Episode 126: Jonas Andrulis on Investment and the Pressure to Deliver

Reza: Hi everyone, and welcome to Project Mindset, the channel for everyone who knows that with the right mindset, anything is possible. Today I have a slightly different mindset episode for you: a fascinating start-up story about a founder called Jonas Andrulis. He founded a company called Aleph Alpha. By coincidence, he also comes from my favourite city, Heidelberg. What a wonderful coincidence! His company is moving at a pace I find extraordinary, and I want to explore his mindset in this episode.

Before we do that, I want to try a little thought experiment so you understand why I find it so relevant to understand what makes him tick. Imagine you found a start-up in a technology field occupied by the biggest players: Microsoft with OpenAI, Google, Elon Musk and many other prominent names.

You can probably tell we are talking about artificial intelligence. You build technology so important to your own country that politics and business place huge hopes on you. Your customers include public administration, the Bundeswehr, large corporations and so on.

A financing announcement on an extraordinary scale

To develop faster, you decide to raise investment capital. Several major companies come together and invest €500 million in your start-up. First of all, that is an extraordinary number: €500 million. We can all do maths; we have studied or completed vocational training. But I still find 500 million a number I cannot really picture.

What I have described in this thought experiment sounds extraordinary, but it is Jonas Andrulis’s reality. He is turning the AI world upside down. I am not simply saying that; you can research it online. His technology has attracted so much confidence that many people in politics and major corporations hope it will allow Germany to compete at a world-class level in future.

I had the opportunity to discuss precisely this with him at my recent Dinner im Schloss event on 23 April. I did not want to focus on the technology — how he built it, what matters technically and so on. There are plenty of technical deep dives online. That day, during the panel discussion, I wanted to understand the person carrying so much weight and so many expectations.

Why Reza asks about mindset

I can recommend looking for documentaries online. Search YouTube for Jonas Andrulis and Aleph Alpha, and you will find excellent documentaries explaining his technology and related subjects.

But I wanted to understand how he handles the pressure. He makes striking statements. In another interview, almost as an aside, he said:

“The AI revolution is the last industrial revolution created by humans.” That sounds extraordinary to me, but somehow I think he may have a point.

I wanted to know what keeps him awake at night, what he struggles with, his vision for the future and the next steps he wants to take. It was mind-blowing for me to have this conversation. I think we will hear a great deal more from him. Today, though, we have some fascinating insights into his mindset here on Project Mindset.

I was delighted to speak with him. Here is the excerpt from the panel discussion. Enjoy.

Opening the panel discussion

Reza: A few months ago, I heard a sentence from Jonas Andrulis, the founder of Aleph Alpha, and I was not quite sure how to interpret it. I read that he said the AI revolution is the last revolution created by humans. I thought: is this a mic-drop moment? Does it describe the reality I am living in? How likely is it to be true? What is clear to me is that AI is here to stay. We can consider ourselves fortunate to have companies and start-ups here working so intensively on it and participating at a global level.

One question I did not ask Ms Hotz was what happens when a €500 million financing round suddenly lands in an account: how the bank reacts and which warning lights come on. I did not want to create compliance problems for her. So I am curious how Jonas reacted when he raised the €500 million.

I would like to invite the panel speakers forward: Jonas Andrulis from Aleph Alpha, and Michael Kleinemeier, whose presence is a great honour. Unfortunately, Adrian Thomas from Gründermotor is ill and had to cancel. We also have Emily Ryan from Google. A warm round of applause, please.

The perspectives represented on the panel

Reza: For me, this is an ideal panel. We have someone fighting relentlessly, someone who spent a long time on SAP’s executive board and knows whether a technology has a chance or is simply the latest hype, and someone with the broader perspective — and the resources behind it — to understand what AI can make possible.

Including in the future. Thank you all for your time.

Jonas, my question was serious: what do you actually do with €500 million? You have just completed a financing round. Many major players are involved — Bosch, the Schwarz Foundation, as far as I know. Perhaps you can say a little about it. Hewlett Packard Enterprise too: some of the world’s best companies.

Jonas: Indeed. My background is as a nerd. I spent my life in dimly lit rooms in front of flickering screens. If someone had told me a few years ago that I would one day have five-figure interest income per day, I would have doubted it too. That was what made this round special. We also received excellent support. In one part of the transaction, Sparkasse Heidelberg effectively beat Morgan Stanley. As a local patriot, I was very pleased.

Growth and high expectations

Jonas: Of course, it is a challenge. We skipped two size categories in financing, and now we have to do that in our growth. Expectations are rightly high. We speak with the boards of the world’s largest companies and at political levels that can hardly go any higher. That still stretches us as an organisation. Everywhere you look, you see stretch marks, and sometimes things go wrong. The difficult part is that we cannot afford to slow down.

We started in 2019, when AI was a very different subject. Now it is the top priority of the world’s best companies, and they all want to dominate. If you ask Microsoft or Google, their goal is for what we are trying to achieve — technological sovereignty — not to succeed.

Reza: I wondered about that. Money alone is not enough; you need the right people. My first question was how you put that power on the road. I do not think it is simply about hiring huge numbers of people, but about hiring exactly the right people.

Jonas: Exactly. We are not a services business, so our aim is not to acquire lots of people quickly. Our model only scales through headcount to a limited extent. Our reason for existing is innovation, and you cannot simply scale innovation through headcount.

People and leadership in a growing team

Jonas: That is why we bring in people from all over the world. We have German citizens who had left Germany, as well as people from India and many other places, whom we have brought to Heidelberg. We need exceptional people, and we offer an exceptional environment.

My last company still had fewer than 30 people when I handed it over. We are now at about 100. I have never led a company this large before. Within Apple, of course, the environment was different. It is a major challenge for all of us.

Over the past 12 months, however, we have brought in some extremely experienced leaders whom I can learn a lot from, including people from Google and Palantir. Some of that was reported in the press. We need them because we are on an unusual mission.

Reza: On regulation, Jonas, I read something from you — or saw it online.

I cannot remember exactly, but you said something interesting: if we had immediately imposed regulation when personal computers were being developed, that would have seemed strange. We would not have wanted that. Today, regulation is being built around a new technology that could be as transformative as the personal computer. What is your view?

Regulation and the speed of innovation

Jonas: If we invented the car today, we would ban it first. Of course, individual rules make sense. Would it make sense to regulate how securely people can stand on wooden planks? Certainly — they can be dangerous.

First, we need an authority to check that. Then we need sanctions in case a plank has been installed with different stability or load-bearing characteristics.

Every individual point can be justified. You can justify requiring toilet paper used on subsidised film sets to contain 90% recycled material, then checking and enforcing that. Each individual case can seem right. In aggregate, though, as Michi said, we slow ourselves down. We end up very busy being compliant.

We participated in the discussions and always said regulation is good, including regulation now.

What hurt me a little was that so much of our creative energy and resources were directed towards compliance. Germany and Europe do not exactly have an excess of AI leadership. The creative energy we have was channelled into compliance.

Creative energy for innovation

Jonas: It comes up with all our customers. I am also active in Bitkom and on its AI board, where AI compliance is always a topic. If we had used that energy for innovation instead — thinking about how we and our companies could build the next generation of empires or digitalise public administration — I believe we would have been better served overall.

Reza: There is an interesting tension. I understand that too much regulation is not good. On the other hand, you work with institutions that face extensive regulation or data-protection requirements.

Jonas: But they already do that. If I work with Bosch or SAP, nobody needs to tell them to consider and minimise risks. They have reputational risks. Their brands stand for good technology and responsibility.

Even without regulation, they think about this. In my experience — and we can discuss it — you sometimes have a 600-page regulatory document that nobody reads afterwards. Separately, people think about protecting customers and taking responsibility for their technology and products. If those happen to cover similar issues, we have been lucky. I do not think we need to force our companies to do the right thing and take responsibility.

Experiences with public administration

Reza: You are also very involved with public administration, which looks at things differently again.

Jonas: Exactly. I have learned two things from my many conversations in Berlin, Brussels, Stuttgart and elsewhere. First, I was genuinely impressed by some of the people I spoke with and their ability to recognise that we are setting the strategic direction for decades to come. Everyone understood that. I was at the research summit yesterday, where people broadly agreed that decisive steps are needed now.

But then I have so often heard: “We would like to help you, but unfortunately there is nothing we can do. It is all so difficult. We have so many great ideas, but unfortunately…” We have also received commitments that were not kept.

I do not think this is because people do not want to act. They want to shape the future in the interests of the country and our values. But we have made it very hard for the excellent people in those institutions too.

Time pressure in AI competition

Reza: Time is a useful keyword. How much time do you have, Jonas? I heard you say in an interview that you worry about this being a race against time.

Jonas: Absolutely. As was described just now, this is not a finished product. We do not have a completed product with which we can comfortably expand. The technology is more like intelligence infrastructure.

That is why cloud providers are so interested: they already have computing infrastructure. Combining that with intelligence infrastructure and data makes economic and strategic sense.

While this whole market is developing, we have the opportunity to position ourselves so that further scaling becomes possible. We need to reach a position that allows us to survive and keep our heads above water. It is not only large cloud and infrastructure providers; all the platforms are looking to use AI to enable new processes of value creation.

Reza: Where is that point of keeping your head above water? In terms of time, or perhaps organisation and technology?

Jonas: Exactly. I can share a little behind-the-scenes detail. Do we have press here? Yes, two? No? Okay — please do not immediately print everything tomorrow.

Revenue, costs and growth

Jonas: The revenue we already have on the books for this year puts us in a significantly strong position by comparison. One reason is that we are among the few players in this area actually achieving positive contribution margins. Operating the technology is so expensive that running something like a consumer chatbot at break-even, or with a positive contribution margin, is difficult.

We are already well positioned there, and I am very pleased with our customer engagements.

The problem, of course, is that we cannot remain at our current level. We need to grow faster than would ordinarily be possible.

One question I regularly ask myself is how we move beyond the standard path. The standard path is hiring a few more people, adding a few more customers and asking each customer about their problem. One says supply chain, another customer support, another an HR chatbot.

That is too slow. How can we reach a stable position quickly enough to continue our mission of technological sovereignty? Let us not fool ourselves: I am very pleased with our partners and financing, but technological sovereignty remains a mission against the odds.

What technological sovereignty means

Reza: Could you explain technological sovereignty briefly? I know independence from other market players matters to German and European companies, but could you expand on that?

Jonas: Complete independence is not possible. We have operating systems, hardware and so on.

For me and our customers, technological sovereignty means being able to take responsibility for the future. The future is changing. Value creation will look very different. The world in which future generations grow up will be built by AI.

Can we take responsibility for that future? Can we use the value created by the most powerful industrial revolution ever for the cultural and social challenges this transformation brings? Or will it all end up with shareholders of other corporations? Perhaps nobody will even ask us which ideas may be considered. How can we support Schwarz Group, Bosch or SAP so that they can take responsibility for their own futures and build their own empires?

That is the business side. In public administration, the issue is even more pressing. There are currently 400,000 vacancies. Anyone who looks at the demographics can see that we are heading off a cliff over the next ten years. If we cannot maintain the ability to act there, even the best companies will not help us.

Reza: May I ask again, Jonas? We could go very deeply into the technology, and you are certainly the right person for that. Perhaps those conversations will happen here too. But what interests me more is your mindset, now that I have the chance to speak with you.

The personal side of being a founder

Reza: What helps you sleep well, and what keeps you awake, when you think about Aleph Alpha’s future and AI in Germany? It is not an insignificant question. A great deal of hope rests on you and Aleph Alpha. There is a risk you will not succeed, and then what? Other countries have companies too, such as Mistral in France, with great hopes attached to them. We also have Paretos here; we heard from Thorsten Heilig today. But what keeps you awake, and what helps you sleep?

Jonas: If it does not work, we will do exactly what we would have done without Aleph Alpha. When I discussed my founding idea with smart people, everyone said: “Do not even try. It cannot be done.” Technological sovereignty from Germany, with German investors — even Mistral did not manage purely European financing. It is difficult, genuinely not easy.

Important tasks and limited time

Jonas: It would be a shame if it did not work, because it is not obvious to me where Germany’s future winners and value creation will come from. We have many excellent established companies, but they need to make it into a new era. I think we all want them to succeed, and that outcome is not obvious.

What makes me sleep badly is that whatever I do at the moment, I neglect things that are incredibly important, including to me personally. There simply is not enough capacity.

Reza: Work on the actual substance?

Jonas: Everything, really. In the first year, I wrote the first pieces of our explainability code myself. Many researchers came because they wanted to work in that environment. I am often told they miss that collaboration with me and would like me to engage more deeply in the technical substance.

But I also have a leadership team, and they would all like as much of my time as possible. I have a responsibility to support them too. These are all important things. Being here matters to me, as did the trade fair and the research summit, which is incredibly important for our location.

In the end, I try to find the best allocation of my time while still meeting my responsibilities.

How dependent is the company on its founders?

Reza: Would everything work without you? Sorry for the question.

Jonas: We often tend to overestimate how indispensable we are. But I think removing either me or my co-founder Samuel would cause pain.

If it had to happen, there would certainly be ways to manage it. You should never consider yourself too important. But with about 100 people, perhaps unsurprisingly, several people — not only Samuel and me — are very important to what we do. Losing them would hit us hard.

Reza: It is sometimes remarkable how much big things depend on small details. That repeatedly blows my mind.

Jonas: That is also what makes it interesting. My team at Apple had a thousand people. You may have read in the newspaper recently that ten billion went into Apple’s self-driving car, and in the end the project was discontinued.

That happens. Those are the big pots — and using a large pot does not necessarily mean the food will taste good.

You need the right people, and above all the right decisions and positioning. If I am badly positioned, adding resources will not fix the position.

Positioning and viable business models

Jonas: This is one problem many generative-AI start-ups are seeing now. They positioned themselves in ways that cannot work, saying they would think about the business model later. That is the Silicon Valley approach: attack at full speed, then work out the business model. It can work, of course, but some people are sweating quite heavily at the moment.

Reza: I think you paid attention early to having the right business model. That also feeds into scalability.

Jonas: Exactly. I feel very comfortable with it right now, but take that with a grain of salt. I am the CEO and I built it. We will see over the next few years whether we positioned ourselves well.

Reza: Perhaps one final question: what are your next three major steps over the coming years?

Jonas: We will probably have a major announcement and launch in June this year, showing how to make real money with the technology. Thanks to Mark Zuckerberg — who certainly is not doing this purely out of love for humanity — the opportunity to capture value from foundation models alone is currently limited. Nobody is really making money with that alone.

Commercial goals and collaboration

Jonas: We also integrate open-source models and add our own innovations. That is great, but it is not enough on its own to earn 100 million. And we need to earn 100 million.

So we have some commercial activities planned.

Reza: By when do you need to earn that 100 million?

Jonas: We do not need to earn all of it this year. But by the end of the year, it needs to be clear that we are on the path towards it. That must be visible. We cannot still be limping along at year-end saying we made 20,000 here and there. That would be bad.

Another subject very important to me personally is our people. During these growing pains, some structures have developed that do not work particularly well. Some people do not feel particularly good. I have brought a phenomenal leadership team on board, but they may feel a little lost because I have so little time and we are all overloaded with work.

Stabilising the team and enabling people to deliver what they are capable of is another important part of my job.

Reza: We have talked a lot about technology, but ultimately it is still a people’s game.

Jonas: I could get a little nerdy about what we are doing technologically. Who has done their linear algebra homework?

Closing the conversation

Reza: First of all, thank you very much for these insights. Looking at the time, I think some of you will still be here for a while, although I believe you need to leave soon. Anyone with questions should go directly to Jonas. Otherwise, thank you all for your time. That was excellent — thank you.