Rise of AI Conference 2025 · Berlin

State of AI 2025

Fabian Westerheide · CEO of Rise of AI and AI.Fund

At Rise of AI 2025 in Berlin, in his opening keynote "State of AI 2025," Fabian Westerheide, CEO of Rise of AI and AI.Fund, surveys the global AI landscape under the year's theme "Good News from Europe." He argues China has proven itself the clear technological number two, pointing to DeepSeek and Manus as evidence it can take American innovation and make it leaner and more cost-efficient, while the United States remains number one and is now largely unleashed from regulatory constraints as government and tech industry interests align. He notes new investment hubs emerging in Saudi Arabia and the UAE, including a deal in which Saudi Arabia is pushed to invest a reported one trillion dollars in exchange for one hundred billion back. He argues Europe cannot compete with the US and China on large language models given its gaps in talent, resources, capital and infrastructure, and must instead become smaller, slower and more specialized, building on many small success stories rather than one big win. He then walks the audience through the day's three stages, Policy & Society, Applied AI and a new offline "Cognitive Labs" format, and their speakers.

Key takeaways

  • Westerheide says China proved itself the clear AI number two via DeepSeek and Manus by taking American innovation and making it leaner and cheaper.
  • He describes the US as increasingly deregulated in AI, with government and tech interests aligned, which he says speeds iteration but strains earlier trust norms.
  • He cites a deal pushing Saudi Arabia to invest a reported one trillion dollars in AI for a hundred billion back, alongside UAE interest shown in Paris.
  • He argues Europe has lost the race for the largest language models and must instead be smaller, slower and more specialized to compete.
  • This year's livestream, run by German company Meetyoo, rebuilt the conference environment so virtual attendees feel present in the same room.

Detailed notes

State of AI: China, the US and the rest of the world

Westerheide says China proved itself the clear number two globally over the past months through DeepSeek and Manus, arguing China is quick to take American innovation and make it leaner and more cost-efficient, and is pushing governments and companies toward common AI standards. He says the United States remains number one and will stay there, now largely deregulated as government and technology interests align, which he says speeds iteration even as it strains earlier norms around trust and oversight; he adds that despite this, American companies operating in Germany, often through local teams, continue business as usual. He also points to Saudi Arabia and the UAE as newly emerging funding hubs, citing a deal in which Saudi Arabia is reportedly pushed to invest a trillion dollars in AI in exchange for a hundred billion back, and UAE investment interest shown at the Paris AI summit.

Europe's "good news" strategy

Westerheide says this year's conference theme, "Good News from Europe," is meant to counter negativity, acknowledging he struggled to define it and that asking an AI tool for reasons produced only "random numbers." He argues Europe has lost the race to build the largest language models due to gaps in talent, resources, infrastructure and capital, so its path forward is to be smaller, slower and more specialized, finding niches and being smarter rather than bigger. He frames the many individual company and network success stories in the room, not one large announcement, as the real "good news."

Walking through the day's agenda

He previews the Policy, Society & Strategy stage, opening with a panel on positive developments in Europe, followed by speakers with international perspective and further sessions including Axel Voss of the European Parliament, one of the figures behind the EU AI Act, discussing criticism and improvements to it, and a newly appointed state secretary for digitalization who has long been part of the conference community. He notes the country's digital minister could not attend due to a change in position, calling the existence of a digital ministry itself "good news from Germany." He says that stage closes with a keynote from Hans Uszkoreit, a long-time participant in the community. He then describes the separate Applied AI Stage as focused on execution, scaling and infrastructure, with sessions on longevity, feminist AI and agent systems.

New formats and closing notes

Westerheide highlights a new offline "Cognitive Labs" format, a closed-door, screen-free space starting with a session on software-defined defense, framing AI as a driver for national defense capability through dual-use knowledge. He also mentions an NVIDIA and HPE masterclass on AI factories, and, for the first time, two live podcast recordings at the PIABO lounge. He notes the livestream this year is run by a German company, Meetyoo, which rebuilt the conference environment so online attendees feel present in the same room, and that virtual reality was not used because, in his view, VR and the metaverse never became mainstream. He closes by describing pre-produced postcards, some containing seeds, meant to encourage offline sharing and personal outreach, including sending some to Germany's chancellor.

Quotes

China proved that they are the clear number two in the world.

Fabian Westerheide

We can't compete with the large language models where we lost the game due to talent, resources and infrastructure and capital.

Fabian Westerheide

We will make it, we will continue, but smaller, slower and more specialized.

Fabian Westerheide

Questions answered in this talk

What does Fabian Westerheide say about China's position in AI in 2025?

According to Fabian Westerheide at Rise of AI 2025, China proved itself the clear number two in AI globally through DeepSeek and Manus, taking American innovation and making it leaner, cheaper and smarter, while pushing standards across its government and companies.

How does Fabian Westerheide describe the United States' AI regulation approach in 2025?

According to Fabian Westerheide at Rise of AI 2025, the US has largely abandoned its own AI regulation as government and technology interests align, which he says speeds up iteration even as it strains earlier norms around trust and oversight.

What strategy does Fabian Westerheide propose for Europe to compete in AI?

According to Fabian Westerheide at Rise of AI 2025, since Europe cannot match US and Chinese scale in large language models due to gaps in talent, capital and infrastructure, it should become smaller, slower and more specialized, focusing on niches and many small success stories.

Mentioned

DeepSeek, Manus, EU AI Act, AI sovereignty, Cognitive Labs

Keep exploring

More from Fabian Westerheide

More from Rise of AI Conference 2025

More on AGI & Future of Intelligence

Transcript

Auto-generated transcript, not edited. · 2,923 words

Show the full transcript

Good morning. Um I have to take this for a second. It's um once in a year together everyone in our ecosystem. It is lovely and great and to see all of you. It has been a busy year so far. I mean do you feel the acceleration out there, the disruption? Are you still able to catch up with everything in AI and politics? Um, today we meet because we are all part of the AI ecosystem. We are all part of this acceleration especially here for the progress in Europe. Um, before I start, I have to mention that we have no internet in Berlin. It's not ours. It's the United Nations next to it. And I assume it's jammers police.

I mean, you saw all the stuff outside there. That is Berlin. Yeah, it doesn't work, but everyone is still coming. I also um would like to welcome our people in the live streams um who are here on this stage and the other stage for the day. We have a German company this time running the stream uh called Meet You. And what I liked is that they rebuilt the whole conference. So the people online you are in the same room like we are. I wanted to do it in VR one day but honestly VR metaverse didn't reach us. Um it was one of the B hypes which didn't make it now. Um I tried each year I tried to talk about the state of AI and uh this time I have to make it a little bit shorter because AI got very very complex.

Um I would say it changes faster than most of us can get. I mean I I lost track and so many iterations, new models here, funding there. It is plenty to do. That's good why we here to catch up to see what is honestly the big news for us. Um so but my my short view on the state of AI globally. So first we have China and what did China in the last couple of months what we learned? We heard about deepseek and manos and it was interesting because all the the media was like what is deepseek what happened here and are we lost now and everyone else knew deepseek before it's like we knew CHP before too and open AI so if you're in the ecosystem it's no big surprises but the mainstream they were overtaken and what China China proved is that they are the clear number two in the world they are very quick in adapting they take American innovation make it leaner cost efficient, smarter.

And what I learned in the ecosystem was the the the the wording, why didn't it come from us? Why were we not able to go mainstream in the world with something like manos or deepseek because technical we could have done it? Sometimes the feeling we always want to do again and original and we don't want to copy and improve on it but the Chinese are doing it and if you talk with the Chinese ecosystem they are heavy in the integration part. they are forcing all the governments, all the companies to work on AI have standards. So, especially in the in the hidden part of AI, we will learn more about them. And then we have the United States still number one and it will stay number one for a while.

And I would say AI is now unleashed. I mean, we talked about EU Act and we need to, you know, maintain a balance for society. Americans don't give a on this anymore. You know this is when government and technology people are now the same. Yes. They get rid of their own regulation. And for the ecosystem it's good because we will see fast iterations and more because regulation is often a burden. Yeah. For for speed. And also what we see is that tech partnered with the right wing. Something which will not happen here. I'm pretty sure. But also it's a very weird mixture um which we're seeing there. Also what we learned about the state of the AI is that we thought this trust and relationship doesn't work anymore.

It's gone within a couple of months decades of building up trust in our systems was kind of dismantled on the government level. But what I also learned is talking with a lot of American companies, business continues. American companies are present here. I mean half of them are Americans and they have German teams here and they want to do so business continues and we sometimes have to abstract and don't blame the US for everything. It's just part of government and governments can change but the business will continue. So I say that we will see the USA continue to lead the field. They still have the most capital. they have the most talents and the most ambition.

And then with the rest of the world and what just in the last year we have seen mainly Saudi Arabia and United uh UAE coming up with a lot of funding. I think Riyad is parallel. Trump is over there is forcing Saudi Arabia to invest I think 1 trillion in exchange they get 100 billion back. It's a good deal for someone. Um um and we also saw it in the Paris summit that the UAE wants to invest heavily in France where again said why don't they do it here why is German government not able okay we didn't have one at the moment able to allocate these resources too but we see how hubs are outside emerging and sometimes um just as an orientation so we labeled last year this conference the good news from Europe because we wanted to break with the negativity or how do um I think Frankman quoted it on on LinkedIn it's time for this and I would say it's not an easy topic I struggled with it so I asked

perplexity perplexity came up with plenty of numbers but no good reasons it was just random numbers and um we've seen this Paris summit huge announcement and what happened after not a a lot. Um so sometimes I feel like this iteration cycle yeah someone is making announcement one would do and then it doesn't happen again and this is kind of frustrating. However the answers for the good news of yours will be given today because it's all these small stories which every one of us can share. We all are success stories with our companies with our networks whatever we do. So we share the stories. We are the good news. It's not the big news we have.

It's many many small news from Europe which are my personal view Europe is a place to live and if you talk with some Americans currently they don't like it they come over you know you always have a switch all four years between the oceans sometimes some people but I would say for the last decades Europe is the place to live and most of us want to continue to live here we are here because we want this to be a good Nice. So what I say is we will make it we will continue but smaller, slower and more specialized. We have to find our niches and sweet spot. We can't compete with the large language models where we lost the game due to talent, resources and infrastructure and capital.

But we have all everything else they are outside of the big big competition. Um we just have to be smarter than the others and we are. So this time I will speak a little bit about the agenda first changes because humans are humans. Yes. And they change a lot and also to guide you a little bit through the day. So we kicked off with the AI topic tables out there and I like them because they are more um like a preview of what are the main topics who wants to talk about something and when I asked all the guests what do you want to hear I got 90% answers agents agents agents and I struggled with this because it's another hypeboard another bus which is not real there I mean we can speak and technical about this but um people want to hear about something they look for the next on but it's sometimes a little bit early but not but we have agents had it already there we talk about sovereignity because this is

again something important for us for Europe to stand on our own feet to be able to defend self-reliant and to build a surve infrastructure we talk about humans and how do they adapt to this times networking jobs how do we adapt to the terms sustainability even we have other thoughts you know climate change is still there and we still have to think it and it can't be that AI is needing more energy and more energy and it's not sustainable at the end and then a lot of topics about transformation transformation of government of industries ourselves evergreen so on the agenda we have some small changes but I will come to this so I structured it today this stage is for everyone who's interested in policies society strategy um the first block will a panel with Peter Zalin, Jonas Andrulis and Lisa Holski which is well one of the most known people we have here in Europe.

So they will talk about the good news in Europe because they produce some um and then it will be followed by Pascal Fong who was for 20 years in Asia is now in here has an international view and is bringing something and then we have the next uh success story which is defin um then we have the second block and there something is missing in contrast to your printed agenda because the minister of digital who wanted to come had to leave his position last and the new guy is not ready. I just got an email this morning. I'm sorry, I can't do it. I'm overwhelmed with my schedule currently and I understand it. Yeah, he will come.

But that's by the way good news from Germany. We finally have a digital ministry. Then we have in the second block Rafael from talking about AGI and etc. And something we lobbyed for years too to have something darker similar. I mean how often say we need this? Well, we finally have it. They have budget and they're doing really, really good stuff. Then we will have Axel Fos from European Parliament and one of the guys be behind the EU Act. So, critics, complaints, improvements towards simply. And then we have Thomas Yachbeck who's talking about how we become a ca and he is the very fresh appointed secretary of secretary of state for digitalization. So he is now the highest ranked member of government who's supposed to know everything here and since he is part of this conference from the early beginning I'm very happy that we finally have someone in government who has been part of our community for a very long time.

Um and then we have a a panel with cast Marcos and Vanessa and Stephanie to to wrap up everything and then go for the third block. The third book is I wanted to have talking about the consequences for society. How do we do the storytelling positive? How do we frame AI as something good for society so more people benefited and it's not getting all these prejudices which we have there out uh unknowingly. And then so there we have Kenza, Aneka, Elizabeth and Nicolas. And then for the last keynote we have Hans Oscorite who is here for every year and who is one of the long longest um professors entrepreneurs which we have in the ecosystem.

So I'm always very happy that he's joining us. Well applied AI stage this is for everyone who doesn't want to get into politics and is not interested. This is for execution, scaling, infrastructure, more transformation. We have a topic about longevity. So Ramen is here and he's very happy if you join him later. Yeah. Um we have feminist AI as a topic. Yeah. By um we have innovation agent systems too. And at the end we close with art and AI and here we have a small update. Dr. Tina Clever contrast to your program will be now at 12:30 because at uh 4:30 she is meeting her new boss for yeah so that's good you know we have people here who are directly now there but that also means if the boss is calling you and it's the new boss you want to show up so we had to change the agenda a little bit that is with having a new government and shuffle so don't complain if she's not there like written program it's 12:30 now

um We have also a new format out there which is called cognitive labs. It's if you go out there on the corner it looks like a classroom. It has no screens. So it's an offline format and let's say you can talk there behind closed doors. Whatever you learn there you can keep but don't quote it in public. So the people on the other side can talk a little bit openly and be honest which honest which we like. Um we start there with a topic of softwaredefined defense uh with uh general major Amen and Yorkbart and this is a topic which is always sensitive to talk about defense and something we are facing. I mean we have a threat out there.

We have to ramp up our national defense capabilities and we as an ecosystem honestly AI is one of the main main drivers. So I asked them what can we serve the military? How can we provide them? How do they have to act so we can help them with dual uses knowledge etc because they need us then we will have during the lunch with Claudia Paulink she mentioned it iconic women in AI so all iconic women in AI I invited but cool men too I asked for this um and then we will have a master class with Nvidia and HPE about AI factory so a little bit more hands-on getting in the practical side and then finally we have Mario Brandenburg Borg who is a friend for years too.

He is the predecessor let's say from Tumas Chumbbe. So he gave up the power and he's today here as a private person but he has been part of the government for the last three and a half years and he has seen stuff. Yeah. So you can ask him for insights which I did yesterday for two hours and is very interesting to look behind the curtains what really happened in the government what his thoughts are what his learnings are. And then we have an another thing um I always wanted to have is live podcasting. There are some podcasts I like to enjoy. So for the first time we have two live podcasts. So if there's one of the podcasts you know or you don't know, you can join the audience and be part of a live recording which will be over at the PABA lounge.

First Elizabeth together with Anika Fumios talking about HR and AI and then we have Aneska and Alexander for the podcast and um will which will be in German I assume because I always produce in German. I have no clue what they would talk about. They don't know it either yet too but it will be wrap up of everything around which we have and oh yes Veronica reminded me to do this. We have pre produced postcards for you because the more we use AI the more we need humans and I baronic had an idea to make it very offline. So we have postcards two versions you will find out there and some of them are with seeds inside.

So AI needs roots and ethics empathy and earth or AI is the seed humanity's garden. So you can plant them and grow something out of it. And I thought like send them send them to friends or your mother. They're happy to receive a postcard. Send them to people who should be here next year and send them a handful of them to fat. I want that. He's receiving a plenty of postcards that he has to show up next year to show and to explain us how he wants to make Germany cool again out there. And because people are lazy and if you don't have it, you can ask me. I have stems with me. Yeah. So, grab them outside, hang them in the fridge, send them out, make this a little bit offline, and share it.

And um now I rushed it a little bit, but Daniel is prepared. Um I want that you enjoy the rest of the conference, meet a lot of people, catch up, go out, talk, connect. Um, stay with the progress, be progress, and let's produce a lot of good news from Europe today. Thank you for being here.

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.