Daniel sent us this one, and it's a genuinely meaty question. Israel is frequently cited as having the highest per-capita startup rate in the world, over six thousand active startups in a country of under ten million people. But the more interesting puzzle isn't just why Israel, it's why the gap is so wide. Japan is wealthy, educated, technologically sophisticated, and produces almost no startups relative to its size. Germany, same story. The Gulf states have capital to burn. So what's actually going on? Daniel wants us to go deep on the full stack: culture, bankruptcy law, military service spillovers, university-industry pipelines, diaspora networks, immigration policy, the role of anchor exits seeding the next generation. Herman, I have a feeling you've been waiting for someone to ask you this.
Oh, I have strong opinions about this. Strong, well-sourced opinions.
Of course you do. Also, quick note before we dive in, today's episode is powered by Claude Sonnet four point six. Just thought listeners should know who's doing the heavy lifting around here.
The friendly AI down the road. So let's start with the statistic itself, because it's worth grounding this properly. Israel has more than six thousand active startups, which at its population size works out to something like one startup for every fifteen hundred people. There is nowhere else on earth that comes close to that ratio. The United States has more startups in absolute terms, obviously, but per capita Israel laps the field. And the figure that really stops me in my tracks is the venture capital number. Israel attracted around twelve billion dollars in VC funding last year, which puts it fifth globally in absolute terms. Behind the US, China, the UK, and India. Countries with populations ten to a hundred times larger.
That's the contrast that I think most people haven't really sat with. Because when you say "startup nation" it sounds like a branding exercise. It's on the cover of a book, it becomes a slogan. But twelve billion dollars in a country of under ten million is not a slogan, that's a structural reality.
Right, and the comparison to Japan is where I want to start because it's the most clarifying. Japan has the third largest economy in the world. It has extraordinary engineering talent, world-class manufacturing, a culture that prizes precision and craft. And yet Japanese startup activity is, relative to its size and wealth, remarkably thin. So the question isn't resources. Japan has the resources. The question is what resources alone cannot buy.
Which is, I suspect, the answer to most of this episode.
Let's actually define the terms, because "startup-friendly" gets used loosely. What we're really talking about is a cluster of conditions that lower the cost of attempting something new and raise the expected value of doing it. Those are two separate levers and countries get them wrong in different ways.
Unpack that a bit.
The cost side is things like, can you fail without being destroyed personally? How long does it take to register a company? Can you hire and fire with some flexibility? Is there capital available at the early stage when you have nothing but an idea? Those are friction reducers. The expected value side is different. That's about whether the upside is real. Are there acquirers in the market? Is there a talent pool that can help you scale? Do customers exist who will actually pay for new things?
A country could score well on one and badly on the other. You could have low friction but no upside, or high potential upside but so much friction that nobody tries.
And I'd argue most of continental Europe is the second case. The upside is theoretically there, but the friction is high enough that the expected value calculation keeps coming out negative for a lot of would-be founders.
Then there's a third thing, which is harder to quantify. Whether the culture actually sends people toward that calculation in the first place. You can have low friction and real upside and still have a society where the prestige gradient points at large employers or government or academia.
That's the one that's almost impossible to policy-design your way out of. You can reform a bankruptcy law in a parliamentary session. You cannot reform what a twenty-two-year-old thinks a successful life looks like in a generation.
Which is why the military piece is so interesting—it’s one of the few mechanisms that actually shifts that gradient at scale.
And that’s where Israel diverges from every other comparison case. It’s under-theorized, even by people who know the story well.
Because the usual version is "Unit 8200 produces tech founders" and then people move on.
Right, and that's true but it's also the least interesting version of the claim. Unit 8200 is Israel's signals intelligence unit, roughly analogous to the NSA in terms of function, and the alumni list is staggering. Check Point, Waze, CyberArk, Palo Alto Networks co-founded by an Israeli who came through that world. The companies founded by Unit 8200 alumni are collectively worth well over a hundred billion dollars. But the mechanism is more interesting than the roster.
What's the mechanism?
A few things happening simultaneously. First, you're taking people at nineteen, twenty years old and putting them in charge of systems where failure has real consequences. Not career consequences, actual security consequences. That produces a very different relationship to responsibility than, say, a graduate training program at a large bank.
The stakes are load-bearing from day one.
From day one. And the problems they're working on are hard. Signals intelligence, cybersecurity, real-time data analysis at scale, these aren't simplified training scenarios. So by the time someone finishes their service at twenty-two or twenty-three, they have solved problems that most people their age in other countries haven't encountered yet. And they've done it in teams, under pressure, with incomplete information.
Which is a fairly accurate description of a startup.
The other thing that doesn't get enough attention is the network density it creates. You come out of a unit like 8200 and you know fifty people who've been through the same crucible. You trust them. You know how they work under pressure. That's not a LinkedIn connection, that's a co-founder pool.
That trust is pre-built before anyone's trying to raise money or ship a product.
Right, and co-founder trust is one of the most underrated inputs to startup success. There's actually data on this. A long-running study on sixteen years of startup outcomes found that co-founder relationship quality, which is partly about trust and partly about complementary skills, is one of the strongest predictors of survival past the three-year mark. The military service is essentially manufacturing that relationship quality at scale, across an entire generation.
Though I want to push on something. Mandatory service exists in other countries. South Korea, Finland, Switzerland. None of them have produced anything like Israel's startup density. So the service alone isn't sufficient.
No, and this is where the cultural tolerance for failure becomes the multiplier. Because the military piece gives you skills and networks, but if the culture punishes failure harshly, those skills and networks don't get deployed into new ventures. They get deployed into safe careers instead. Israel has a cultural norm, unusual, where having started something and failed is not a mark against you. It's almost a credential.
Which is the opposite of Japan, where the social cost of business failure can follow someone for years. And the legal cost, historically.
The bankruptcy comparison is stark. Japan's corporate bankruptcy procedures have traditionally run well over a year, and the personal liability exposure for founders in a failed company is significant enough that many people simply don't start. Estonia, by contrast, has streamlined its process to around three months. Israel sits closer to that end of the spectrum. What bankruptcy law is really doing is pricing the downside of trying. If the downside is catastrophic and prolonged, the expected value calculation we talked about earlier shifts hard against attempting anything.
Japan has the talent, has the capital, has the infrastructure, and then the legal framework essentially taxes the attempt itself so heavily that rational people don't make it.
The cultural norm reinforces the legal reality. The law says failure is expensive. The culture says failure is shameful. Those two signals pointing in the same direction is very hard to overcome through sheer individual will.
Which also explains why the Gulf states haven't replicated this despite throwing serious money at it. You can fund a startup ecosystem. You can build incubators, you can create state-backed venture vehicles, NEOM is the most expensive example of this logic running at full scale. But if the cultural substrate and the legal substrate don't shift, the capital is swimming upstream.
The Gulf case is interesting because it adds a third variable, which is that the failure tolerance issue there is partly about what failure means for family reputation in ways that go beyond the professional. The social stakes are higher than just career. So even very well-funded founders are operating with a much heavier downside weight.
You can't fix that with a better term sheet.
Which brings us to the university piece, because that's where the infrastructure side of the equation starts to look very different from the cultural and legal story.
Technion is the obvious place to start.
Technion, the Israel Institute of Technology in Haifa, has been running startup incubators and commercialization programs since the nineteen eighties. What makes it unusual isn't just that it has a tech transfer office, most research universities do at this point. It's the density of the pipeline. Technion graduates are founding companies at a rate that, per student, rivals MIT. And there's a structural reason for that. The university has historically had very permissive IP policies compared to, say, most European universities, where the institution retains ownership of anything developed on its premises and the commercialization process can take years of committee review.
A Technion student who builds something actually owns it.
Or has a clear, fast path to owning it. That distinction sounds administrative but it changes the incentive entirely. If I know that spinning out my research into a company is going to take three years of negotiation with the university's legal department, I'm going to think twice. If the path is clear and the timelines are short, I start the company.
Europe has world-class research universities. Cambridge, ETH Zurich, Delft. The research output is excellent. And yet the commercialization rate is a fraction of what you see out of Technion or Stanford.
The Global Innovation Index data from last year actually captures this gap pretty clearly. European countries rank well on research inputs, patents filed, academic citations. But the translation from research to commercial venture is where they lose ground. And it's not a talent problem. It's a structural problem. The pipeline between the lab and the market is clogged.
Which is partly the IP issue and partly, I'd guess, the employment regulation issue.
Very much so. If you're a researcher at a German university and you want to leave to start a company, you're potentially giving up a tenured position with significant employment protections. The cost of trying is a stable career. In Israel, and to a large extent in the US, the expectation is that people move between academia and industry and back again. The career path is porous in a way that European systems often aren't.
Stanford and Silicon Valley have basically built that porosity into the geography. You're forty minutes from Sand Hill Road.
The density of that proximity matters more than people give it credit for. It's not just that the capital is nearby. It's that the norms around what's possible get transmitted through proximity. You're at Stanford, you know three people who left to start companies, you see the outcome, the feedback loop is short. In a system where nobody around you has done it, the imaginative leap required is much larger.
To the capital access question directly, because twelve billion dollars a year into a country of under ten million is one thing. But where does that capital actually come from, and why does it come here rather than somewhere else?
A piece of it is the diaspora network, and I think this is underappreciated in the standard "Startup Nation" narrative. The global Jewish diaspora has a few specific properties that make it unusually effective as a capital and mentorship network. It's large, it's geographically distributed across major financial centers, New York, London, Los Angeles, and it has a long tradition of institutional trust-building through shared community structures. That's not a mystical claim, it's a network topology claim. When you have a diaspora that is densely connected across multiple high-capital cities and that has pre-existing trust infrastructure, capital can flow through it faster and with lower friction than through a cold institutional channel.
The Indian diaspora in Silicon Valley operates similarly. Same basic mechanism.
Exactly the same mechanism. Which is why Silicon Valley has become so effective at funding Indian-founded companies. The diaspora acts as a bridge, providing not just capital but pattern recognition. Investors who've seen what this kind of founder looks like, what their trajectory tends to be, what the cultural context means for how they build teams.
That pattern recognition is hard to fake. A European venture fund trying to evaluate an Israeli cybersecurity startup is doing it from further away, in every sense.
And the cybersecurity angle is worth dwelling on because it's where the military and the capital story converge. The research from WTN Insider earlier this year put cybersecurity at thirty-six percent of Israel's total tech funding, despite representing only seven percent of its startups. Seven percent of the companies capturing thirty-six percent of the capital. That's not random. It's because Unit 8200 and similar units produced a generation of founders who understand offensive and defensive cyber at a level that almost nobody else in the world does, and the market for that knowledge is enormous and growing.
The military pipeline doesn't just produce generalist entrepreneurs. It produces founders with a specific, deeply valuable technical edge in a sector that happens to be one of the fastest-growing areas of enterprise spending globally.
The anchor exit story feeds back into this. Check Point went public in nineteen ninety-six. That single event seeded the next generation, because the people who were early at Check Point, who made money, who understood the cap table, who knew what a successful exit looked like, those people became the angel investors and mentors for the next cohort. Waze sells to Google for over a billion dollars in twenty thirteen. The recycled capital and the recycled knowledge from those exits created the soil for everything that came after.
Europe's relative scarcity of those anchor exits is part of why the ecosystem there hasn't compounded the same way. You don't get the recycling.
It's a flywheel problem. The first few rotations are the hardest. Once the flywheel is moving, the exits produce the angels who fund the next wave who produce the next exits. Europe has had exits. But not at the density or the scale that generates that self-reinforcing cycle at the national level. The US has it, Israel has it, and increasingly India has it. Most of continental Europe is still waiting for the flywheel to catch.
The Gulf states are interesting here too because they're trying to buy their way into that flywheel. Import the exits, import the founders, build the ecosystem through capital injection rather than organic compounding.
The jury is still out on whether that works. The honest answer is we don't have a clean historical example of a country successfully importing a startup culture wholesale. Singapore is probably the closest case study, and Singapore took decades and involved very specific policy choices around immigration and foreign talent that are politically difficult for Gulf states to replicate at scale.
Singapore is worth a beat here because it's the counterargument to the idea that culture is destiny. Singapore doesn't have Israel's military pipeline. It doesn't have a diaspora network of that kind. And yet it built a functional high-density startup ecosystem.
Through very deliberate institutional engineering. Singapore essentially decided it would become a node in the global capital network and then systematically removed every friction point that might deter a foreign founder or investor from choosing Singapore as a base. Tax structure, IP law, employment rules, residency pathways for founders. The government treated it as an infrastructure project and executed it with the kind of precision that Singapore executes infrastructure projects.
Which is its own lesson. You can get there without the organic cultural preconditions if the institutional will is strong enough and the execution is disciplined enough. It just takes a very long time and a very specific kind of government.
It probably helps that Singapore is a city-state, which means the coordination problem is much smaller. You're not trying to reform a federal system with competing jurisdictions. You're reforming one city.
There's something almost unfair about how well that scales. Estonia is the same story in miniature. Small, coherent, fast to move.
Estonia's bankruptcy reform is a good example of what speed looks like when the political will exists. Three months from filing to resolution, versus Japan's year-plus. That's not a marginal difference. That's a signal to every potential founder in the country about how seriously the system takes their time and their risk. But it also raises the question: how do other countries, without Estonia's advantages, even begin to approach that kind of efficiency?
So what do you actually do with all of that, if you're a policymaker sitting somewhere that isn't Israel or Singapore or Estonia?
The bankruptcy law point is probably the most underrated lever available, because it's concrete and it's changeable. It doesn't require you to rebuild your culture. Japan's lengthy insolvency process, over a year of proceedings in many cases, functions as a tax on trying. You reform that to something closer to Estonia's three-month window and you've immediately shifted the expected cost of failure for every potential founder in the country. That's not a cultural intervention. That's a legal one.
It sends a signal even to the people who never go bankrupt. The existence of a clean exit from failure changes how people think about entering.
The second lever is the university-industry pipeline, and specifically the IP ownership question. If European research universities moved to Technion-style permissive IP policies, where the path from lab to company is short and clear, you'd see commercialization rates move. The talent is there. The research is there. The clog is institutional.
For listeners who aren't policymakers, which is most of us, I'd say the practical implication is knowing what to push for. Employment protection laws that make it catastrophically costly to leave a stable job to start something are a drag on the whole ecosystem. Immigration rules that make it hard for a foreign founder to stay and build are a drag. These are specific, nameable policies, and they're things that civic pressure can actually move.
The diaspora network insight is also worth sitting with, even if you can't legislate it. What it tells you is that trust infrastructure and warm capital are not luxuries. They're load-bearing. Communities that invest in building dense, high-trust professional networks across geographies are building something that pays compounding returns over decades.
Which is a very long game.
But the countries that are winning at this started playing it a long time ago.
Most of them didn't know they were playing it at the time. That's the other thing. Israel didn't sit down in nineteen forty-eight and design a startup ecosystem. The military pipeline was a security necessity. The diaspora network was a survival structure. The tolerance for failure was born out of circumstances where failure was the baseline condition and you had to try again anyway. The ecosystem is, in a strange sense, a byproduct of adversity more than a product of planning.
Which raises the hardest version of the replication question. Can you engineer the preconditions for something that emerged organically from pressure that most countries would rightly prefer never to experience? Singapore gets close, but Singapore had its own version of existential urgency in the early decades. Estonia had Soviet occupation as the forcing function. There's a pattern there that's uncomfortable to name directly.
The countries that built the most resilient innovation cultures are often the ones that had the least choice about it.
I don't think that means other countries are out of luck. It means the lever that's actually available to them is probably the institutional one, not the cultural one. You can't manufacture the urgency. You can lower the bankruptcy timeline, open the IP pipeline, clear the immigration path. Those are real. They compound slowly, but they compound.
That's where I want to leave this, actually. The question of whether any of this is replicable doesn't have a clean answer, and I think the honest position is that the full package probably isn't. But pieces of it are, and the pieces that are happen to be the ones that policy can reach. That's worth something.
It really is. And the global trend is moving in that direction. More countries are watching what worked and making at least some of the right moves. Whether it's fast enough, and whether the compounding kicks in before the window closes, that's the open question.
A good one to sit with. Thanks to Hilbert Flumingtop for producing this one, and to Modal for keeping the infrastructure running. This has been My Weird Prompts. If you've been enjoying the show, a review on Spotify goes a long way. We'll see you next time.
