iTranslated by AI
re:Invent 2025: How Adobe and Salesforce Drive Sustainability with AWS CCFT
Introduction
By transcribing various overseas lectures into Japanese articles, we aim to make valuable hidden information more accessible. The presentation featured in this project, conceived with that goal in mind, is as follows!
For re:Invent 2025 transcription articles, information is compiled in this Spreadsheet. Please check it as well.
📖re:Invent 2025: AWS re:Invent 2025 - How Adobe & Salesforce enable sustainability initiatives with AWS CCFT (AIM332)
In this video, Alexis Bateman from AWS, along with representatives from Adobe and Salesforce, explains their sustainability efforts using the Customer Carbon Footprint Tool. AWS aims to achieve net-zero carbon by 2040, driven by three pillars: carbon-free energy, water positivity, and circularity. In 2024, AWS published Power Usage Effectiveness and Water Usage Effectiveness by region, and in 2025, it introduced a verified methodology covering all Scope 1, 2, and 3 emissions. Salesforce is pursuing net zero by 2040, emphasizing improved data transparency and granularity, and promoting migration to AWS cloud and cost/carbon optimization. Adobe targets net zero by 2050, aiming to reduce Scope 1 and 2 emissions by 42% and Scope 3 by 52% by 2030, leveraging Customer Carbon Footprint Tool's Scope 3 data to enhance accuracy.
※ This article was automatically generated while maintaining the content of the existing lecture as much as possible. Please note that there may be typos or incorrect information.
Main Content
Session Start and Introduction to the Customer Carbon Footprint Tool
I think we're just about ready to get started. First of all, thank you all for joining us here for this session today. Out of the many choices you have, I'm really delighted to have you join us here to talk about sustainability. Before we dive into the meat and potatoes of the conversation, I'd like to just do a quick poll of the audience. How many folks here are in a sustainability job? Do you touch sustainability or sustainability data? Yeah, a few folks. What about operations? Do we have any software development engineers here? Yeah, a few folks. What about FinOps? Yeah. And I can see about two hands raising over there because of these bright lights. Okay, great.
So hopefully this talk touches on all of those aspects and is a little bit interesting for our builders, for our FinOps folks, for our operations folks. And of course, when you leave here, we hope that you are more engaged in sustainability or at least thinking about it a little bit more. One last question before we dive in. Does anyone know what the Customer Carbon Footprint Tool is? Oh, we got three folks. Okay. By the end of this talk, you all will know what the Customer Carbon Footprint Tool is, and hopefully you'll go check it out, see what it is, and what information it can provide to you. So let's get into it.
So for a quick recap, you are here at session 332, where we're going to be talking about how our colleagues from Adobe and Salesforce are enabling their sustainability initiatives with the Customer Carbon Footprint Tool. My name is Alexis Bateman. I am with AWS, and I lead our sustainability tech team. We enable all of our sustainability goals here at AWS. Whether that's our net-zero carbon goals, whether that's our water positive goals, or our renewable energy goals. We're sort of the backbone for those products and initiatives. Today, we're joined by Eric, who's the Director of Technology Sustainability at Salesforce, and Rao, who's the Senior Program Manager for Cloud FinOps at Adobe. Both of you, thank you for joining us today to talk about how you're using the Customer Carbon Footprint Tool to help enable your goals.
So let's dive right into it. The agenda is super simple. First, I'm going to be talking a little bit about sustainability at AWS, some of the goals that we have, and of course, how you actually get insights and visibility into those goals through the Customer Carbon Footprint Tool, which is available in the AWS billing console. Then, I'm going to hand it off to Eric to talk about Salesforce's sustainability goals, what they're doing, and of course, how we've partnered with them on some of those sustainability goals and the reporting and insights that we're providing to help them. Then, I'm going to hand it off to Rao to talk about similar initiatives at Adobe and how they're driving those initiatives with us. And then we're going to have a few targeted Q&A sessions. I've had a few conversations here at re:Invent in the short time that I've been here, and I've curated a few questions that we're going to talk about. It's in the topic area of generative AI and sustainability, and that's how we're going to close it out. So that's the session for today.
AWS Sustainability Goals and Shared Responsibility Model
Let's get right into it. AWS places sustainability at the core of much of our operations and at the core of how we serve you, our builders, and our partners. We are a signatory of the Climate Pledge. This is the Amazon Climate Pledge for all of Amazon. It's to achieve net-zero carbon by 2040. Amazon co-founded this with Global Optimism, and we are one of the signatories of the Climate Pledge, and of course, we adhere to that same goal.
But how do we actually operationalize this at AWS? We have a couple of different pillars that we're working towards. One of them is carbon-free energy. So how do we actually power our data centers, what energy sources do we use to power them, whether it's solar, wind, nuclear, so we're working towards achieving our renewable and carbon-free energy goals. It's also important that we're as efficient as possible in how we operate our data centers. And then at the core of how we build things is circularity. How do we take back critical raw materials, critical rare earth materials, and component circularity initiatives? Fun fact, last year, in 2022, excuse me, since 2023, we've recycled or sold for the secondary market over 23.5 million components.
We also have a water positive goal. Let me explain this. We're working towards being water positive by 2030, and we're currently 53% of the way there. What does water positivity mean? It actually means that we are returning or replenishing more water than we're consuming in our data center operations. So these are sort of the three pillars that we're driving at AWS sustainability.
And we have company-wide partnerships to help us achieve those goals.
But of course, how does this look at scale? We aim to deliver a sustainable infrastructure at scale across 120 availability zones, across over 245 countries and territories, and over 38 geographical regions. And we've seen a couple of studies that we're actually about 4.1 times more efficient than on-premises. So we're actually providing that efficiency to our customers who are moving to the cloud. But we know we can't do sustainability alone. So that's why, of course, today we also have some of our customers joining us to talk about their story.
So sustainability is a shared responsibility. We at AWS have talked about our goals across carbon-free energy and water positivity. As the infrastructure provider of those cloud services, we can support how we build our servers, what components go into them, how we cool them, where we source our water, how we dispose of our waste, and the energy that we source. So we at AWS control those and can provide those to you. But there's also a partnership with our customers. You, our customers, are leveraging our services. Here, we actually need to support how you use AWS. And that's whether it's through the services that we provide or the reporting and insights that you can use to make decisions. So this is what we call our shared responsibility model for sustainability.
If any of you are familiar with the Well-Architected Pillars, this is one of the foundations of the sustainability pillar in the Well-Architected framework. You can read more into it, but essentially, it's also that shared relationship between AWS and our customers. And with that in mind, our sustainability team really needed to meet our customers where they are in understanding what it is that we're providing, what goals we're working towards, what transparency, what visibility do you need to actually achieve your sustainability goals? And again, not only just the partnership model, but of course, our customers operate directly on AWS, so we are a part of their footprint. So how do we actually start to build out that relationship?
Evolution of the Customer Carbon Footprint Tool and Addressing Customer Feedback
This has been a long time coming, but one of the key initiatives was launching our carbon footprint tool in 2022. Our carbon footprint tool, which I'll show you what it looks like and of course where you can learn more about it later, actually shows customers what portion of AWS's emissions you're responsible for based on your usage. So you log into the billing console and you can see that allocated portion. I'll talk a little bit about how we're building that out and providing that to our customers. But one of the learnings along the way is that we weren't listening to our customers enough. We really needed to know what was actually holding them back in their initiatives and how we could listen better.
So in 2024, we started a customer survey with our largest, most vocal, and most sustainability-active customers and brought them together to gather their feedback. Through this mechanism, we were able to ask what could we do better? A lot of the early voices that we heard were, we need more transparency from AWS. We need to know what our footprint looks like running on AWS. We need more visibility into your operations. So we did a couple of things to answer that feedback. One of them was that in 2024, we released our Power Usage Effectiveness ratings across each region. So this actually provides visibility into how effectively we're using power at our data center sites.
And earlier this year, we also released Water Usage Effectiveness. Similar to power, this also looks at how effectively we're using water for cooling in our data centers. You can see this at a regional level. And then in 2025, as part of some of the things we heard from our customers, we leveled up our Customer Carbon Footprint Tool. Customers wanted to know, how are we calculating these numbers? Can you tell me the methodology? Is this third-party verified? So we actually released a verified methodology of how we're calculating these numbers, and we had a third party come in and make sure that it meets standards. We did this earlier this year, and we also improved our export functionality to make it easier for customers to get their allocated portion.
We also released location-based emissions. This is a different type of methodology. So we have both market-based emissions, which takes into account energy attributes and other credits in our overall footprint, and then location-based emissions, which basically indicates the grid. This is our Scope 2 reporting. And then just recently, about a month ago, we released what we call Scope 3 emissions. I'll dive a little bit more into what that looks like across the entire scope, but now we have the full picture of the AWS footprint.
So our tool includes all Scope 1, Scope 2, and Scope 3 emissions so that customers can actually see where their emissions are coming from, whether that's from operations, whether that's from energy sourcing, or of course Scope 3, which is things like the hardware that goes into our server racks. And we have more plans coming this year. We've hit our success metrics of having fully verified reporting. And then in 2026, we're looking to enhance the customer experience and also further granularity, continuing to provide services to enable our customers' sustainability journey. I'll touch a little bit more on that later, but it's really continuing to hear from our customers like Adobe and Salesforce and respond to them.
How Scope 1, 2, and 3 Emissions are Calculated and Tool Implementation
So I talked about what Scope 1, 2, and 3 are, right? Some of you may be like, oh, that's old news, I know what that means. Some of you may not know what these scopes are, so I'm just going to give you a quick recap of what Scope 1, 2, and 3 are. When we think about carbon accounting, it's very similar to financial accounting. This is actually how we're accounting for the emissions that we're generating as a business. We have what's called Scope 1 emissions. These are our operational emissions, so things like our backup generators, things that we are fully responsible for. Scope 2 is basically where we're getting our energy from the grid, so the energy sources that are actually powering our data centers. And then Scope 3 is basically all of the components, materials, hardware, cement, all the things that go into our actual physical data centers to actually power all of the compute services that we're providing to our customers. So this is sort of a recap of all of that and to provide that visibility.
So how do we actually calculate this assembly footprint, this carbon footprint? We have various data sources across our business. Whether it's hardware components and processors, whether it's data center building materials and equipment, all the way down to cement, energy consumption, grid sources, water usage, all of these different data sources that we're actually taking across the business to calculate this footprint and actually provide it back to our customers. And all of this funnels into what I was talking about, the Customer Carbon Footprint Tool that we actually built out to provide to our customers.
So how do we do this? We go through a couple of different steps. Obviously, we have to ingest millions of data points from across the company to actually generate those numbers. We orchestrate it. So obviously there are numbers coming in on an hourly, daily, weekly, and monthly pace, right? So how do we ingest those numbers and generate them and actually aggregate all of that to our overall footprint? We also often have to apply emission factors because it's carbon, and we also have a team of scientists that are actually modeling the carbon emissions of things like the components that go into our server racks. And then of course, we have to persist that data and govern it. Because it is data from across the company that's going into downstream analysis and action. So how we actually plan and forecast to meet our 2040 goal and of course how we report monthly through the Customer Carbon Footprint Tool.
And I talked a little bit about how we break down all of these different components into stages. So this is how we break down the top-down allocation of our carbon emissions. We actually look across all of our regions, all the way down to our server racks, and we allocate it to foundational services like EC2 and S3, and then we allocate it to non-foundational services, which are running directly on services like EC2 that are on hardware. And then we basically allocate it back to our customers. So across all of your AWS accounts, we actually run an allocation model to say, hey, Adobe, across all of your accounts, this is the portion of emissions that your usage is generating based on the AWS footprint.
So here's a quick screenshot of the Customer Carbon Footprint Tool.
And I also want to quickly introduce Marta Fraga, our product manager, who is here. She actually supports the Customer Carbon Footprint Tool and she's going to be in our demo. We'll get to that a little bit later, but as you can see, this is a screenshot. You navigate from your billing console, you find the tab, and you can see both market-based emissions and location-based emissions. And then there are various data exports that you can get in terms of the scope that you need. So this is actually the market-based emissions for a sample account. This is not for anyone's particular account, of course, but you can see an example here. And then here are our location-based emissions.
So that was an overview of how we actually got to where we are today across all of our sustainability goals and down to the Customer Carbon Footprint Tool. So with that, I want to hand it over to Eric to talk about how we've partnered and worked on this.
Salesforce's Sustainability Strategy and Four Pillars
Yeah, thanks Alexis. That's fantastic. I appreciate you all being here. It's definitely been a journey. Salesforce has been an AWS customer for quite a while, and hats off to you all for consistently pushing the envelope, bringing more sophistication to some very complex problems. So thank you all who've been working on that. Emery Katz and I lead our tech sustainability program at Salesforce. I want to give you all the whole picture of what we mean when we say Salesforce sustainability. I'm going to talk about four simple pillars and then I'll dig in a little bit more on our infrastructure program.
Emissions reduction is probably one of the biggest first starting points that we focus on. We also want to reach net zero by 2040, and that's looking at long-term Scope 1 and 2 targets, total carbon, plus Scope 3 targets for carbon intensity. You can see those reflected up there. To do that, we have to significantly reduce our value chain emissions. That's Scope 1, 2, and 3, and the things that we can control in the shared responsibility model, meaning internally, we want to be more efficient. But we're also dependent on working with key providers like AWS to push the envelope and continue to decarbonize.
We have our Agentforce for Net Zero Cloud product, which is essentially a sustainability management console to help clients track, analyze, report, and reduce their environmental and social impacts. So if you're in that industry, it's definitely something you should check out. It's a great way to do aggregation and footprinting, and then there are agents. As you know, Salesforce is big on agents, and a lot of you are too. There's a lot of great internal and external agents doing some fantastic work on sustainability. If we want to talk about that towards the end, we can.
In terms of energy transition, we're really pushing hard to continue to move the industry forward. For ourselves, we've been 100% renewable energy since 2021, and that's Scope 1 and 2. We're looking beyond that. We're looking at location-based grid emissions. We're looking at Scope 3. We're looking at market-based emissions by time and space, as the market looks, to try and figure out what is the right thing for the world and for the economy. We're helping to innovate specific technologies at scale, be it efficiency, fuels, renewables, removal, etc. We may not have direct control over it, but we're pushing the industry where we can have influence.
We want to be nature positive. Nature positive doesn't necessarily show up in everybody's sustainability reports, but it's important for us. We believe that business health is dependent on nature's health. It's a no-brainer. We're all floating around on this rock together. But studies from the World Economic Forum and others suggest that if we enact nature positive policies, there's enormous economic value to be gained – up to 10 trillion dollars a year in business value by 2030 and about 400 million jobs created. So big stuff to think about.
We want to reduce our impact. We've talked a lot about carbon, but water is a huge focus. This is a very localized resource. We're thinking about it in our data centers. We're thinking about grid resilience. Grids use a tremendous amount of water. This is something we're going to be focusing on a lot more. And we're also thinking about literally investing in upstream watersheds to ensure resilient sources of water to continue to power our data centers and our offices.
We also have a 100 million dollar nature and sustainability fund. We're looking at nature-based solutions for decarbonization through nature solutions like the One Trillion Trees initiative. A lot of you have probably heard about that, OneT.org, a very important initiative. And we recognize that large corporations alone can't do this.
We have to foster an environment where ecopreneurs, meaning smaller mission-driven businesses who are doing a lot of great sustainability work, can succeed. And we're leveraging our entire capital stack to support them through our philanthropy and our venture activities. So when I talk about our infrastructure sustainability program, what we're focusing on for this year and next year are some of the things on the right-hand side. And I want you to focus on data transparency and how data plays into each of these things. I'll talk about that a little bit more in the next slide.
Salesforce's Infrastructure Program and Data Utilization
First and foremost, we're doing public cloud migration. We have a lot of colocation data centers right now, and over the last few years, we've been aggressively rolling out into AWS. That takes a lot of data to understand the right economic incentives, the right sustainability impacts, and then to make the case for continued rollout going forward. But to do that, we also have to right-size our utilization. We don't want to just blindly throw all of our capacity into it without making sure that we're doing it in the most appropriate, most thoughtful, most considered way. So we're right-sizing utilization and we've gotten a lot more efficient over the last few quarters, actually a few years.
In the shared responsibility model, we've thought a lot about cost-driven optimization. For those of you who are FinOps professionals, which it sounds like there are a few of you out here, you're thinking a lot about keeping costs low, cost to serve, unit cost. We're doing the same thing with a carbon lens, and they are very similar paths. It's okay to work together, and one can piggyback on the outcomes of the other, and often, by working together, we can double down on the impact and the incentives, the business drivers. But this includes things like instance choice, using S3 Intelligent Tiering, upgrading to EBS GP3, suspending or releasing on-demand or reserved capacity, things like that, that will increase efficiency and reduce cost and carbon.
We're also looking at our own infrastructure that will deploy more efficiently and effectively for our customers. We think about efficiency, we benchmark where we should be based on what we see in similar customers, and we're trying to give them tools to deploy most effectively. This, of course, means we become more efficient ourselves. But we're also region-aware in terms of not just carbon intensity, but water stress. This is a huge issue. And we provide our customers the option to actually deploy or migrate from colocation to cleaner regions or from AWS to AWS regions. If they prioritize being in more environmentally friendly regions, we're trying to develop software code that is more efficient and more elegant, that avoids anti-patterns that could really degrade resource utilization.
It all starts with that foundational code. That's both human-generated code and AI-generated code, which we really have to think about going forward. We're tracking water, we're tracking e-waste, we're trying to set performance expectations. When we talk about WUE and PUE, we're thinking about climate, we're thinking about the trade-offs between carbon and water, what's the right strategy? Data is super important for that, and of course, we're engaging with various stakeholders, our customers, regulators, investors, and data is really important for those purposes.
So let me reiterate a couple of things. Different types of data have different purposes, and we need a lot of different data for various purposes. Location-based grid emissions are important for a specific use case. Market-based is required, and we appreciate you continuing to give us more types of data. Scope 1, 2, 3, supply chain, as we as customers, as suppliers, our decision making becomes more sophisticated, it really matters. Our customers, this is a huge ecosystem, and we all have to think together and collaborate. That's broken down by master payer account, sub-account, by region, by service level, and I talked about WUE by region, total water intake across all regions, and in the future, various other things that are in my head, and I know you've been thinking about them too.
The four categories on the right-hand side, these are where we use things. Things like footprinting. We have to track our historical data. We do it internally, and we track where we are, where we're meeting our goals, but we also have to do external reporting through our stakeholder impact reports, through CDP, or our 10K filings.
All of that takes the Customer Carbon Footprint Tool and data from AWS to do it accurately. Forecasting and trend setting involves taking historical data and thinking about the future. We have science-based targets that we have to hit in 2040, and we need to know what trends we're on, where we are in our current trajectory, and what our predictions are going forward.
Internally, this overlaps with our internal analysis, but we go one step deeper, because we think about each of our business units. As many of you know, Salesforce is a CRM company, and obviously it is, but we have a lot of other divisions. We do marketing and commerce, and now we have Slack under our umbrella, and Tableau, and MuleSoft, and a lot of products that people know and love, and they each have different businesses and different metrics. And when we think about that, we need AWS data that maps to that to find patterns and opportunities.
The last one is stakeholders. We talk about customer engagement, and this is super important to us. Our customers are always asking about their usage of our cloud, our products, and of course, we are often dependent on AWS to hand over some of that data. And when we set expectations with suppliers and engage with suppliers, we think about industry benchmarking. All of this is needed, and we're thinking a lot about how to ingest that data in the richest way possible.
Adobe's Sustainability Journey and Three Pillars
So with that, I'll close out here. I'm going to hand it over to Rao from Adobe to speak, and then we look forward to the discussion after that. That's right. Thank you, Alexis, for the wonderful context and introduction, and Eric, for explaining Salesforce's sustainability journey. It's truly inspiring to see how this collaboration is driving collective progress across our industry. This is a perfect example of collaboration.
Hello, good afternoon, good morning. My name is Rao Charahala, and I serve as a senior program manager at Adobe. I lead our efforts in cloud FinOps and sustainability. At Adobe, sustainability is thoughtful and intentional. What we want to establish is that sustainability is not a destination. It's actually baked into Adobe's business DNA.
One of the things that I want to focus on here is that Adobe's sustainability journey is built on three pillars. The first is our product pillar, and in this product pillar, we are developing digital workflows that enable our customers' process flows. The second is operational sustainability. In operational sustainability, we are enabling our internal and external operations and efficiencies. The third is, as we collaborate, this is advocacy. We collaborate with our partners and our customers, and we are moving along together as one partner model.
To achieve this strategy, we have actually developed five guiding principles. The first is empowering our customers through our products. Many of you probably know Adobe Acrobat. Most of you use Adobe Acrobat and Adobe Document Cloud, and through these, we are enabling digital workflows and eliminating billions of pages of paper printing. We also enable digital signatures through Adobe Sign, and that reduces logistics.
The second area focuses on building sustainable and innovative workspaces for our employees. 80 percent of our facilities are LEED certified. For example, one of Adobe's recent buildings at our headquarters features on-site renewable energy and a unique water recycling system, which reduces the burden on municipalities.
Our third area is Scope 1 and Scope 2 emissions reduction and productivity improvement within our own data centers. This is achieved by reducing PUE and reducing emissions. The fourth area is employee engagement. We have numerous green teams, with thousands of employees participating. As part of their green teams and their activities, they actually handle 90 percent of our waste recycling.
And the last one is what we call transparency, and this is achieved through our annual sustainability report. The annual sustainability report not only provides all of our environmental parameters, but it also addresses ethical AI, sustainable AI, and data transparency. So these are the five guiding principles that drive our overall strategy. And of course, this is not just Adobe; it's also our partnerships and our customers.
Now, regarding our commitments, our North Star is to achieve net zero by 2050. This is very similar to many of your organizations. However, the milestones you see here, the commitments you see here, are not just aspirations; they are milestones for our progress. The two key milestones you can see here are to reduce our operational emissions, meaning Scope 1 and Scope 2, by 42 percent by 2030. The second goal is to reduce our supply chain emissions by 52 percent, which includes our cloud and supply chain in this category.
Beyond that, we have four different operational milestones: 100 percent renewable electricity by 2025, 90 percent global waste diversion recycling, 80 percent of our buildings maintaining LEED certification, and a 25 percent water consumption reduction target within our own data centers and per employee overall. Driving sustainability is a shared responsibility. And of course, partnership is a very critical aspect.
There are three points to note. The first is data, deriving insights from data. This is one critical aspect. The second is the reporting aspect.
Moving from data to reporting, we have several types of reports. Sustainability disclosures and detailed internal reports to understand the carbon footprint of each product. The third is how we will actually get programmatic data access from the perspective of the Customer Carbon Footprint Tool.
In our partnership with AWS, Adobe's inventory is standards-compliant and also adheres to the Greenhouse Gas Protocol, but where the Customer Carbon Footprint Tool data adds value is in enhancing the data. We create our own data in the form of expenditure-based estimates, so Scope 3 data plays a vital role here in increasing accuracy. In summary, sustainability is a shared responsibility, and it's progress we are making together. Thank you for all the support from AWS and for compiling that information. Let's cherish every pixel, every partnership. Thank you.
The Importance of Collaboration and the Opportunities and Challenges of Generative AI
Awesome, thank you. So, there's actually a funny story: last year, Rao and one of his colleagues, Tridar, said they wanted to come on stage this year to talk about how they're delivering Scope 3. And now we actually have that full-circle moment of announcing that here. Just a little bit more from me to close out here.
As you saw in the title of our talk, collaboration is key, right? In sustainability, we cannot do this in isolation. I'm sitting here with two of our customers, and they're telling us that we need to hit our commitments and goals and push for transparency. And at the same time, they're also driving their sustainability initiatives and demands. There are things you can do. For example, conducting Well-Architected Pillar reviews with a sustainability aspect. You can work with your account manager or members of our technical field community to conduct such sessions.
And of course, as I said at the beginning of today's talk, only three of you have seen the Customer Carbon Footprint Tool. So, by the end of this conference, I want all of you to see it. Please go to Marta and our wonderful colleagues at the Venetian showcase. There's a big sign that says sustainability, and there are lots of cool demos. Please check it out. I promise it will be worth your time.
I apologize for the plug, but I'm also giving another talk tomorrow, where I'll actually talk about how we built some of the core data infrastructure using AWS native services. If you want to see a little bit more of the inner workings of the core data and AI services that are actually built into the Customer Carbon Footprint Tool, please join us there as well.
We have a few minutes left, and I've heard some conversations at this early stage, and there's a topic I'd like to bring up. We have about nine minutes left, so I'd like to address some of these questions and topics I've heard. I'd like to ask a few questions while you're still with us. I think a big theme in this room, which I've heard from several of you, is generative AI. We're here as sustainability professionals, but large organizations like Adobe and Salesforce are starting to address the growth of AI, and we know that comes with increasing environmental pressures, and I'd like to get your thoughts on that. So, Eric or Rao, would you like to comment?
Yes, let me start. I think there are a couple of points, but generative AI is truly a force multiplier. It's certainly a double-edged sword. One is the positive side from an opportunity perspective. There are several opportunities, and one is how we can bring unstructured data together and move from data to insights.
This is definitely achievable from a sustainability data perspective, and we can also take it to the ESG disclosures that we publish. So, this is what we're doing, and it's actually happening. But on the other hand, on the challenge side, there are clearly some problems accumulating.
One is that the more GPUs you introduce, the more energy and water you consume. So, for organizations like AWS, their carbon footprint is increasing. And another area is that these GPUs are refreshed very frequently. Unless we extend the internal life of those GPUs, e-waste is being generated. So, when you look at that accounting part, from hyperscalers to customers, customers are also accounting for it.
So, there are opportunities and challenges here. I think ultimately, when we look at ROI and cost, it also depends on that. We know that sustainability is one of the pillars of the Well-Architected Framework, but what we were trying to do was to introduce sustainability as part of the region selection process. That's definitely something that helps.
Yeah, I mean, obviously we see a lot of scary things that might come out of AI in the future, but I'm pretty optimistic. I think AI is actually going to help sustainability quite a bit. Yes, we have to be careful about data center energy usage, and we will, but honestly, I think AI has the opportunity to transform these infrastructure systems.
When we talk about data centers, being able to have better, more intelligent automation can actually reduce costs, make them more reliable, and more efficient. And even further upstream, for example, in the power grid, I think there's a huge opportunity to predict supply and demand in time and space to get more out of it, and to integrate more renewables into the grid. So I'm excited about it, and it will also accelerate innovation.
I think it was the IEA that said that 50% of the carbon reductions we'll see in 2050 will come from technologies not even invented yet. So, we need to get these technologies online, and the sooner the better. And I think AI will drive that forward significantly. And also for modeling, I think AI will do this really wonderfully. It will help us take massive amounts of data and model things.
And when we think about mitigation, whether it's short-term things like flood risk or changes in sea ice, or longer-term things like drought, or canopy growth and water retention, those are important. And the same goes for policies, like how certain programs, such as carbon pricing, change behavior. AI can generate models of what is most impactful and help policymakers, businesses, and other organizations make better decisions.
These are truly exciting times. Like other industries around the world, there are all kinds of opportunities, but I think sustainability really needs to leverage AI as much as possible. And I think the entire industry will be better for it.
Yeah, great, great comments. Speaking as an engineering manager myself, our developers are building products at an even faster pace than before, leveraging core developer products like Q Developer. And tomorrow, I'll talk a little bit about Bedrock, but we're actually using it to make insights and feasibility more accessible to internal stakeholders and ultimately external stakeholders as well. So, while we have the difficult challenge of needing to continue to operate more efficiently, at the same time, we also see how to leverage AI to build and deliver more sustainably, transparently, and with progress. So, we're experiencing that at AWS too.
Future Outlook and Expectations for Improved Data Granularity
So, I just want to ask one more question. Maybe we'll have time for a third question, but we clearly talked about our journey this year, right? And where we thought we were last year, we thought, "We'll never get there," but we've made it this far, standing on this stage, and in a way, we're saying we've reached this moment. Do we think we're done? Have we checked all the boxes? Are we ready to go home? Or what else do Adobe and Salesforce want from AWS in this area? Could you talk a little about some of the initiatives you're driving and how we'll continue to collaborate?
Yeah, I mean, that's kind of a leading question, isn't it? We'll never be done. As sustainability professionals, I think we have to continue to push the envelope, and again, we've made tremendous progress. It's not mission accomplished for any of us yet, but I want to reiterate that customers really care. And I think the world cares. I know we're in a very interesting political headwind right now, but I think now is the time for business to lead, to show leadership, and to be true to our values. The pendulum will continue to swing for the next 100, 200, 500 years.
So the importance of sustainability is not just about hugging trees. It's really about providing social value, economic value, business value, long-term business survival, risk reduction, improving various things related to business, and reducing costs. So for us, this is not the end. Thank you for the partnership. We want you to continue.
What we want to see is continued accessibility, more access to data, and perhaps better tools, easier APIs, and more up-to-date data. No time lag – I know this is happening, and it's exciting. Data granularity, right? We sometimes look at things at a very high level, but going forward, we're trying to go deeper, deeper, deeper into the stack to understand where the carbon footprint is coming from. That will help us analyze at a better level and make better decisions in our sustainability programs.
And perhaps identifying opportunities within our deployments on AWS, or feasibility, is an area where we'd like more support. We'd like you to help us understand our place in that shared responsibility model. I know there are very clear pink and blue categories, but in reality, you know so much about infrastructure. You know so much about running workloads on the cloud, so if you can help customers find ways to be more efficient, that would be great. So, these are things we can talk about next year, and the year after, and the year after that. We're keeping busy.
Yes, well, I can say the same, but certainly among our customers, more and more are asking, what is the carbon footprint of product X? We're not there yet. I think we're still trying to put it together, break it down, and be able to allocate it to the product level. That takes a little more effort. That's one use case.
And another is that, again, primarily from our European customers, they're starting to ask about the footprint of AI, generative AI. And that, again, I think ties into everything Eric was saying about data granularity and data accessibility. Those are also top items on our agenda, I think. And we're actually excited about the Customer Carbon Footprint Tool console because it provides more access from that perspective.
And of course, one of the exciting things that was talked about on the roadmap is monthly data, the availability of monthly data, right? Yes, really, I think tremendous progress has been made over the last 12 months or so, but we'll continue to push the boundaries on data accessibility and granularity.
That's great. It's always good to hear that my engineers will have a lot of work for a while. So, it sounds like a lot is in store. We have many plans. And I'm looking at Marta, who will be one of the key orchestrators in 2026, so please keep an eye on this space. There are many interesting things as we drive granularity, so we will continue to enable partners like Adobe and Salesforce to innovate on AWS with sustainability in mind.
Thank you very much for joining us today. As I said before, let me plug it one more time. Please come to our demo desk in the showcase, there's a big sustainability sign. We hope to see your faces again at another session tomorrow, where we'll actually do a chalk talk and build out the infrastructure. About how we built what I call the internal workings of the Customer Carbon Footprint Tool. We'd love for you to come there as well.
Also, please give our session a five-star rating. Sustainability is still fundamental, right? Still very important. So we want to continue to bring you these insights and perspectives. We really appreciate your time today. Don't forget AWS SkillBuilder. This is the obligatory slide I have to show you, but we also have a pre-curated set of questions, and I'll be around for a few minutes after this, so if you want to talk to us, please come. Again, thank you very much.
※ This article was automatically generated using Amazon Bedrock, maintaining the information from the original video as much as possible.
Discussion