- 4 hours ago
How climate intelligence moves from models → decisions → protection of real assets and ecosystems.
Artificial intelligence is transforming how we understand climate risk — but how do we turn predictions into real protection for nature and biodiversity?
This panel brings together climate AI innovators, risk leaders, and biodiversity experts to explore how advanced data, geospatial technologies, and institutional action are reshaping climate resilience. From forecasting environmental change to deploying AI at scale, speakers will discuss what it takes to move from insight to impact.
Artificial intelligence is transforming how we understand climate risk — but how do we turn predictions into real protection for nature and biodiversity?
This panel brings together climate AI innovators, risk leaders, and biodiversity experts to explore how advanced data, geospatial technologies, and institutional action are reshaping climate resilience. From forecasting environmental change to deploying AI at scale, speakers will discuss what it takes to move from insight to impact.
Category
🤖
TechTranscript
00:00I
00:36Hi friends, thanks everybody for making it here.
00:39I know we're getting a little bit later in the day and everyone's a little tired, but
00:43I believe this panel is going to be very much worth your while.
00:45We have some absolutely fabulous speakers.
00:48My name is Shane Tchaikovsky.
00:49I'm the editor-in-chief of Diplomatic Courier, and I'm just going to tell you a little bit
00:54about the panel, a little bit about the panelists, and then I'm going to let them kick it off.
00:58So, when we talk about AI for nature, by nature we're also talking about AI for good.
01:05We have to measure the costs and the risks of AI.
01:09Those are economic, those are ecological, they're societal.
01:12We have to measure those alongside how AI can be mobilized for positive impact.
01:18And that's very complicated.
01:20We have energy use, we have water use, we have efficiency, we have efficacy.
01:25Is this the right tool for the job?
01:27How do we navigate those kinds of questions?
01:29It's very complicated.
01:31But in fact, these are the kinds of people that can help us answer that.
01:35They're at the forefront of thinking about this, about predicting, measuring, deploying
01:40the best solutions for the problem at hand.
01:44So today with me, I have, just to my right, I have Carolyn Gray.
01:47She's the co-founder of Treferro.
01:50Beside her, we have Shashin Mishra.
01:53He's the senior vice president of AI Dash.
01:56Beside him, we have Sarika Naik.
01:58I think I said that right.
01:59I hope.
02:00Thank you, Sarika.
02:01She's the Group Chief Corporate Responsibility Officer at Capgemini.
02:05And then on the far end, you may recognize him if you've been here for the last couple
02:09of hours.
02:09We have Ashpuri, a partner at Light Rock Capital.
02:13They're going to be sharing some of what they're about professionally, organizationally, and
02:17from their personal perspectives.
02:20Basically, we're going to be looking, as I guess the title says, how to use AI to predict,
02:25how to protect, and how to preserve.
02:28So to kick this off and to let them actually introduce themselves, a quick open question
02:33to all four of you.
02:35How do you ensure that the potential ecological cost of your AI usage is quantified in a way
02:42that you're not just doing a net good, but that you can prove it?
02:46You can prove it to your customers.
02:48You can prove it to regulators.
02:50You can prove it to investors.
02:52And you can prove it to nature advocates.
02:54And let's just sort of go from left to right here.
02:57So we'll start with you, Caroline.
02:58Please.
02:59Thank you very much.
03:01I really appreciate the introduction.
03:02So Caroline Graves said from co-founder and go-to-market lead from TreeFairer, I think
03:08from our point of view, and I've got a confession at the start of this talk, which is never good,
03:13never a good start, but my confession is I am naturally not a frugal person.
03:18But when we started TreeFairer, my co-founder, Jonathan, one of our values at the core of TreeFairer
03:27is all about frugality.
03:29And it's absolutely run through the course of the company.
03:32So my previous background is from SaaS companies working in SaaS businesses, which is all about
03:37people munching and crunching on your platform 24-7.
03:42TreeFairer is a data company that we serve up data insights on nature-based assets.
03:48And so for us, it's frugality in how we have architected our platform, how we utilize our
03:55platform, how we work and co-engineer solutions with our customers to ensure that the amount
04:02of data and insights we're serving up are given to them on a timely basis when is most relevant.
04:08And that, for us, is the absolute critical thing of making sure we manage the uses.
04:13And then, of course, we do manage it, and we give ourselves various challenges, both from
04:18customers and our own internal point of view, to structure the data in a way that we can
04:23collapse things very quickly or we can bring it back up very quickly to minimize the impact.
04:28As we've said at TreeFairer, it would be kind of ridiculous to do something and that then
04:35puts more damage into the atmosphere utilizing our AI than it does than what takes away and
04:43the benefit it delivers from the data insight we provide.
04:48Thanks, Caroline.
04:49Shashen, please.
04:51Thanks, Shane.
04:52I'm Shashen Mishra.
04:53I lead the nature and environmental solutions business for AI-Dash.
04:58Okay.
04:58Sorry.
04:59I'll start again.
05:00I'm Shashen.
05:00I lead the nature and environmental solutions business for AI-Dash.
05:04When we first started, we got a lot of questions about the good that our AI can do, the impact
05:10of the AI.
05:11From my own background, before I joined AI-Dash, I wrote a book on responsible use of AI.
05:18And as part of our work, at the end of last year, we launched an industry alliance, which
05:25is also called AI for Nature.
05:28So, it's an industry alliance we've launched with some of the largest environmental solutions
05:32companies in the world, like Rambol, Atkins, Arup, as well as with the Steward of Nature
05:40in the UK, Natural England, and the Chartered Institute of Ecology as the founding members.
05:45And the goal has throughout been to build technology that can not just make an impact, but also make
05:51an impact while doing it at a lower cost to the nature itself, lower cost of operation,
05:59and better results than whatever alternatives exist.
06:02So, for us, the frugality or the good of AI has been around, are we using data that can help
06:09us create value that can be far more and far more efficient and also far less impactful than
06:17anything else that goes to replace it or anything that it is replacing?
06:21One example which we calculated early when we first started and we realized when we were
06:27trying to understand the impact and still stays with me is if you have a helicopter flying for
06:32an hour collecting data, it's going to emit about a ton of carbon.
06:38But instead, if you replace it with a satellite image and you calculate the impact of that image
06:43throughout the life of satellite and what that image then costs, it's about instead of a ton
06:48of carbon, it's about 80 grams. Now, that's the kind of impact that the use of technology can bring
06:53in. Same goes towards use of AI as well. So, when you're using AI to apply on that image that's
07:00captured by the satellite, you're replacing possibly a flight, you're replacing diesel trucks being run,
07:06and you're replacing a lot of travel being done by people to go and walk the site, take the time,
07:11while you're making it faster and more impactful. So, that's how we approached it. Happy to be here.
07:19Thank you, Shahjan. Surika. Thank you, Shane. And yes, you got my name right.
07:26It is one of the easier Indian names for sure. My name is Sarika Nayak. I lead Capgemini for the
07:34global corporate responsibility. What this means is that I'm responsible for all the initiatives that
07:39we do globally for our people, planet, and society. On a day-to-day basis, this means that what we
07:45do
07:45towards, you know, how do we embed environmental sustainability, you know, social impact and
07:51inclusive innovation in our day-to-day work and embed it as we grow as an organization. For those
07:57of you who may not be familiar with Capgemini, we are nearly a 60 years young organization spread over
08:04more than 50 countries and over 420,000 people. Frankly, when we look at AI, you know, our brand
08:14promise is really using technology to build an inclusive and sustainable future. And clearly, our
08:19clients count on us to help them in their AI journey towards business and, you know, technology
08:26transformation. So, for us, AI is something that we breathe, eat, hear every single day. Having said that, we
08:33recognize that this is one technology since the evolution of the modern, you know, technology
08:38evolution way back in the 20th century, that this is one technology revolution which can go both ways.
08:45So, one, it can create unlimited possibilities for the human race. At the same time, it can really bring
08:51the human race to its knees. And which is why, as an organization, we have been very thoughtful in the
08:56way we use AI in our own, you know, operations. We have been doing all that we can to understand
09:04the
09:05impact of the AI usage when we, you know, implement it for our clients. At the same time, we are
09:10also
09:10being thoughtful to make sure that we are using more targeted or specialized models. One example,
09:17what you gave, you know, Sachin, how can you use specialized models instead of large language models for
09:22everything. So, I think, you know, as an organization, clearly, AI is something that we really live in.
09:28And at the same time, we recognize that we have a responsibility towards the people, planet, and society.
09:36That's fantastic. Thank you, Sarika. And now, Ash, please.
09:39Yeah, thank you. Good afternoon, everyone. My name is Ash. I'm a partner at Light Rock. And we are a
09:46global
09:47platform investing in technology companies, essentially, with a big focus on sustainability.
09:53And we measure this across three dimensions, which is people, planet, and productivity.
09:58And our intention is to find companies which are solving really complex problems
10:02and take them on a global stage with the capital and also the network. We're a global platform,
10:07like I said, so we're based in the Europe, with offices in the UK, but also offices in India,
10:13in Africa, and also in Latin America, so fairly global. And our ambition is to take these companies
10:18across the world. To your question about AI and sustainability, I'll give you a two-pronged answer.
10:25We do it in-house, and we also do it through a proxy of the companies that we're investing in.
10:30When it comes to AI, I think it's an inevitable change, and it's an important change. In some ways,
10:36on my previous panel, I was mentioning, I think AI will drive sustainability faster.
10:41And I can explain that. But more importantly, if you think about the source of the AI,
10:46which is the data center, the data center is where we are focusing. So we're looking at different
10:51angles of how to make AI more sustainable. And so from that perspective, we're trying to see all the
10:57feeders that go into the AI requirement. So AI, the data center, how do you ensure that the data center
11:04can become greener? So the electricity that goes into the data center, we're looking into solutions
11:08that allow the data centers to become more green. Two, about talking about large language models
11:14versus specific models. We're trying to understand how do you use certain architectures or certain
11:19type of technologies to allow for specific use cases, as opposed to a blanket use of an LLM.
11:25Three, we're trying to understand, even from an electronics perspective, what kind of certain
11:29electronics can be used to produce these tokens, which are used as a currency for these models, to reduce
11:35the consumption of those tokens, and try to understand that. And then four, when electronics
11:41will use the electricity, you will have to cool the data center. So how do you have more sustainable
11:46ways? And lastly, the heat that is produced, can you use it for certain applications? So we're trying
11:52to find all the angles that will allow the net carbon footprint that the data center is producing
11:58to be dramatically lower. And as a result of which, the token that you're using for running an AI model
12:05will have a lower carbon footprint, because you're also mitigating it by putting in other areas.
12:10So it's a very holistic approach, because we do believe AI will be good for society. It depends on
12:16how you apply it. But at the same time, if we can reduce its footprint and improve its applicability,
12:23that will actually have a more positive effect on the planet that we live on.
12:28Fantastic. Thank you, Ash. So this kind of situates who everybody is and why they're on this panel.
12:35So I hope that's really whet your appetite. A quick note before we get into the questions,
12:39we are trying to do a quite conversational format. I've obviously got a notebook here,
12:43it's got some questions in it. But as we go along, as somebody finishes answering a question,
12:48there's a very good chance that a panelist is going to want to have a comment. And by golly,
12:52they're going to comment. So things might get a little bit hectic here. Who knows? But let's get
12:56crazy. All right. Caroline, we're going to open up with you. So you've talked a lot about first
13:02mile visibility. That's very core to what you're doing at Tree Faro. And it seems to be a really
13:07big value add, but something that's been really missing historically, because it's been a blind spot
13:12until, I guess, just recently. So what I'd really like to know is how has this first mile visibility
13:18changed in the last year or so? Why does it matter to sustainability? And very specifically,
13:25how can this concept and this practice of first mile visibility really figure into leadership
13:30decision making? Yeah. So there's a few things really that we looked at when we're starting
13:37Tree Faro, but also just thinking through first mile visibility. So 67% of cost sits in the supply
13:45chain at the first mile. But as you rightly pointed out, there is literally up until now,
13:51it's been used retrospectively, people on the ground, you know, two years later, organizations
13:56get information about their first mile crops. And also there hasn't been the same level of
14:02weather and climate significant events, as well as the food security challenges we're facing,
14:08as well as the degradation really of land and crops and so on. So all of those things have become
14:17now a very significant factor for organizations to look at, given, frankly, the cost base.
14:24And when we started Tree Faro, we looked at it from a couple of reasons, a couple of factors. One,
14:29you know, from a personal point of view, Jonathan, our chief executive, he was at JP Morgan running
14:37global risk and AI. I was just about to be the bankers, but sorry for any bankers here. No. And
14:45he was, you know, for him, he felt very passionate about he could see the assets where people were
14:52investing in these nature based assets, and how significantly important they were to so many
14:57things. And what we could also see is that there's a massive change in technology, satellite technology,
15:05which we've talked about, which personally, I still think is a very underutilized under talked about
15:09thing. And frankly, it's thanks to all the space work that he's enabled and satellites now becoming
15:14a commodity. So it's data that we can use to see what's going on on the ground. Also, the cost
15:22of
15:22compute power, as we know, has gone down, we can argue if that's good or bad. But of course,
15:28fundamentally enables huge amounts of data to be wrangled with. And then, of course, the improvements in AI.
15:33So they are the big technological changes that have enabled something like progressive
15:40technology at Tree Farrah of us, because the difference is, is we're not a workflow platform,
15:45which obviously everyone's relying, there's nothing wrong with that. But previously, organizations
15:49have relied on that to sort of type into the computer what's happening. Now, what we're doing
15:54is we're actually assessing using satellite, radar, infrared, and we are engineering science at scale,
16:01which is, I have to say, I've been in technology for 30 years. You know, I've worked in robot software
16:07before I've worked in, you know, workflow technology before I've worked in lots of different technologies.
16:12And this is definitely the most complex technology and progressive technology I've worked with.
16:17But fundamentally, what, you know, they are the big changes. And then that coupled with all of these
16:23changes in climatic events, if you look at Nestle alone, they lost 16 billion dollars off their market
16:30cap. So their stock market price due to the fact of cocoa and coffee becoming the price of cocoa and
16:38coffee
16:38going up. So it really has now got board level attention, which personally, as somebody, and we're
16:45all I think we're all here for probably the similar reasons. I'm a great believer that, you know,
16:51with organizations, organizations have to be commercially successful. But I sort of what I'm
16:57excited by is I finally think proving where the flow of capital comes from and trying to influence the
17:04flow of capital by really demonstrating, you know, at a coffee location, what's going wrong and what's
17:10happening, can then and now with the connection of boards realizing that it knocks money off their
17:16share price, and they are actually competing with hedge funds, when they are buying this stuff.
17:22This type of data and this type of thing is really bringing sustainability now to much more of a
17:28central business. Now, we all know the political situation, we all know that sustainability can sometimes
17:34be seen as a bad word right now. It's not popular everywhere we go. But frankly, when you attach it
17:42with that type of losses, or those type of changes happening, and thanks to the technology allowing us
17:47to see what's happening on the ground, I'm super excited about what this ability and what this is doing.
17:55Beautiful. Thank you, Caroline. I really enjoyed those examples. And I feel like I know a little bit more now.
18:01Shaxing, we're going to move to you now. So you lead B&G AI, and this represents a pivot within
18:07AI-DASH
18:08just over the last year, as you've explained to me, towards really environmental solutions.
18:13You've described that pivot as involving two primarily things. One is preserving the public trust,
18:19and the other is creating long-term scalable climate impact. So when you're in, can you sort of walk us
18:27through how this pivot is handling these considerations? Sure. Yeah, so until last year,
18:34we were providing solution for no net loss of nature, and it was a technology coupled with AI where we
18:41would baseline the land and provide a solution that the users could use to do their planning, and then
18:48work on delivering it. But then we realized that there is a big question of trust on the data.
18:56So when somebody is trying to demonstrate, and they want to be able to show that there is a no
19:01net loss,
19:02traditionally that term has been associated with offsetting in carbon where you are doing a carbon
19:09emission, but then you're just protecting it somewhere else, and then that has eroded trust over time.
19:14So our goal was always to be able to help our customers, users achieve that by being able to build
19:21annual data, very high-resolution data that they can use to show how the land has been changing.
19:27And the other benefit is that we, through the technology, they are able to do it on the land where
19:32they are creating the impact in the first place. So we realized that there's the best way to put that
19:40forward and best way to help our users stand behind our data and the trust is by showing the trust
19:46ourselves.
19:46So by becoming an environmental solutions company, the responsibility and the liability on us shifts.
19:52We are not just a tech provider anymore. We are taking responsibility to take our users from the
19:58start to the end of the journey. And then we become a deeper partner with them to bring that trust
20:03into
20:03the data. And then in terms of response, what we're finding is that this is probably what the users were
20:09also looking for. They may not have asked us that explicitly, but then, in a way, our response was
20:16in response to what they were asking us for doing. Sorry, the second part as well.
20:27We just asked sort of how the pivot was handling the two considerations. One was preserving public trust.
20:32First, the second, they're creating long-term scalable climate impact.
20:36Yes. So in terms of long-term climate impact, I think there are a few ways in which that does
20:41it.
20:42So every time when somebody's, an organization's looking at a land, they're not just trying to find
20:48out what's there today. There is always a plan or a goal behind it. The goal could be to create
20:53a value
20:54by improving the nature and delivering that for the community. It could be a habitat bank. It could be,
20:59or it could be a development plan where the development needs to go in. It could be a data
21:04center. It could be a development plan that the local government is looking at. But as they do that,
21:11the first part of long-term impact is being able to make sure that when the development's done and a
21:19large part of what was there in the nature is gone, the rest of the land is uplifted year on
21:24year to be
21:25able to bring it back. And those are not short-term plans. They typically last from 15 to 30 years.
21:30So going back every couple of years, if not more frequently, showing how that land is improving
21:36and eventually demonstrating the full known net loss being achieved. So that's the first part of
21:43the long-term answer. The second is any organization that's operating at scale, looking at large number of
21:49sites, large amount of land, has other sustainability targets, other reporting that they need to do.
21:55So often this reporting would then fall back into other areas that have traditionally supported it,
22:03so through consultants. And then data sources will shift. The quality of reporting goes down.
22:08But the engine that we built can produce that high-quality information that they can still utilize.
22:14So our engine's not changing as a result. What we're doing is just adding capability to build more
22:20things on top of it, support more reporting on it, so that if they have to do things like CSRD
22:25reporting
22:25starting this year, TNFD reporting, all of that can be natively supported and utilize the same
22:31high-resolution information. The third part of the long-term impact is around the climate change itself.
22:39So we know that as users are trying to plan for that change, being able to support more reporting,
22:45they're also planning change on a land which has changing climate. So the vegetation as it exists today,
22:51or as has traditionally existed on that land, may not be the best vegetation to do long-term planning.
22:57And we see that in many places. For example, in Brussels, in the city of Brussels, they've been
23:02replacing native trees with trees that are native to south of France. And that's the kind of planning
23:07and vision that needs to go in to bring vegetation that can improve the quality of nature, but can also
23:13withstand the test of time, given the climate change. So supporting, again, from long-term change
23:18of view, that's another thing that we're building in to help our users. It seems to me like a lot
23:23of this
23:23is really about having a relationship with your client, with your project, and with the land.
23:28That's actually really beautiful. Thank you. Thank you.
23:31Sirika, I'd like to move on to you now. So Capgemini is really interesting because it's
23:35quite uniquely situated as both a consultancy and as a developer and deployer of solutions in and of
23:43itself. So I'm thinking a little bit of your Morocco example earlier, but also if anything
23:49similar comes to mind, these are projects that seem to have quite specific use cases.
23:56And so I'm wondering, how do you decide what to develop, how to go about developing it,
24:01but then very sort of significantly from this sense of specificity, how do you look to scale from that?
24:09Thanks, Shane. And yes, we do sit at a very interesting intersection where we get the opportunity to work on
24:19the
24:20sustainability transformation of our clients. And at the same time, we also use technology
24:26for building use cases such that we are able to create more solutions that can be more broadly used for
24:33environmental sustainability, for building resilience and, you know, making sure that we are using
24:39technology for good. Frankly, a lot of these, you know, discussions start with our clients,
24:47you know, conversations that we have with them, we look at our own operations.
24:53Conferences like this also help us get ideas into what are some of the areas that we can leverage our
24:59expertise in building solutions and use cases. And most, also most interestingly, like I said,
25:06we have a dynamic workforce, very young and very, you know, vibrant about wanting to use AI.
25:13So a lot of the solutions also come from our own employees. You know, in the green room,
25:18we were talking about one of the examples, which I called out, which is the algae blooms in Morocco.
25:24I wanted to talk through on that. You know, we have a challenge, which is called Tech for Positive Futures.
25:30This is a global challenge where we encourage our employees to think of, you know, how to use technology
25:37for solving real world challenges around sustainability. And we are in the fourth edition of this challenge.
25:44And, you know, here is where our employees work across with NGOs, with local governments.
25:51They partner amongst different countries to come through with solutions. And one of the solutions
25:57that I love to talk about is, you know, the one that we created in by the Morocco team.
26:03As you know, Morocco sits on has has a lot of water scarcity and water challenges. Many countries have now,
26:09frankly,
26:10but Morocco in particular, and the teams, they found that, you know, the water was getting contaminated
26:16in the reservoirs because of the blooming of algae blooms. You know, these are algae, which are not
26:22necessarily good. And they actually leave a harmful impact on the ecosystem. And the team worked with the
26:29local water body. They worked with a university, local university for the local context. And they built,
26:36you know, a water-based drone with eco-sensors. And it has an integrated AI capability such that it can
26:45identify the harmful algae. And it can send the signals back to the team, giving them three days
26:53advance notice before the water gets contaminated. So this is a great example of how you can predict in
26:59advance so that the relevant teams can take necessary action. You know, another example that comes to my
27:06mind, which was built by the India team, it's called Invasive X. Many of us think that all plants are
27:11good. I'm assuming many of us think that, right? All vegetation, all plants are good. Well, actually
27:18not. We have a lot of plants which are invasive in nature. These are plants that actually, you know,
27:24spread very rapidly. They cover the healthy plants. They are not edible by the, you know, animals in the
27:31in the forest. And what they do is they push the animals out from their ecosystem, thereby, you know,
27:38upsetting the fragile, you know, man and animal balance that we have closer to the forest region. So
27:44long story short, the India team recognized this challenge in a forest in Erkut, which is in the south
27:51of India. They worked with the forest department and one of the NGOs. They used AI to identify what kind,
27:58what looks like, you know, among all the other vegetation, what looks like invasive species.
28:04And they used a mechanical bot to help pull that out and, you know, manage it. Now, you know,
28:11one might think that, oh, it's, it's fine. You know, it's not such a great, it's not a very big
28:16thing.
28:16But imagine in a forest where there are hundreds and types of species. It's a very, very interesting
28:22way of, you know, using AI to pinpoint exactly what needs to be removed. Now, if you look at
28:28all these solutions, you know, there's a common thread through them. One, you know, while they have
28:35the local context, because they are, they are challenges which were built for a local challenge
28:42that the team faced. But the solutions that are built, you know, the underlying capability
28:47of prediction, the underlying capability of environmental intelligence and the corrective
28:53action, it's, it's actually, can be applied in a broader sense to many other, you know,
29:00such challenges globally. And the other part which we found very interesting as, as the Capgemini team
29:05is that none of these, none of these solutions were built in isolation. They were built with strong
29:12collaboration with local bodies, with, you know, universities, and of course, our tech teams.
29:17So clearly, you know, when, when we look at such solutions, we feel that, when not only should,
29:24technology should not be the starting point, it should be the local context, but in partnership
29:29and collaboration with the local agencies, such that then we can really build lasting solutions.
29:38That's actually very cool. I really love this sort of internal hackathon thing that
29:42Capgemini does. It seems like it's obscure employee benefit, huh? I love that.
29:48So, Ash, we'll move on to you now. You've been warning for some years that we're far too focused
29:54in our environmental conversations on mitigation at the cost of rejuvenation and adaptation.
30:02But then you've also noted that this does seem to be changing, perhaps, especially when we talk about
30:08restoration. So from your perspective as an impact investor, what do you think is, is sort of behind
30:15that change? And how will that change our sustainability story in the years to come?
30:21I think the, it's a very interesting question, actually. The reality is, where mitigation was the
30:30core agenda behind climate investments for the last decade. The reality is, we haven't really met our
30:37climate goals. And as the climate changes, the need for adapting and living in this ever-changing
30:45climate world will become even more important. So investments into climate adaptation themes are
30:52becoming ever so more important. And you're starting to see that as a strong VC theme. You're starting to see
30:59a lot more companies come in this space. And you're starting to see different facets of society building
31:06different things for different angles. And let me give you a few examples. I mean, TreeFair is a very good
31:12example of adaptation to a certain angle, because the way we are living now, we need climate intelligence data.
31:20And how an AI-2 is helping in that regard, right? How does that intelligence data then be used?
31:27Because what is happening is, we're starting to see microclimates form. Microclimates in return. And by the
31:33way, our radar systems today on Earth cannot do that, which is why you're having satellite-based monitoring
31:38done to create and understand how microclimates have been formed. But that climate intelligence data is
31:43super valuable to understand how do you do crop, for example, if you're growing certain things.
31:50How is that used for transportation? How is that used for airline industry? How is it used for insurers?
31:56How is it used for the hedge fund industry? So climate intelligence is becoming a way because
32:00people have realized the climate is going to have its own patterns. We can't predict that necessarily,
32:06but we've got to live with it. And so having more real-time information is becoming a booming industry.
32:11At the same time, as the Earth is heating up more and more, we need to adapt the way we
32:18grow food.
32:18We need to adapt and see how our crops can become more resilient in this infrastructure.
32:24And so from drought-resistant seeds to the way we are watering our agriculture to the way we are actually
32:34growing plants without actually using that much energy are all fascinating ways of adapting to the
32:40way we're doing. The third thing is we all need to live somewhere. We all need, and this is one
32:47of the
32:47largest carbon footprint creators. So how do you reduce the carbon footprint of the very infrastructure that
32:54we use whilst we adapt to this is construction materials? You know, the way we're actually
32:59building now has a lot less carbon footprint. The way planning is done, I mean, Airdash is doing quite
33:05a bit of work in that space. We're actually restoring the entire ecosystem after something has been built
33:12or is so that nature can kind of continue to thrive. It's another area. So the point I'm trying to
33:18make
33:18is climate adaptation is becoming a very, very important theme for investors. And because
33:25sustainability used to be a choice, but it's actually become a growth opportunity. You can actually
33:32build really large businesses today and you can actually make money. But whilst you're making money,
33:38you can also do sustainability properly and add something positive to the planet. This creates a
33:45nice flywheel effect for improving the earth whilst also businesses are to thrive. And that's the beauty
33:50of where climate adaptation is going. It's interesting. It feels as though that's the promise that we'd
33:56been hearing for time that eventually tech was going to catch up and we'll be able to make money doing
33:59good. Totally. Well, thank you. We were a bit optimistic in the idea that we'd be able to jump in
34:04and do
34:04conversations and we're going to have all the time in the world. So we're going to try and keep this
34:09next
34:09round a little bit shorter. We've got about 11 minutes left just for your reference. Shoshan,
34:15we're going to come back to you to start this one off. You started to mention no net loss as
34:19sort of
34:20a foundation in the last answer. And when we'd spoken about that in the past, you also spoke about
34:27a two-prong approach with this. So this was on the one hand, predict the risk, and then on the
34:33other,
34:33actively make the land more resilient to that risk. So my question is, could you just walk us through
34:39how that actually works in practice? And how far can that approach actually be taken when we're
34:45talking about not just resilience, but also the rejuvenation and restoration of nature?
34:49Sure. So it all begins with understanding what's on the land. So to predict the risk,
34:56we need to know what's on the land, what condition it is in. And traditionally, that has always been done
35:02by sending ecologists on the ground. So they would do the ground survey. If it is a large site or
35:08a
35:08large area, they would sample select and go to just some areas and then extrapolate it to the rest of
35:15the land. So the key of what we do and where we begin work for every user is by understanding
35:22what is on
35:23the land without going a sample-based approach. It has to be a full population. So if you see some
35:30of
35:31the other solutions that look at measuring biodiversity, people usually by satellite are
35:35today measuring at 30 meter resolution. What that means is one pixel on the output being produced is
35:42a 30 meter by 30 meter square. That makes it a 900 meter squared square. When we do the when
35:50we do
35:50measurements and these are all automated, we're looking at data which is 10 to 15 centimeter in
35:54resolution. And that makes the output roughly 40 to 90 thousand times higher resolution than what is
36:02done traditionally using tech. That also exceeds because that then does not get tired. The machine
36:07just runs. That also becomes far more better coverage than a human can do over thousands and
36:14thousands of hectares. So it scales really well. So once we understand very, have a very high quality
36:21view of what is on the land, then we can understand the impact of any change being done. Whether it
36:28is
36:28a building being put on or a habitat being changed to something else. We can go and repeat the same
36:34level of measurement year on year. And that's where the whole predict and then preserve comes in.
36:40It is the same land. So if it's a data center being built, that's probably a half a billion dollars
36:47being put into land. Apart from the cost of the land, that's a massive capital expenditure. So
36:51anybody who's building that, they're putting in a lot of capital that they're going to then
36:56get a return out of over a very long tail of time. So the resilience becomes really important there.
37:02So as we are helping our users look at the land around, help them bring the nature back to what
37:08it was
37:08before they started the development. On the same land that is now also exposed to any climate changes,
37:15fire, flood, storms, they can make changes as they're making nature improvements. They can plan
37:21them in such a way that the land itself becomes more resilient. And there the year-on-year measurements
37:27then can be done to find out what's the risk today and then to prove in a very data-driven
37:33and
37:34evidence-based approach how that risk is consistently going down. And that's where the long-term resilience comes in.
37:43Thanks so much for that, Shashin. That makes a lot of sense to me. And I've noticed a few times
37:48again and again we're starting to get to talking a bit about data centers. So it works really well
37:53into my next question, which is for you, Caroline. So AI frugality, you mentioned that shortly before.
38:00We're going to dig in a little bit more. To me it's really interesting just because of the way that
38:04we use it
38:05where I work as well. A lot of choices go into that. But it's very core to your philosophy and
38:10I'm very interested
38:11in how climate intelligence plays into this, how they sort of build each other. And so that leads to my
38:16question of
38:16how do you actually ensure that you're maximizing impact of just sort of the efficacy of your solution
38:23within this frugality framework? And this especially for those in the audience that are concerned with
38:29just the energy appetite and the ecological impact of data centers, because this seems to be very, very core.
38:35Yeah, I think a very good question. And as I said, frugality genuinely is one of our core business values.
38:44And I think there's a number of ways in which we do that or which anyone can do it. The
38:50first one is
38:50obviously looking at your tech stack and the way in which now we have a, we are, we're effectively
38:55serving up what we call discoverable, understandable and trustworthy data products. And we sell them
39:02sort of in market intelligence, which goes to hedge funders, traders, all that, and banks and that type of
39:07thing. We also sell risk intelligence, which is all about, you know, the how resilient is your
39:13first line commodity, which typically sells, sells to both banks and also supply chain organizations.
39:19And then we sell environmental intelligence that sells again across to supply chain organizations,
39:25but also to carbon project developers and so on. And so therefore underlying that is a very complex
39:32data fabric with raft amount of data, but I'll give you a sort of a fun fact. Um, we pull
39:39in sort of,
39:40as I said, some of our, um, legacies from JP Morgan and during the people, the architects around their
39:47daily reconciliation, which pours in $27 trillion a day. So processes $27 trillion a day. Our data fabric, um,
39:57processes four times amount of that, uh, information. So we're talking about a huge complex thing.
40:05However, our, um, costs, uh, our AWS or MS Azure costs are actually one hundredth of that. And it all
40:15comes down to the way in which we've architected the platform. And we were exceptionally mindful
40:21given that Trufera is, you know, the idea of the, a commercial tech organization, but obviously with
40:27sustainability at its core, going back to the point, it would be ridiculous if we input more carbon than
40:33we removed, um, with our data. So that's the core thing. I think the second thing I would then come
40:41on to, which is probably a newer thing that's happened, which I'm sure you're all, everyone,
40:44all of us are seeing this in this room is then the use of how we personally consume AI, you
40:51know,
40:56and people, obviously I, I say more now than ever, I think good people are the core and central to
41:04anyone's business. Um, because things are moving. I mean, I came from a technology company that was
41:10seen in 2021 as being the fastest growing, most progressive organization in tech there was.
41:16And I now look at it and think, God, we were quite old fashioned and it's five years. I mean,
41:22five years, I suppose it's a long time in these days. But, um, the thing that I've noticed is one,
41:28how our people use it. Obviously people are our core of our business because, you know,
41:32as much as we like to say AI is game changing, it definitely is, but it doesn't do all the
41:36things
41:37we need it to do. And one of that is making these conscious decisions. So we also give our game,
41:43our teams on, um, making sure that how they use in their tokens. And actually it's not done from
41:49a financial point of view. It's generally done from we, and what notes, what bought it on is we
41:55noticed a couple of our mega AI experts that are on it all day long. We're actually using like a
42:0220th of what a couple of people in our go to market organization we're using or our product.
42:06And so now we've given them the job of gaming them to say, okay, use X number of tokens. And
42:11it's
42:11almost a point of shame if you get ahead of it. Um, and we found that quite fun, but also
42:17very,
42:17very important because again, it is so important that we use it very purposefully. And there's so
42:23many good ways in which, you know, in hacks that you can learn, if you will speak, if we will
42:27speak
42:27to one another about how to use it safely without being ridiculous. So I think now it has both a
42:32technical, but also a human impact around the frugality points. Beautiful. Thank you, Caroline.
42:39We're, we're going to do our best and get through this, uh, Sirika. When Capgemini is consulting with
42:45its clients on how to bring AI to bear on these sustainability issues, where are you finding the
42:51biggest friction points to sort of successful adaptation and deployment?
42:57So, you know, in our conversations with our clients, you know, and of course we, every client is
43:02different, you know, the realities are different. So the challenges often are different, but, uh,
43:07there are a few common areas that we see, you know, as barriers or limitation or friction, uh,
43:13in our conversations with them. The first, uh, is really the way they look at sustainability. Uh, here,
43:19you know, uh, till date, sometimes organizations look, they look at sustainability more as a compliance
43:25or a cost and not yet evolved to what you said, you know, as a driver for, uh, value creation.
43:32Uh,
43:32and whenever you look at things for a cost, uh, or just compliance, uh, the less is better. So,
43:38you know, we need to, we, our conversations with clients are really how to show them that, you know,
43:43investments in sustainability are really about creating long-term value and it's right for the, uh,
43:48planet as well. The second friction we often face, um, is really around data. And I, you know,
43:54McKinsey had thrown this very important study last year where they showed that, you know,
43:58more than 80% of the organizations are embarking on AI, but, you know, one less than one third can
44:04take it to scale. The real issue is on the data. And when we look at environmental data,
44:09of course, it's even more fragmented. It's in silos. It doesn't, you know, and many of the data on
44:15environmental sustainability are, uh, third party estimates. So if you look at that, you know,
44:20the quality of data becomes a huge challenge when we are looking at finding long-term solutions,
44:25uh, for environmental sustainability. Last but not the least, like, you know,
44:29it was mentioned a little while ago here, uh, there is an evolving nature of the way the measurements are
44:34done. So, you know, there's too much of ambiguity and, you know, people are learning. So I think these
44:39are the three main factors that we find as frictions when we are engaging with our clients.
44:43Thank you, Sarika. I, that, that was very, very efficient. I appreciate you. Uh, Ash,
44:49to close things out, I do think this is something that people would be quite interested in. So I,
44:53I'm glad we could get to it as an investor at the intersection of AI and sustainability.
44:59What does a young company actually need to demonstrate to do to demonstrate that their
45:02impact is real, but also that their business model is sort of sustainable and scalable,
45:06uh, durable, I should say. And where is the sticking point that you're finding the most often
45:13as innovators and startups try to move from ambition to scale? I promise you I'm going to be
45:19concise. Um, so any good business needs to be self-sustaining. So what does that mean? First of all,
45:26whatever the company is building, it needs to have, let's say it's a technology company, it needs to have
45:30technology which is unique because that will create the moat it needs to survive and, and sustain
45:36against this competition. B, if the technology is unique, you need to have unit economics.
45:41We can talk about sustainability all day long, but you will not pay a green premium for a company
45:46because that's better for the planet. So unit economics become a very, very important factor
45:51in deciding whether a company will survive in the long term or not. And so for us, when we are
45:55doing
45:55our work, this is a very big, um, aspect that we look into of unit economics because that's what will
46:01help it compete and phase out the non-sustainable technologies over time. The third thing
46:05is how does this moat actually compound over time? You know, how does adoption actually create?
46:11So what we call go-to-market, right? And we don't understand how will this fit into certain ecosystems
46:15because some ecosystems probably adopting new ways of doing things. Some already got very established
46:20ways of doing things. How do you break into those markets and how do you allow a young company to
46:23kind
46:24of grow into that and create scale which will eventually become global by nature? Um, requires marketing,
46:29requires investment, requires, uh, thinking about adoption. And so another area where we look into massively. So
46:35net-net, if you're building a company which has to be sustainable by nature, it has to be a good
46:41technology, first of all. It has to be what you can accomplish. And then the three is you go to
46:45market.
46:46How do you kind of build this at scale? And how do you protect yourself against the ever-changing
46:50landscape of, because with AI, the landscape of technology will also change dramatically every few
46:57years. So you've got to be ahead of the curve. It might be a little cliche to say that, but
47:00it couldn't
47:01be truer than now. Um, and at the same time, you've got to be able to measure. You have to
47:06be able to
47:07measure what are you actually saving on. If it's a water technology, how many gallons or what are you
47:12saving as a result of that? And demonstrate, be able to show that. Because when you do that, that's when
47:17people actually believe in your purpose and will also support your purpose. And that's what investors
47:21want to look for. That's beautiful. Thank you so much. And thank you, everybody, for being with us
47:28for all this and for your patience. I hope you really enjoyed it. And please give these fabulous
47:32minds a round of applause. Thank you. Thank you.
Comments