SoftBank’s 75-billion-euro investment in France – Europe’s AI sovereignty is taking shape, but not through Europe

Source: Startup Insider – Original Episode (episode in German)

SoftBank is investing up to 75 billion euros in data centers in Hauts-de-France—45 billion euros hard committed over five years, the largest infrastructure project in European history. Europe’s AI sovereignty is emerging, but not through Europe itself. Björn Rieckhoff and Jan Thomas analyze the deal: why French nuclear capacity provides industrial electricity for 58 euros per megawatt-hour instead of 91, and how SoftBank is simultaneously betting on Stargate, ABB Robotics, ARM, and France—with roughly $130 billion in debt against a market capitalization of $270 billion. Also: Anthropic with $47 billion in annualized revenue in its fifth year, a comparison to SpaceX’s valuation, and Germany’s 96% dependence on foreign AI technology.

Key themes in this episode

Transcript

This transcript has been edited for readability. The content and statements have not been changed. The original conversation was in German; this is an English translation.

Jan Thomas: Björn, hi, nice to meet you!

Björn Rieckhoff: Hi Jan, likewise, it’s great to see you.

Jan Thomas: I hope you're doing well. You were on vacation—I hope you had a good rest.

Björn Rieckhoff: I’m feeling great. I got to enjoy the Italian sun; I was in northern Italy at Lake Garda. It was fantastic.

Jan Thomas: And we’ve already got our batteries charged, so to speak, for SuperVenture, SuperReturn.

Björn Rieckhoff: Right, right, here we go again. I’ve been biking a lot this past week and racked up some elevation gain. My legs are tired now, but my mind is refreshed. And next week it’ll be the other way around: my legs can rest, and my mind will be put to work.

Jan Thomas: I was thinking about you yesterday. I spoke with Enrico Ohnemüller from Bunch, and that’s when the term “European Dynamism” came up for the first time. I actually thought that was pretty cool. He presented it as Bunch’s mission—we have to somehow hold our ground here—and he referred to Andreessen Horowitz and this “American Dynamism” that we’ve been talking about. And I thought: Oh cool, this “Dynamism”—it’s going to take hold now, like a kind of conflict between tectonic plates.

Björn Rieckhoff: Yeah, I have to smile a little. On the one hand, I think it’s pretty cool to use words like that, especially “European Dynamism.” On the other hand, it’s obviously such strong language. But I’d like to avoid becoming a hardliner in that direction in the future. Maybe a few European equivalents to Peter Thiel and Co. would do us some good, in a way.

Jan Thomas: Yeah, although luckily—or should I say, interestingly—he’s now emigrated to Argentina. I found that really exciting too; we just had it as the top story in the newsletter. So there’s really a lot going on in the world right now. Hey, maybe just a quick note before we dive into today’s topic. That actually fits right in with the topic of European Dynamism. But I was really thinking of you—as I mentioned in our preliminary chat: If I’d had to bet on what we’d be talking about today, I would’ve said: We’re talking about the Oura IPO, which is slowly starting to take shape. And the Oura Ring Gen 5 just came out, too. So there’s a lot happening on that front as well. I’m really curious—do you have any thoughts on that?

Björn Rieckhoff: Actually, I’m still thinking about our conversation from, I don’t know, September or so. I’m still missing what you brought up back then: Where is this all heading? What’s actually the long-term play here? That’s still not 100% clear to me. It’s absolutely clear that it’s an incredibly good consumer company, but I haven’t quite grasped the full picture yet.

Jan Thomas: You’re the market leader, right?

Björn Rieckhoff: It’ll be interesting to see what the capital markets make of this. Of course, it’s still a different matter whether you can sell a certain story to private investors, or whether it’s actually viewed as promising on the trading floor—that is, across the broader market.

Jan Thomas: I think it was an 11-billion-euro valuation—I’m not buying that just yet. But at the same time, I saw that Google released a new wristband, the Google Fitbit Air. And there you can already see that this health-tracking sector is also super important to the big players. Maybe one of the big players will snap that up at some point, too. I wouldn’t be surprised.

Björn Rieckhoff: Hey, it always comes down to this—we often talk about proprietary datasets. Building that up is, of course, the gold of the future, and the companies that have access to it or create it are incredibly well-positioned, regardless of whether it’s in a B2B or B2C context. That’s why it makes total sense that the big players are getting heavily involved.

---

Jan Thomas: In that context: There’s a new startup now—I’ve mentioned it here on the podcast before—and I think it’s totally wild because you’re talking about data collection. There’s a startup in New York that will now clean your apartment for free. In return, though, the cleaners who do this for you for free have to wear a camera and record virtually every move they make, which is then used anonymously to train robots. I think that’s totally wild.

Björn Rieckhoff: It’s kind of that topic or analogy that people used to draw back in the day on social media: If it doesn’t cost anything, then you’re most likely the product yourself. And I view such trends with a bit of skepticism.

Jan Thomas: Well now you’ve got me there, Björn—if we could just touch on that briefly... That’s really fascinating, because I was thinking the same thing in that context. All of us are using ChatGPT or Claude right now, and they hardly cost anything. And think about why. I think that’s close to what you just said. I believe we all fail to realize that we’re actually the product there too, and that what we’re doing is actually making the AI smarter, so that it can eventually take over our jobs.

Björn Rieckhoff: Yeah, I mean, at the end of the day—you have to be careful there, too. Especially when it comes to Claude Cowork and the like: What are you granting access to and so on? It also helps to be aware of what’s actually being done with the data. Can it be used for training or not? There are different approaches to that. It always makes me feel pretty good, or reassures me in a way, when I see Anthropic’s intentions—they’re structured in such a way that they have “The Greater Good” in mind. It’s a bit like Google back in the day, which originally had the vision: “Do No Evil.” They eventually dropped that.

Jan Thomas: It just quietly disappeared.

Björn Rieckhoff: Exactly. In that regard, I hope that the major players will uphold this intention—especially from an ethical standpoint—for as long as possible. But you’re right. I think the same applies here: If you were to set ethics aside, a whole lot would probably be possible with the data shared by the average consumer.

Jan Thomas: There’s this great saying: “The Google search bar knows more about you than your best friend.” And that’s probably true. And I think with ChatGPT and Claude, it’s even more extreme—because there, you just upload… with Google, you can’t even upload a contract or your health data or anything like that. So one person uploads—essentially uploads their deepest emotional turmoil—while another uploads the contract for their house purchase and says, “Check this out.” And ultimately, you’re just training an AI with that the whole time. It’s actually kind of insidious in a way.

Björn Rieckhoff: You know what surprises me a little: Sometimes it’s not even that I’m consciously doing it on purpose, but just things like—there are those standard use cases from Claude Cowork or whatever: “Sort my screenshots,” “Clean up my Downloads folder.” Fun, right? It contains all your insurance policies and, for all you know, every bill you’ve ever received. So it’s simply a data pool that you have to handle very, very carefully.

Jan Thomas: Yeah, the data—but let’s take an example: You’re working as a consultant, and at some point you want to start with branding or have a marketing strategy. And just this dialogue you’re having is, in itself, training an AI. The AI shows you something, then you give feedback—and you know, that’s worth its weight in gold later on. It’s just that people don’t really realize that, in this case too, they’re actually the product. It’s just not obvious.

Björn Rieckhoff: That's right, absolutely.

---

Jan Thomas: Hey, while we’re on the subject, maybe just briefly: Anthropic, IPO plans. We’ve been discussing it here on and off over the last few days—how they had that massive latest funding round and have now surpassed OpenAI. We’ve also discussed SpaceX here, which wants to go public. Which would you rather buy—one SpaceX share or two Anthropic shares?

Björn Rieckhoff: Two Anthropic shares.

Jan Thomas: Yeah, well, obviously, right?

Björn Rieckhoff: Even just half. I don’t care—I would never buy SpaceX.

Jan Thomas: Yes, that's right.

Björn Rieckhoff: Yeah, yeah.

Jan Thomas: I find it interesting that SpaceX actually wants to be worth twice as much as Anthropic. And, interestingly, it’s worth just as much as the entire DAX.

Björn Rieckhoff: Oh, it’s so absurd how this person can just spin a narrative out of thin air. You could learn a thing or two from that—but also from the fact that there are people who just buy into it. That’s unimaginable to me. I’m probably too short-sighted when it comes to doubt. But when someone seems to change their entire business model or the company’s vision every ten minutes—I find it hard to see the big picture in the long term.

Jan Thomas: And you just said about Oura: Let’s see how the public markets react. What would you say about Anthropic if they went public?

Björn Rieckhoff: You know, what’s so exciting about Anthropic is that they’ve generated a huge amount of revenue in an extremely short time. This isn’t just a hype play in the sense of “We’re just selling a story”—there’s real substance to it.

Jan Thomas: They’re just getting started.

Björn Rieckhoff: But I saw that Anthropic officially stated: $47 billion in annualized revenue. That’s just...

Jan Thomas: ...not even five years.

Björn Rieckhoff: Five years after founding—just think about the scale of that.

Jan Thomas: That’s why I’d say: I’m crossing SpaceX off my list entirely. I’m curious to see how that turns out. It’s just too long a shot for me, somehow. Maybe in ten or twenty years they’ll be a monopoly and own the whole of space. But I think Anthropic, in the meantime, will become the most valuable company in the world.

Björn Rieckhoff: Yeah, I could definitely see that happening. I think that at some point—we discussed this in the last episode, I believe—you might see this shift from companies like Microsoft and Apple toward companies like OpenAI and Anthropic or the like. And I think that’s taking shape right now. I don’t necessarily want to downplay the Apples and Microsofts of the future; they’ll remain very relevant players. But you just have this speed that these supposedly smaller players bring to the table. It’s so immense that it’s hard to replicate. Even if you see them miles away in your rearview mirror and you signal to change lanes, you still can’t get into the left lane before they do.

Jan Thomas: I think Pip mentioned in the podcast—I’m not entirely sure if I’m quoting him correctly—that they actually managed to increase revenue per capita fivefold, meaning per user. That’s so crazy. And I had Henri Kühnert and Daniel Höpfner here yesterday, and they also said: If you were to take Anthropic away from someone now and just say, “From now on, this costs 500 euros,” then you’d be at a point—you’d at least start thinking—where you’d say: Okay, I’d really miss that. I think the price would be bearable, albeit grudgingly.

Björn Rieckhoff: Yeah, of course—all the things you can do with it. If I look at it from my own perspective in this classic consulting role: Would I personally pay 500 euros a month? Well, sure, if it means I can serve one more client, then those 500 euros are, so to speak, peanuts. And that’s just based on me, a small fry, so on an extremely small scale. The actual intrinsic value behind it—that’s way, way beyond that. You’re totally leveraging yourself—and that’s why those 47 billion in annualized revenue, so to speak, are actually a joke. If you look at the people currently moving in the circles of Anthropic, OpenAI, and the like—it’s almost become a kind of branding in the tech scene to be an ambassador for such companies. Hosting random events, applying that leverage, spinning the whole narrative further, without getting much else out of it besides merch. And then you walk around with a cool Anthropic cap, right? But there you just see—and these are all highly talented people doing this—there you see that leverage. It’s really just like it was decades ago: “I work for Facebook, I work for Google”—now you have: “I work for Anthropic.” That’s the changing of the guard.

Jan Thomas: And actually—with OpenAI, I was always curious to see what Johnny Ive would eventually come up with, some kind of hardware. I think if—no matter what it is—Anthropic were to announce tomorrow that they were releasing some kind of gadget, I’d buy it. No matter what. I never had that feeling with OpenAI.

Björn Rieckhoff: I think it’s simply this different approach, this different way of communicating, and so on, that creates a different foundation of trust.

---

Jan Thomas: Let’s move on to today’s topic. And now, of course, it’s actually difficult—the question is: Where does Europe fit into all this? That’s kind of the bridge, really. We just mentioned “European Dynamism”—we don’t want to brand it too simplistically here, but there are certainly some very exciting things happening, at least in France. And you’ve picked out a player who’s particularly active in this area.

Björn Rieckhoff: Exactly, right. So today’s topic—since we’ve already been on the podcast for 20 minutes—is actually SoftBank, which has announced a mega-deal with France and plans to invest up to 75 billion euros in data centers in France. The whole thing is happening in Hauts-de-France, somewhere between Paris and Belgium. They’ve actually hard-committed 45 billion euros over five years—probably the largest infrastructure project in Europe’s history. And I think that’s good insofar as it helps establish Europe’s AI sovereignty. But of course, I don’t think it’s so great that it’s not happening across Europe. You still need the global markets, players like SoftBank, to provide the necessary funding. And we might as well discuss right away where that money actually comes from—because it’s not just Japanese money behind it.

Jan Thomas: Well, first of all, I hope that the money was taken from the Stargate project and was just soft-committed there—hopefully. So that those 500 billion that were originally supposed to be invested in the U.S. for the AI strategy. I’m actually wondering—there are several questions here: SoftBank, you know them from the past as kingmakers and so on. They make big bets, and I find that exciting. But I don’t really know—why is this a bet? I also don’t know if this is even a VC case or if it’s their VC arm. But I would have thought this money could be found elsewhere.

Björn Rieckhoff: I think this infrastructure component plays a role here, along with a shift in power in this direction. We’ve been talking a lot in recent months about how capital is becoming politicized, regardless of where it is in the value chain. And I think SoftBank’s commitment in this area simply reflects that. This isn’t necessarily a VC case—to me, it’s a classic infrastructure project. They’re just building data centers. From that perspective, no, it’s not a VC case. But the way it’s being implemented certainly sends a signal of global political significance.

Jan Thomas: Well, there was that 500 billion special fund, that special federal debt. That would actually be the kind of project that could be financed with that. Not in France, of course, but following that logic where you say: We’re building infrastructure that will be relevant for the future. I don’t know about a data center like that—I’m not familiar with the calculations at all. It’s clear, of course, that they’re relevant and that they might eventually become a bottleneck, but I wouldn’t even know what kind of margin a data center like that has afterward. I’m not an expert on that.

---

Björn Rieckhoff: To be completely honest, I can’t really break down the business model behind it for you either. I couldn’t say right now that there’s an X percent EBIT margin behind it or anything like that, because I simply didn’t prepare for that. There are a few other reasons why I chose this topic. And that’s also (...) this idea of European sovereignty. If you look within Europe, then again: How is Germany actually positioning itself here? You mentioned the special fund, which is good. But even there, you see different approaches from Germany compared to France. There were several reasons that SoftBank made public, more or less officially: What drove the decision in favor of France? On the one hand, it was a personal visit by Emmanuel Macron to Tokyo, which made his personal conviction tangible once again. On the other hand, the decisive factor is, once again, energy. Specifically: In France, a megawatt-hour of industrial electricity currently costs 58 euros, while in Germany it’s 91 euros. This is simply because France has operational nuclear capacity that provides a stable baseload. And we decided a few years ago to phase it all out.

Jan Thomas: We’ve decided—though we’re almost putting a question mark on it—whether that was really us or whether it was a quick Merkel decision. But whatever the case may be.

Björn Rieckhoff: Yeah, I agree with you there.

Jan Thomas: I have rarely seen a decision of such far-reaching significance made so quickly behind closed doors.

Björn Rieckhoff: Yes. You know, I sometimes get the impression that when it comes to this... I’m a huge fan of renewable energy—let me just say that up front. I recently installed a balcony power plant on the railing of my own rental apartment. But there’s more to these decisions than just that aspect. This is becoming particularly clear right now when it comes to a certain degree of self-sufficiency: In a tense global political situation, you can no longer make a one-dimensional decision in favor of renewable energy just because it will be valuable in the future. You might also have to ask: Was nuclear power dangerous back then, to a certain extent? Yes. Do we have problems with final storage? Yes. But is it worth investing in it so that we can become more self-sufficient in the future? I would have written a big “yes” behind that, too. And that is precisely the short-sightedness that was present back then; you have to acknowledge that. From today’s perspective, that was an absolute mistake.

Jan Thomas: As I said, we’re moving too fast—but that’s also mentioned in the article as one of the reasons. I can also imagine that France’s branding or positioning has shifted. I think it’s happened relatively quickly—over the last three or four years—that they’ve positioned themselves as the AI stronghold in Europe, right?

Björn Rieckhoff: Yes, absolutely. I mean, it’s not just about infrastructure now—to be honest, France was able to play this card very well because other major nations in Europe, like Germany, had already opened the door. And that’s where Macron can step in and say, “All right, we can make the nuclear capacity available so that all of this can be built here.” At the same time, Paris has seen an incredible push toward the tech scene, essentially using places like Station F as a flagship—which is heavily subsidized and has become the one relevant hub in France. So not just Paris itself as a hub, but also the specific location. And of course, that’s something missing in Germany, since you don’t have that centralization. Pros and cons—there’s a lot to be said about that—but I have to say, France has positioned itself very, very sensibly in that regard.

---

Jan Thomas: When it comes to data centers and the speed at which they’re being built—I’m totally torn. I mean, I see what kind of megaprojects are being planned in the U.S. They’re basically the size of Manhattan or bigger. And then I read that they have an energy demand equivalent to the entire city of London. So, a data center the size of Manhattan with the energy demand of the entire city of London—that’s unimaginable. And then I find myself wondering: We’re already back in this kind of gold rush, this phase of greed, a bit like 2020 and 2021. With “look, you could even say the data centers will have to be in space” and so on. It’s all moving a bit too fast for me. So maybe we should pause for a moment and ask: How much capacity do we really need? And what are the right locations for it? You know, once something like that is built—they’re incredibly loud. There are a lot of residents protesting against them, a lot of citizens’ initiatives. But there are also measurements from people who had one of these things placed right outside their door—and half a kilometer away you still have 60 decibels; that’s totally crazy. You have this constant whirring in the air the whole time; it’ll drive you nuts. You’re out in nature right now, but just imagine if there were that kind of whirring all the time. That’s too much for me.

Björn Rieckhoff: No, that’s no joke; I don’t see it that way. But I do agree with you: it’s often a bit of a compromise you have to make.

Jan Thomas: Nah, nah, that was just my... I was actually going to say: I’m totally undecided too. On the one hand, especially right now with Anthropic—the speed at which it’s... Every day, when a new update comes out, I think: Now this thing can even draw Miro boards or something. I think that’s totally crazy. On the other hand: Okay, if it could do that in five weeks, that would be fine too. And then maybe we could take a step back and rethink data center planning on a global level. I don’t know. It’s not really up to us to have an opinion on this, but everything is moving pretty fast right now.

Björn Rieckhoff: Yeah, that's right. You know, another point I'd really like to bring up is SoftBank itself. You mentioned that Stargate project earlier, where they’ve committed a lot of money. That’s one of many topics. And SoftBank itself has become incredibly active; I believe it also holds an 11 to 13 percent stake in OpenAI.

Jan Thomas: But some of it has been sold, right? I read that we’ve already exited on a small scale at some point.

Björn Rieckhoff: Exactly, that’s right. That has already been scaled back. I think we actually covered that topic on the podcast back then. And yet, based on the current valuation, their stake is worth over 90 billion dollars, which of course has enormous collateral potential. The investments SoftBank is currently making aren’t necessarily on the balance sheet; instead, they’re all backed by collateral assets like the OpenAI stake. And you have Stargate, a $16 billion commitment, ABB Robotics—I think you had $9 billion there—and now this major deal in France. I believe SoftBank itself—I looked it up once—has over $130 billion in debt and a market capitalization of $270 billion. So there are also analysts who say: They’re taking on way too much. We’ll see how it plays out, but to follow up a bit on the point you raised earlier, along the lines of: How sustainable is all of this? At some point, we’ll just have to see how the whole thing actually generates margins. On the other hand, the whole thing is being financed super aggressively. So if it doesn’t go in the direction of eventually becoming profitable—these infrastructure projects, these Stargate commitments, these ABB Robotics plays—then you’ll also see globally active mega-investors getting into a bit of trouble, because a lot of their commitments are leveraged.

Jan Thomas: Yeah, at the same time I was reading—there’s ARM, and SoftBank has invested heavily there; they’re worth around 400 billion on the stock market. That’s also exciting. I don’t know how Michael Burry views this whole scenario right now. You see him say every now and then: “I’m shorting Nvidia here and there.” But this whole growth-at-all-costs approach, financed through debt, in the hope that all the pieces of the puzzle will fall into place—because that’s what’s happening with SpaceX right now, for example: People are now saying that water could become an issue again for all those data centers. We’ve already been there: data centers are a bottleneck. Then energy becomes a bottleneck. And now, suddenly, water is a bottleneck. That’s a pretty fragile chain. And that’s largely due to this speed. That’s why I keep saying: let’s pause for a moment.

---

Björn Rieckhoff: That probably wouldn’t be wrong. In that sense, that’s also why I often take a somewhat relaxed view of things and am actually quite happy to be in Germany. That’s a statement you might disagree with given the current circumstances. But this caution we often have—whether it’s regulatory in nature or in the way we weigh risky investments, etc.—has its advantages in that it gives you a kind of inherent pause. But it just goes on and on.

Jan Thomas: But I think it’s great how you’re turning that into an advantage now.

Björn Rieckhoff: Yeah, yeah, right. But when it comes to first-mover advantages—it’s not necessarily always what ends up being the most successful. What we’re already seeing is that Germany is also an enormously important market for international companies like OpenAI or Anthropic. Because AI adoption here is relatively high, because everyone is thinking about it—especially the broad middle class. And while you naturally had initiatives like Anthropic’s—which we discussed last time—with a software suite essentially for SMEs—and especially for micro-enterprises in the U.S., since they make up the bulk of the economy—in Germany you have a very broad SME sector, which shouldn’t be overlooked. And naturally, they adapt more cautiously than, say, the U.S. or U.S. customers. But once you’ve implemented the whole thing sustainably, it naturally gains tremendous momentum across the board. And here, too, we’re building on a certain level of expertise—the topic of proprietary datasets from earlier—which is exactly what benefits the use of AI. And I believe there’s now a massive wave of startups that are structuring this proprietary data so that AI can work with it. I think that’s also something we can expect to see over the coming months and years.

Jan Thomas: Yeah, well, I think a certain amount of time pressure is definitely important—and putting that pressure on yourself, creating a sense of urgency to act. I recently had someone on the podcast who works in AI, and he said: Well, German SMEs—keywords: AI and implementation—are only just now starting to learn about digitalization in the first place. That’s putting it a bit harshly, of course. Of course, there are all kinds—but I can already think of many companies I’ve seen where you get the feeling: Okay, they’re still mostly using pen and paper. On the topic of “What will become of Germany—is that an advantage or not?”—the lovely term “indigenous peoples” has popped up here recently. We just have to be careful that the world doesn’t look at us the way we look at indigenous peoples and say: “These are the ones who eventually turned their backs on progress.”

Björn Rieckhoff: I have to laugh at that too, and of course that’s a picture you can paint. I think if you consider what you actually see in practice—of course, I know people in my circle of acquaintances and friends who are active in the SME sector, and they say: We’re experimenting with a co-pilot that helps us draft emails and the like. That’s certainly true, but there are also other areas of application in research and development where I’m genuinely impressed and say: It’s amazing what’s being developed here—including in Germany with the existing engineering expertise. That’s definitely an asset we have there. Is it already widespread on a large scale? Probably not. But I do believe—I recently looked at a Bitkom study—that they found a 60 percent adoption rate of AI among companies with over 500 employees. And that’s not 95 percent, but it’s not zero either. 60 percent is probably already relatively good.

Jan Thomas: But there’s also a Bitkom study that says many employees are rebelling against the prescribed AI strategies internally—and especially against the tools. That’s also interesting. So that means: Not every tool implemented by management is the right choice—at least from the employees’ perspective. So it’s also clear that this is a phase of exploration. It shouldn’t sound cynical or skeptical at all, but... What’s important to me is this: I’d hate to see people spin this hesitation into an advantage, because that might lead to a kind of laid-back complacency that we really don’t need right now. You know, that’s what I’m getting at.

Björn Rieckhoff: No, I agree with you there. Earlier, that wasn’t so much in the sense of “We’re on the right track,” but rather a nuanced perspective along the lines of “It can also have its advantages.” But one thing is clear: capital velocity is a major issue—it’s almost become a unique selling point. You just have to be able to close deals quickly. And of course, a visit by the president to attract private capital to France is different from when we decide—with all due respect to the work of KfW and all the ministries handling grant applications—to implement those semi-governmental solutions again and say, “Yes, we’ll fund that.” But with application processes that take another two years. That’s different from tapping into the speed of private capital.

Jan Thomas: Yeah, and you just have to stay investable. Capital velocity is one thing—but you just have to somehow... Government debt is one issue, then demographic change, which we’re still dealing with anyway. That means: A company like SoftBank doesn’t just go to France because of Macron’s visit; that’s only part of it. They’ve probably also looked at Poland or the Nordics or whatever, and then they go to France for whatever reasons. And you have to understand those reasons and work on them so that next time they say, “Now let’s build this in Germany, too.”

Björn Rieckhoff: Yeah, that’s right. And I also think—“we’re building this in Germany”—that actually used to be a thing. At the World Government Summit, it was said that Germany is 96 percent dependent on foreign AI technologies. What an absurd figure. Sure, we live in a globalized world anyway, so this dependency also exists in the U.S. and China. There, too, you kind of—I read it in the same study—the U.S. has a 45 percent dependency on foreign technologies. Fair enough, they aren’t exactly independent either—we live in a globally interconnected world. But 96 percent is just too much. You’re really hanging on a lifeline from others, and that’s become a geopolitical issue by now.

---

Jan Thomas: And yet it’s also exciting—I wouldn’t even know what to do if now... I mentioned the downsides of this data center earlier. I mean, that’s really the kind of thing I don’t think anyone wants in their neighborhood. And I wouldn’t even know if you could pull that off in Germany. We’ve been looking for a nuclear waste repository since the ’70s and ’80s, after all. And that’s the thing—you just can’t pull something like that off politically in Germany. The will might be there, but then there’s always someone who throws a wrench in the works. I think you’d have to—federalism aside—at some point start looking at the country as a whole and say: Okay, if we really want digitalization and AI, then there will inevitably be some kind of consequence, and then maybe one region will be a little disadvantaged, but might get other benefits in return—you can manage that, after all. But then it always turns into this endless debate.

Björn Rieckhoff: I’m sometimes surprised by these debates, because of course each of us would like to have one foot in the countryside and the other in a major metropolitan area. But often, you can’t have both. At the same time: What has really driven Germany forward in the past? It was the Ruhr region—and if you look at the environmental devastation in the Ruhr region, you don’t exactly find yourself on a green path. And of course, when you look at it this way: the coal phase-out and so on, what happened to the region—not everyone is a winner in that. That's absolutely understandable. But still, you have to say: For that very important period, it was simply a necessary evil that people accepted in order to make progress possible. But we can't now make the whole thing acceptable to the general public if we consequently accept that the general public will be absolutely disadvantaged in the future. That would be the wrong approach, in my view. We need to figure out how we can work together to make certain compromises—so that, when in doubt, people say: “Okay, I’ll accept the hum of the data center two kilometers away, but I know that this data center ensures I’ll still have a place in the world down the line.” That would be the right compromise for me.

Jan Thomas: Like that guy Palmer—I just forgot his first name—the politician was on Lanz’s show once and said: For a solar farm like that, they had to build it right next to a highway on-ramp. So, right where it really doesn’t bother anyone. That was a ten-year approval process. And things like that—you just have to... I was just talking about this the other day with Caro Gabor. I like this image of—you probably know curling, that ice curling. There are these sweepers, I think they’re called—these guys who guide the stone exactly where it’s supposed to go, and at the right speed. That’s my image of the federal government. That’s how it should be. That’s the only job they have: to get the stone there quickly, where it belongs.

Björn Rieckhoff: I think it's awesome because I have such a negative view of curling as a sport. I hate curling, haha!

Jan Thomas: Curling is awful; I’d never watch it. But I like the image of this person—you know what I’m getting at—just bringing speed into it and making sure the ice is so smooth that there’s no friction, so you can get where you want to go as fast as possible.

Björn Rieckhoff: Hey, you know, I think both analogies fit pretty well. I mean, my contempt for curling as a sport and for the federal government—its current work. I’m going to stick my neck out a bit here. No, but I agree with you on that. The approval processes—they just need to be reformed. We talk about cutting red tape all the time. And I get the impression that no matter what you do, it just leads to more red tape. Should I give you a personal example?

Jan Thomas: Yeah, go ahead.

Björn Rieckhoff: There's actually something that's currently in the works, and maybe I'm shooting myself in the foot here. You might be familiar with the transparency register.

Jan Thomas: Sure.

Björn Rieckhoff: Anyone active in VC is familiar with this transparency register. It’s about reporting who the beneficial owner is behind the capital structure. I have an asset management LLC, and I got married. I changed my last name from Loose to Rieckhoff. Yes—now I’m facing administrative proceedings.

Jan Thomas: You shouldn’t do that either; it’s not intended.

Björn Rieckhoff: What’s that supposed to mean? I’m being accused of trying to cover something up because I changed my name. And so much for reducing bureaucracy.

Jan Thomas: There’s a preliminary suspicion here, Björn. Sorry about that.

Björn Rieckhoff: Especially when it comes to the sums of money I’m shuffling back and forth every day. It’s the height of absurdity. Everyone I tell this to starts laughing out loud. Even the employees involved probably have to do the same, and then they think: Yeah, but we’re just following orders.

Jan Thomas: And you know, that’s the exciting part. You’ve probably seen this with lots of startups—but at least that’s how it is for me, and also with many of the people I talk to on the podcast. We’re always talking about focus, prioritization, and focus. Those are the two most important things you need to master as an entrepreneur. So: Work on the right things. And you know, I think the government should also—there should be some kind of focus area where you say: Then maybe government employees will understand that they don’t necessarily have to be working on certain issues right now—like, for example, writing to Björn Rieckhoff, formerly of Loose.

Björn Rieckhoff: That would be desirable.

Jan Thomas: This isn’t just about me personally, but I really want people to feel that we’re working on the right things and simply setting aside other matters that are more of a nuisance—things that aren’t exactly “curling.”

Björn Rieckhoff: Right, yes. So Jeff Bezos—whatever you may think of him—recently argued that 50 percent of income earners in the U.S. shouldn’t be subject to income tax at all, because the tax revenue generated from them is negligible. You just have to focus on the relevant, big issues; then you’ll make progress faster. And you’ll save yourself so much effort on the smaller things. A confession of guilt—did I report the name change to the transparency register? Officially: No, I unfortunately missed that. Of course, I made a formal mistake there. But you could also come up with the idea of asking: Why isn’t the transparency register synced with the commercial register? Why do you need this separate notification? That’s just the kind of pragmatism I sometimes wish for. And yes, I’m very much in favor of protection against terrorist financing and the like. In that respect, I understand the intention behind it. But as you say: It just has to stay proportionate somewhere.

---

Jan Thomas: And that’s what makes it so exciting. Maybe a podcast recommendation: I just listened to a new episode of Ronzheimer, which featured Martin Greive from Handelsblatt as a guest—and it was a really good conversation about the question: How much money actually needs to be saved in Germany, and where can it come from? And a few points do come to mind. For example, handling data twice, three, four times in the chain, as you just mentioned. Gunter Dueck once said: These are the “screen-stare-ers” who look at the screen all the time and copy data from A to B. You enter it once, and then it gets printed out, retyped, re-entered, retyped again. And those are just the kinds of processes that really need to be addressed—now that we have a Minister of Digitalization—that would be such a great topic to tackle. Well. Björn, we have to be careful; we’re running out of time. I think this is more suited for a pub or a bar sometime. We won’t be able to resolve this today. But back to SoftBank: Could you briefly summarize your take on that? Will we see more of this? Are there more SoftBanks out there looking for targets like this? Should Klingbeil or Merz go on a little tour? Or would you say this was a one-off deal, and we’re now just tagging along with France and kind of hanging on by a thread?

Björn Rieckhoff: I would actually say: A European approach would definitely make sense, so we don’t all end up doing our own thing again. I don’t think that makes much sense. Then we’ll just end up with a competitive situation that we really don’t need. Of course, it’s kind of about—I’ve seen now, for example, that Schneider Electric is the partner for construction and hardware in this major project. It would just be nice to see a Siemens there, from a German perspective. You have to try to protect national interests there. But I’m not a fan of everyone going their own way and then us ending up saying: We’re not doing a 5-gigawatt project now, but a 10-gigawatt project. And that doesn’t benefit anyone. It would probably make sense to further open up and develop the European electricity market, to collaborate—so that we can present a united front as Europe. As for parochialism, that’s a thing of the past—it’s never done anyone any good.

Jan Thomas: Good closing remark, Björn. I certainly hope so. Perfect. Thanks a lot—it was an exciting episode, I must say. A lot packed in again. Yeah, we don’t want to fall into “European Dynamism,” but I think we can already see: things are happening. That’s a good start.

Björn Rieckhoff: Yeah, we’re just not such a small player as the Americans and the Chinese always like to claim. You can also see from investments like this that we play a relevant role geopolitically, and that’s good to see.

Jan Thomas: And here too—we’ve talked about Recursive, then AMI Labs in France, and then Inevitable Systems from London and so on. There’s also a whole bunch of amazing companies raising massive rounds. So I think this is kind of a wake-up call right now.

Björn Rieckhoff: Yeah, that feels good, that’s right.

Jan Thomas: Cool, thanks so much for joining us, and I look forward to the next episode.

Björn Rieckhoff: Thanks. See you later.

Jan Thomas: See you then. Bye. Ciao.

---

About Björn Rieckhoff

Björn Rieckhoff is an independent advisor and business angel with nearly ten years of experience in early-stage venture capital. He helped build Cavalry Ventures as its first employee and later became a partner of the fund. Today he supports founders more directly with fundraising — sharpening their story, stress-testing business models, and setting up lean financing processes. With over 80 transactions and board seats from seed to Series B, he brings this perspective as a sparring partner for entrepreneurs.

About Startup Insider

Startup Insider is the industry portal for the startup scene in the DACH region. It covers news from all regions and industries, along with an overview of key players and events in the German-speaking startup world.

Links

• Oura Reaches $11 Billion Valuation (episode on Oura)
• Self-training AI and the Rise of the Application Layer (previous episode)