what are the things that we know are the most critical for our business strategy to our stakeholders to either help mitigate risk or capitalize on opportunities?

- Yeah, I'll second that. I would say what's really changed in the last 12 months is that we are seeing an increased focus on action rather than evidence building. I mean, I've been in the sustainability sector for over 25 years, and there was such a long time where we were supposed to build the evidence of making the case of why we need to act and why we need to have climate goals. And the good news is that with all the noise and negative zeitgeist that's going on, people are really focusing on action. And it's also true that what you just said, Rachel, is that because there is so much competition as well on capital allocation, you need to make a better business case of what the sustainability is going to bring to you, and that means more action and more tangible results to scale that.

- I will continue towards that. So the first thing on our side is what we get to ask for. So CO2 AI, I will introduce who we are because I'm not sure you're familiar with us. So we are a AI sustainability global company, and we help a large company solve their complex issue toward the supply chain. So what company ask us for, and it echoes what Philippine just said, is that before they asked for us for measurement to answer the reporting question. And today, what they're asking us for is more business case-led use case. What they want to achieve is sustainability impacts, but also profit impact. And the second thing that I see also is who is in the room. Beforehand, we used to work mostly with sustainability team. Over the last year, we are working across function, sustainability partnering together with our procurements colleague, R&D, and even legal and finance when you look at packaging decision, EPR, and so on.

- Breakthrough helps leading shippers manage their freight, fuel, and sustainability decisions with data-driven influence, really helping them get a better understanding of what's happening in their transportation. Much of that here in the North America sits within Scope three, meaning they don't have direct visibility to that. So we are helping them collect that information, being able to ingest that into our system and use that for not only the emissions calculation, but also for building decarbonization roadmaps going forward. And what I would echo is that shift towards that cross-functional team. In my world, the decision makers sit within the operations, transportation, procurement organization. Those are the folks that are paying for the solutions to help them reduce their emissions, but it needs alignment with the sustainability team. In addition to that, it's the fact that the conversation is shifting away from just calculating and reporting emissions into, "How do I reduce my emissions? What is the cost for that? Who are the carriers that can help me execute that? And what are the trade-offs that I might have to consider as I make those choices?"

- So that points to Severine. We see one of the striking findings in our report is the increasing use of AI in sustainability, particularly in the measurement space. So where do you see are the quickest wins in your interactions with companies, and what differentiates companies that use AI to measure emissions versus those who use AI to reduce carbon?

- So I will start with one fact. One of the quick win is unglamorous because it's working on data. I'm pretty sure around you, like, the data is quite messy, unstructured, fragmented, and scattered around different system, file, and so on. So what AI does the best is it ingests, it harmonize those different fragmented data set that you have, and it fills the gaps where it's missing. So that's quite the quick win that I see. Let me share with you a concrete example. So we're working with one of the largest food and beverage company in the world, and they have about 30,000 material to categorize. So materials, they are spread across different system, different country, different formats, and no shared logic. So every reporting cycle is done by hand, and you can't carry over what you have learned from one year to another. So what we have done, AI has built a model, and the model is able to judge the sources of the data to fill the data gap where you need to estimate it, but also to escalate where it's unclear. Because at the end of the day, it's quite important to keep a human in the loop to make the final decision on which data you want to keep and which is not the right one. So to give you a key number to keep in mind, they had 60% of those 30,000 material categorized. With that AI model, we were able to bring them to 96%, which enables them not only to showcase the effort they have done in term of recycling their material, but also to claim that from a packaging and tax liability when speaking about EPR. In term of measuring versus reducing here, I will say it's not a technology gap. It's more of a mindset gap. The companies that are coming to us, either they are looking into data for reporting, or they are looking into data for decision making. And let me give you a concrete example with that. I work a lot with my clients on how to engage their supply chain. When you ask your supplier about data, what is your end goal with it? Do you only want the data capture the primary related data for reporting, or do you want to make a decision out of it? Think of a decision from, "I want to develop a low-carbon alternative for my material. I want to redesign it. I want to have an impact on my packaging." And for me, I see it a clear benefit.

- My next question is to you, Philippine. When we hear from senior leaders, a lot of them say that they want to invest in sustainability initiatives, but when those investment decisions go to the CFO, oftentimes it's about the business case, right? So what stops companies from transitioning from climate goals to action today? Is it because of data and technology challenges, or is it because they're failing to understand the economic value of those investments?

- That's a very good question. It's not a question of technology gap so much as it is a question of mindset, really, and whether or not they want to use the data that they have access to today through virtual twins, for instance, or through a generative AI embedded in virtual twins. It's how they want to use that data. Are they going to use it in a holistic manner that allows to embed sustainability from the design stage of the product through the production, distribution, and until the end of life of the product to ensure that there's more circularity? Or are they going to use it just to report how their supplies are doing? To your point today, what the technology can do, what virtual twins can do, for instance, is that they allow you to track and measure what's going on all through your supply chain. They allow you to adapt because the granularity of the data that you're able to get today means that you're able to make strategic arbitrage in real time. And so you can decide to go for one supplier because it's going to be more sustainably sourced, or decide on another material because it's going to have increased recyclability at the end of the product. For instance, for product, it allows to optimize the weight-material ratio. So enabling you, when you produce a given product, to meet your constraints in terms of regulation, in terms of safety, in terms of what you want your end product to be, but at the same time, make sure that you use the least amount of resource to do that. And it allows you also to imagine. So imagine, for instance, new material, and I think that's where the groundbreaking increase in sustainability can occur as well, is through innovation in new materials.

- Rachel, to you. One of the interesting findings from our study on the freight decarbonization theme is the customer and tender requirements as one of the strongest drivers of freight decarbonization, 68% in North America and 67% in Europe. However, what we also see is 43% of the companies say that their fleets run on diesel. So if there is strong customer drive, why hasn't the fuel in the tank changed? And from a shipper point of view, is it long-term commitments or higher price for a lower-emission fuel, or is it something else?

- First, I'll say I'm thinking back to even probably like three or four years ago being in rooms like this, specifically on the freight side, and the conversation about bringing procurement into the discussion around sustainability wasn't a best practice at that point with freight. And so it's nice to see that this is gaining traction because I do think it's an important piece. So I'm with C.H. Robinson. We are a logistics service provider, and I say that just to kind of set the stage for what our business model is, which is we work with all modes, air, rail, truck. We're global. We don't own any of those assets. And so 99.4% or more of our emissions are Scope three. And we also work with 90,000 customers a year that are shippers, and that's a really, really wide variety. So again, there's some of them that are very mature and sophisticated in their sustainability work, and there's some that they just aren't there. So there's a big range, and when I speak to this, it's going to be kind of the average. So to your point about the tenders, there's a whole piece to that that I think we need to unpack, which is it's very nice to hear that you all are seeing more alignment internally with customers. I would say that's still a challenge. We have some customers are great with that, or shippers are great about that, but getting that alignment before you get into the procurement process around, "What are our priorities, and what are we actually going after?" And having the sustainability people and the transportation people, and everyone needs to be in the conversation ahead of time. There's a range in what we're getting asked. Sometimes it's kind of a, "Tell us about your sustainability program or policy," all the way through, "Price out every single lane you have for EV or for alternative fuels." So I think that that can be a really important driver, and there has to be the follow-on, which is making sure that those are asking the right questions, and then how are you engaging beyond that? So I just think there's more nuance that needs to happen beyond just the RFP, and that's where forums like this or making sure that you're collaborating and having discussions with the shippers, with the carriers, with everyone in the room through whatever venue that is, is really important. And it goes beyond just what the shipper has to be doing too, like, how are we thinking? As an LSP, what can we be doing to educate our shippers and our carriers on what works so they're not just looking at every single lane, like, "Here's where the EV could work really well. Here's where alternative fuels could work really well," so that we're having more of a strategic conversation beyond the RFP.

- Heather, to you, my question is, there are a number of freight decarbonization solutions that are available out there. So what do you think should companies focus on? Should it be a portfolio of those solutions, or should they look at lane by lane?

- Certainly it's a lane-by-lane choice with a portfolio of solutions. There is no one-stop shop when it comes to alternative energies. A shipper is moving goods across regions. They need to understand where their routed lanes are going to pass by electric charging stations, where they're going to pass by renewable diesel fueling stations or renewable natural gas or compressed natural gas stations. They can't just use one solution and apply it nationwide or regional across entire regions. So that is absolutely a key strategy here, is being willing to flex and being willing to use the data to decide where these opportunities might exist. A big part of what my team does is make recommendations for carrier partners that are executing alternative energy today. I'm guiding our shippers to ask more specific questions about alternative energy. It's important for them to know if their carrier partners do have sustainability goals, but what is that going to do to help them actually reduce their emissions? Very little. The questions that a shipper should be asking carrier partners is, "Do you invest in alternative energy equipment? Where do you have this equipment? Would you be willing to partner with me on an investment in this space?" And certainly the longer contract periods is a key point here. When a shipper wants to use renewable natural gas to move goods, that requires a truck with a CNG engine. There is a premium on those trucks. So to ask a carrier to invest in that equipment, it's a big ask, and it's a question that needs to come with a longer-term commitment on that freight. So that's a big piece of this strategy. To use a sports analogy, people need a lot of at-bats in this space. They need to practice. They need to set up the right tracking and audit systems to be able to identify where an alternative energy is being used, what is the feedstock, what's the right carbon intensity to be applying on their emissions calculations. And that's not something that just happens naturally. It takes effort, it takes planning, and it takes a strong partnership with the carriers. So certainly it's going to be a portfolio approach.

- Heather, can I just a follow-on to that? One thing that I think has also changed from your first question over the last 12 to 18 months is for so long when we would talk to shippers, they were so dead set on EV, and that was the only thing they ever wanted to do. It'd be like, "I need all of my freight on an EV tomorrow." And then you look at their freight, and it's in rural Alabama. And now I feel like shippers are more understanding of the fact that there are limitations, and there's a plethora of options that they can talk to, and that comes from education, I'm sure from groups like yours as well. But that has started to shift that people are seeing kind of more of a portfolio solution.

- 100%. And I would add, in North America specifically, I know this was one of the responses in the survey here, we are still seeing shippers want to take advantage of alternative energy despite some of the rollbacks from a federal policy standpoint. Renewable diesel is the first option that I lean into with shippers if they're moving any freight across California, Oregon, and Washington. Those are LCFS states where the cost of renewable diesel is typically at parity with ultra-low-sulfur diesel. Great first solution. Renewable natural gas is available across the entire US, also a great solution if they're willing to make a longer-term commitment with their carriers. EV, much tougher, as I think everybody knows. Works really great in port dray situations, works really great in short-haul situations. It's a little bit trickier with longer-haul movements and heavier freight too.

- Philippine, to you, the question is, companies are increasingly finding supply chain visibility and traceability to be less of a challenge. How do you use digital twins and other data-driven technologies? How do you use that to be able to identify the right product or make better decisions even before the emissions occur?

- Yeah, we can use virtual twins today at a very granular level. It really goes at the school level. It's really super macro, so you can optimize at every stage. And today, one of the main argument as well for moving to more sustainability is the adaptability to the disruption of the supply chain. And what we're finding is that through virtual twins generated with generative AI, you can actually adapt to changing circumstances much faster. In the last 18 months, because of geopolitical uncertainty mainly, disruptions to supply chains have increased by 40%. And so we've had clients turning to us to help them identify how they can fix those disruption and how best to be agile and resilient in reacting and adapting to these changing circumstances. I think we do have the tools. What I was saying earlier, it's a question of mindset, but it's a question as well of knowledge and know-how and sharing best practices. Because even if this technology are available and can have a tremendous impact, people don't necessarily know about it yet. And there's a lot of know-how and knowledge sharing to be done.

- My next question to you, Severine, we see that the collaboration among companies to manage their Scope three emissions have gone up significantly, but still companies point out supplier data to be the hardest challenge when it comes to managing Scope three emissions. How much of this challenge could AI solve? And how much of this does it boil down to building supplier trust, investing in your supplier, training them, and things like that?

- I think it's a mix of both. From a capability point of view, AI is answering the question. And let me give you a concrete example. So when we are working with your customer, the key to engage supplier, and especially their low-maturity supplier whose emission matters, is to enable them with methodology, tooling, and so on. And AI is able to answer that. So we have equipped them with an agent. Something similar is that a chatbot you're all using on a day-to-day basis. Pretty much the supplier is able to feed that agent with the data that relates to their product carbon footprint, bill of material, energy, transportation, and so on. And the agent will enable them to compute a PACT-compliant PCF. You're going to tell me, "Okay, but then how am I sure I can trust the data?" Then this is important to have an expert, a human in a loop, to verify not only the methodology, but also the accuracy of the data that they feed the AI agent with.

- My next question is for all the panelists. We see, of course, AI being increasingly used. So what it takes from using AI in different use cases to deploying at scale, right? Is it data? Is it governance? Is it people? Is it a business case or the broader ecosystem?

- I think it's all of the above, but more importantly, it's really tangible results. And I think sustainability, the tracking, the measurement, the enablement of more holistic strategic decision making through that tracking and measurement of trustworthy data and connecting of all the ecosystem means that sustainability is a positive AI application. And for that reason, it's providing tangible results. And as such, it has a potential to scale quite quickly.

- I think also all of the above. I mean, that's a really hard one. They're all such important inputs to make sure that you're doing this well. Certainly, the way that C.H. Robinson would think about it as we start is a business problem that you're solving is critical first because you're not getting into hobby AI. And when we're thinking about sustainability and the impacts, the net positive, if you're really focused on the things that are most important, that's going to help with the efficient use of it. I think the governance component is huge to make sure that you're doing this in a responsible way from the start. We're all learning, but to have something in place to be able to flex with is huge. And I think you've said it so many times, that importance of the human in the loop, all the process flows that are around that and the change management with it. If what you're trying to do is execute on a broad scale across an enterprise, having people come along with you is critical. You need their feedback.

- I think what strikes me about this conversation is that I actually believe that AI is a contributor to the collaboration that we're seeing across teams internally. So it feeds right back into that initial conversation that we had about the decision makers, the procurement, transportation, operations teams connecting with sustainability. It facilitates those discussions, answers questions, gives practical recommendations for options to reduce emissions in the network. But on the governance side, bad data just means bad decisions faster, which is not a good solution. So certainly data and governance is really critical here too.

- To close the loop on that question for me, I will advise you to keep it narrow. I mean, AI is unlocking so many use cases, but don't try to solve it all at once. You are not doing something over with your company. So name a problem, name a pain point where you can drive sustainability impact out of it, but also bolt your model and look for proof point. And with that in mind, that direction, keep in mind three things. There is this framework that for me puts quite of a guidance, is 10% on your algorithm, 20% on your data and technology, and the remaining 70% is around the people and the process. So I echo what my colleague was saying here. Do not underestimate the change management that goes with your AI use cases, but really keep it narrower.

- All right, thank you so much, panelists, for an insightful discussion. Thank you so much, everyone, for joining us.

- That wraps up this episode of "Supply Chain Frontiers." A big thank you to the panelists for sharing their expertise and insights into the current state of sustainability and supply chain. To learn more about the lab, visit sustainable.mit.edu. This episode of "Supply Chain Frontiers" was recorded live at Climate Week New York City 2026. Our sound editors are Dave Lashinsky and Danielle Simpson and David Benjamin Sound. Our producer is myself, Mackenzie Berry. Be sure to check out previous episodes of "Supply Chain Frontiers" at ctl.mit.edu/podcasts or search for us on your preferred podcast platform. I'm Mackenzie Berry. Thanks for listening, and we'll catch you next time on "Supply Chain Frontiers."