Please come up, let's have some fun. Hi, thank you. Excellent. Thomas is at the board of SAP, Germany's most valuable company, the head of customer services, which is really at the heart of SAP. He studied business IT, I think, and has an MBA, and he is only 39 years old, which makes him, I believe, the youngest C-level person in the DAX companies since five years. Since five? Oh my God. Okay, so you had quite a way to go. You're a youngster in this group. And does it take a youngster to drive digital change?
No, by no means. I think but diversity is an asset in all aspects, and we talked about diversity earlier, and also age diversity is important. And for sure, in the end of the day, if you think about digital transformation, it's about the mindset. And I mean, it's actually fascinating if you think about also, I mean, quite frankly, my parents, how they use digital technology is fascinating in this sense. So it's not about age or young people who drive the change. We need conviction, we need assertiveness, as also it was mentioned, to really drive the necessary change. And I think one of the things that is underrated is that if you drive meaningful change, you will take unpopular decisions, you will see critics, you will go through a tough time, and you have the perseverance with the conviction to drive such a change. This is critical, and this is something where I hope that also SAP is a good example about driving change in organizations forward.
Would you say that you have been or are a nerd?
So then why did you start working with computers at this early age and all this fascination?
Actually, if you think about it, it was basically back then, so history lesson, it was a 386 back then, so we had four megabytes of memory, which was not that much, and the game which I played actually needed entirely four megabyte main memory, which means in order to play it, you need to deactivate the drivers, for instance, for the mouse, otherwise you cannot play the game. That was the challenge, and that was my first coding challenge: okay, how do I develop a startup disk which is deactivating every driver I don't need, a full 4 megabyte memory? I was fascinated by that because I learned what can you actually do with that technology. And also back then, quite frankly, if you think about when the first rocket went to the Moon with so many, with so little computing power, what the humanity was capable to do, and now think about the computing power you have in your hands. All of you have a smartphone. This smartphone has factor 10, factor 100 more computing power than flying to the Moon. Now the question is, what do we as humanity do with this power? We talk about AI, we talk about the next revolution, and still we talk about some very rudimentary problems where all of you in your business life will still say, 'Why do I need to do that boring process?' Actually, in this time, man can fly to the Moon, but we still need to do boring expense management.
Why don't you... for that, let's divert there for one second. AI needs more and more electric capacity, needs more and more electricity. Is AI going to keep us from keeping our climate goals?
No, I mean, for sure, it's fair to say the energy consumption for AI is going up as we speak, and it will be a very major driver, absolutely true. The question is what we do with that technology in order to reduce actually our carbon footprint. And if you think about the pure optimization potential in the supply chain, optimizing the supply chains is a basic mathematical problem. I mean, if you talk about Siemens, they have products which have thousands, hundreds of thousands of entries in the manufacturing BOM, and now they source it from different countries, different regions. If we can optimize that with AI, we actually will reduce the carbon footprint. So fundamentally, the net effect of AI hopefully should be positive towards the energy consumption, and that's certainly something where we need to work on.
Thomas, what do your customers want from you these days? Do they want just AI in some sense? Do they have specific goals? What do they want?
Actually, our customers don't want an AI for the sake of AI, but they're looking for productivity gains, they're looking for enabling new business models, they're looking for how can we help drive business transformations at scale. And here, for sure, we use technology for the greater good to optimize it. That's also the reason why we basically, as a company, said we want to embed AI as part of our products naturally. So it's nothing special, it's nothing at the side, it's at the heart, at the core of our products in itself, and with that, we help drive customer needs. That's what we focus on.
SAP has tried to deliver AI in your project for a long time, also with the HANA development platform. I believe already HANA was the first in-memory database at scale in the enterprise, which we delivered in 2010, and then for sure it already embedded some AI capabilities back then, but for sure back then different technologies, and that for sure had an evolution. Last year proved to be difficult, and then today SAP offers an AI called Joule, I think, but only for cloud clients, right?
Absolutely, yeah. So to give you the context, I mean, what we have, we have already more than 130 classical machine learning services as part of the portfolio, but last year alone we added 100 generative AI cases as part of the product, including Joule, our digital co-pilot, which also serves fundamentally as the multi-agent system in the agentic AI world what we're living in. And with that, for sure, we help our clients around the globe to release, thrive, and that's actually a very important notion because in the end of the day, AI is only as good as the data. Let's be real about that. And if you think about all the global large language models, at the end, they all base on the same data, the publicly available internet data. So by definition, no large language model differentiates by the content of the data, and that makes it now even more important to use proprietary data. And where's the proprietary data of companies? In SAP systems. And that's exactly what we exploit in our AI capabilities to make the most relevant, reliable, and responsible AI that you really can trust to make business-critical decisions, because that's the business we are in.
But do your clients let you use their data?
Yes, more than 30,000 of our customers gave us the consent in an anonymized way and opt-in way to use the data to train also our own found model itself.
But you have 300 million users, right? Or 300 million cloud users and customers?
And only a fraction of them allows you to use that data?
No, I mean, all the customers in the last years who started the cloud journey, that's why we started with the AI and the consent, and that's the largest portion of our customers who actively actually give us the approval.
Do you have to consent to that in order to be part of it?
Okay, European law sounds like extortion to me. There are much younger companies coming out of Silicon Valley and other places that offer AI-based data services to companies, also very broad-based, trying to put the whole company at a level where they can actually work with their data in a meaningful sense. Companies like Palantir and so on, are they a threat to SAP?
No, I think it's not. You know, we always worried about our only major company in that field, so I think you don't need to worry about SAP. I think you see it with the traction that we're on a good path, and we're not worried about the Palantir and the others. And by the way, we love competition. Competition gives everybody the motivation to do better.
