Subscribe
Share
Share
Embed
Born & Kepler is named after the mathematician and scientists Max Born and Johannes Kepler. This bilingual podcast, offered in both German and English, dives into the expansive world of Artificial Intelligence (AI), exploring its foundations, evolving technology trends, academic search, and its impact on businesses and society.
Born & Kepler will feature a diverse lineup of experts from academia, venture capital, private equity, journalism, entrepreneurship, CTOs, and policymakers. Each guest offers unique insights into how AI is reshaping their sectors and what we might expect in the future.
Our goal is to provide a deep understanding of the core principles and breakthroughs in AI, enabling you to stay updated with the latest advancements in AI technologies and how they are transforming industries. During our episodes, we will explore how AI is influencing business strategies, optimizing operations, and driving innovation. We will also explore the ethical, social, and regulatory aspects of AI in everyday life.
Andreas Deptolla (00:01.918)
Jeff, welcome to the podcast. It’s great to have you.
Jeffrey (00:06.012)
Thanks for having me.
Andreas Deptolla (00:07.854)
Perfect. Jeff, you’ve had a lot of success as an individual investor and as an investor with Golden Gate Ventures. I think our research counted nine unicorns you’ve invested in. And all of that can quickly translate into value. But maybe tell us a bit about the history of Golden Gate Ventures. You started in 2011—what gap in the market did you see back then?
Jeffrey (00:43.677)
If we go back to that decade: I launched the first Founder Institute program in Asia. That brought us into Southeast Asia. It’s a four-month evening program with three to four mentors every night over roughly 14 weeks. What you’ll see is each chapter brings in many interesting mentors. That’s when I met my co-founder Vinnie—he was backpacking in Asia after selling his company in the U.S., and I roped him in. Every chapter needs mentors, and he was in.
After about 10–11 months we realized the teams were great, the mentors were great. If mentors are starting companies, they’ll probably invest. So the ecosystem in the region was forming… but there wasn’t much organized capital. There were many family offices, some corporates dipping in and out, but not many early-stage venture funds. That’s how the idea for our fund was born.
Andreas Deptolla (02:13.742)
So being there at the beginning—one of the first VCs in Southeast Asia—probably gave you access to better deals. Could you put us back in that time and how the model has evolved since 2011?
Jeffrey (02:34.3)
Back then, consumer and fintech dominated. You could call it a copycat culture—everyone was trying things that worked elsewhere. Our timing helped. We made good bets and investments.
But there was also a big catalyst: the Singapore government around 2007–2008 reallocated a slice of its annual R&D budget into startups. That set the stage.
They then launched a Technology Incubation Scheme that we used. Everything was lining up. The government was supportive and treated us very fairly. They did this because they were trying to develop the local ecosystem—so more people would join startups. I think we got a lot right, and the government helped.
Entrepreneurial energy in the region was rising, and more capital followed. Then interesting deals started coming.
Andreas Deptolla (04:02.836)
When you say the government—was that as an LP in the funds, or how did they invest?
Jeffrey (04:13.82)
Back then they had the Technology Incubation Scheme: if we invested, say, around 87,000, they would match it up to 500,000. If the company did well, we could buy out their stake; if not, we’d write it off as debt. That gave companies over half a million in equity capital.
Funds might own a fifth or a sixth after that. It gave us the confidence to do deals, and it made the ecosystem optimistic. Many founders came online, built companies, then found capital. It became a real ecosystem.
Andreas Deptolla (05:12.074)
And can you share some numbers for Golden Gate Ventures—how many funds, how many investments?
Jeffrey (05:20.442)
Sure. We’ve been at it for 12 years. We’re investing out of Fund IV now, a $70 million fund. Still early stage: pre-A, pre-seed and seed, and Series A. About 85–90% in Southeast Asia.
Around 10–15% outside the region. We’ve also invested in Europe.
Andreas Deptolla (06:02.062)
And for the four funds, can you share how they rank in terms of performance for LPs?
Jeffrey (06:10.395)
I can say this: Fund I is definitely top quartile—if not top decile. Fund III is top quartile. Fund II is very close to top quartile—we’re a bit over the hump. And Fund IV is very [unclear]—but tracking well.
Andreas Deptolla (06:33.742)
When you say top quartile, you’re comparing against all VCs globally?
Jeffrey (06:39.898)
Yes, correct—within our region and asset class, but that’s the global reference set.
Andreas Deptolla (06:45.774)
Okay. So would it be fair to say 3x–4x gross return on investment?
Jeffrey (06:57.53)
On paper, yes. Realized DPI—not yet. We need time. On paper, yes. Fund I, yes—much higher. Fund I is much higher than Fund IV.
Andreas Deptolla (07:02.382)
On paper because not all companies have exited yet and there are questions around timing.
Andreas Deptolla (07:18.926)
Is that mainly timing—exits pending—or changing market multiples, or different vintages?
Jeffrey (07:28.718)
Fund IV is a 2023 vintage. Our first checks are still fresh. It’s on track—many have followed-on rounds. Fund IV looks very good. Fund III is 2018, Fund II is 2015.
I’d say most are well distributed. It’ll take a bit more time, but it should happen.
Andreas Deptolla (08:00.942)
Congrats on all the success. You mentioned “copycats”—taking an idea from the Valley or elsewhere and localizing it to another region. What’s your view on copycats? Would you invest in them? How do you see it?
Jeffrey (08:31.161)
Copycats are fine. Whether you’re cloning or doing something entirely new, both paths are equally hard. Copycats are often more thought-through—you can observe mistakes and improvements elsewhere.
If you see it working in another country, you can adapt it. That gives you more confidence that a western competitor won’t just roll into your country and crush you—because you already see what’s needed locally. When you look at the market you may spot things they don’t.
Andreas Deptolla (09:15.214)
No!
Jeffrey (09:28.921)
It’s a good way to help people get into company-building more easily. Many of those founders weren’t highly technical. A lot came from consulting or business school. So you raise quickly, build a technical team, and start executing. That’s what many founders did—and that was perfectly okay. The only time it wasn’t okay was when you copied the wrong thing. If you didn’t do enough research—especially regional research—you could pick the wrong thing, burn money for 3–4 years, and struggle because the original idea didn’t fit your market in the first place. You then try to pivot it into shape—which is hard. Sometimes it works, but locally it can take more time. Maybe your timing is a few years early, maybe revenue expectations are too high.
Andreas Deptolla (17:09.416)
Yeah, it’s interesting. In Germany we saw a lot of copycats initially, especially with Rocket in Berlin. As a result, an ecosystem emerged—because once you have early traction, you can recycle talent and capital. But there’s always a bit of debate on how “original” those things are. If we look at the unicorns you invested in—like Carousell, Carro, Funding Societies, Ninja Van, FlowAccount—just to name a few: how many of these were copycats versus truly original ideas?
Jeffrey (18:04.071)
I’d say some were original. Koda is original, Carousell is…
Jeffrey (18:25.655)
It’s mobile-native first.
For Carro—very similar models existed in Germany and also in China. But in execution, things differ across countries. Emerging vs. developed markets are different. Conceptually similar, but execution gets more complex.
I’d say it’s a mix, but there’s a lot of localization you must do.
Andreas Deptolla (19:23.07)
If you look across your investments that became very successful—is that an accurate representation or just a random sample of what I happened to see?
Jeffrey (19:40.038)
We’re pretty good at marketplaces, communities, consumer, and fintech/payments—B2B as well. That fits the nature of the market, and the timing was right.
Andreas Deptolla (19:59.31)
Why was that?
Jeffrey (20:10.758)
Some companies resembled what you saw in the West, so it was easier to understand and diligence.
Andreas Deptolla (20:19.054)
What market conditions were attractive? It’s always tough because you need both supply and demand—and supply-side applications are expensive.
Jeffrey (20:30.341)
I think back then anything you could put online—B2B or B2C—had white space. Most platforms were new, and there were no dominant incumbents. Over time you saw specialized vertical platforms. It was mostly demand-driven at that time.
Andreas Deptolla (21:07.569)
What methodology did you use to source startups? Was it data-driven? Did you have a technical algorithm to find the right ones?
Jeffrey (21:25.957)
We looked at demographics—we knew demographics mattered. We watched the U.S., China, India—when something happens there, trends rhyme. We knew certain categories were coming. We had a view that a company solving this would be needed—and that’s where we invested.
Andreas Deptolla (21:55.352)
And today—is it more data- and AI-driven? How do you pick the right adventures now?
Jeffrey (22:02.277)
Yes. It’s a bit different now. Previously, consumer and fintech dominated. Now there’s a lot more B2B. We didn’t have many B2B successes historically, so we’re doing more research and being more data-driven. We have AI models—company-scoring models—in a different way. We size markets more precisely; we do things others might not. There’s a rationale: you don’t want to be too early, and obviously not too late. You also want to get market size and adoption curves really right. Sometimes you can back the right company, but it takes 14 years—that wrecks the math. If your fund size is big and the market is smaller, you get into trouble. So now we’re more mathematical: entry valuations, how much capital to raise over four rounds, dilution models, when you…
Andreas Deptolla (23:08.91)
Hmm?
Jeffrey (23:34.66)
…and when a company reaches a certain scale. It all matters—and of course fund size matters too.
Andreas Deptolla (23:42.863)
Given your fund size—what makes a market or TAM attractive? Is there a number? And for a company—do you need X revenue, growth, adoption? Are there specific hard filters you look for to decide?
Jeffrey (24:11.267)
I think the sector needs to be understandable—to the buyer. The problem statement must be crystal clear. Then different cities and countries have different adoption curves, cultures, startup experience, and budgets. Some can build in-house; some need to buy from you. You have to map all that.
There are different customer classes: some have budget and are actively searching; some have budget but don’t understand what you’re talking about yet—you need to educate them; some love your solution but have no budget. So in each country, in each city, at that moment in time, you must really understand how demand dynamics work. If you’re going across three or four countries, you need playbooks for each. If you’re focusing on one country, you need to understand it deeply—and sometimes you’ll still find the market is too small or cycles too long.
Andreas Deptolla (25:14.51)
Mhm.
Jeffrey (25:40.322)
Then you need a counter-strategy: add two or three more products to sell to the same customer, or find other rails—like fintech or lending—around your solution. Things evolve. But what we’ve learned is that markets here can be hard to break and then to scale to certain revenue thresholds. Not every sector can reach those thresholds—you need to be careful.
Andreas Deptolla (26:26.35)
Do you have specific founder-pattern preferences? Basics are clear: right market size, product-market fit, clear problem statement. But do you look for solo vs. co-founder teams, serial entrepreneurs, anything like that?
Jeffrey (26:54.754)
In the past, complementary co-founders were ideal. If three co-founders covered product, go-to-market, etc.—great. Sometimes a strong solo founder who can hire around them also works. Now it’s a bit different. It’s less negotiable that someone must truly own product. I’d say most teams need product leaders. Sometimes you need product lieutenants early—before you scale sales. Product iteration must be really fast now, because everyone knows B2B with AI is the path forward. Verticals or job roles are shifting everywhere. Everyone has the same idea—so it’s competitive. You need to iterate quickly and know customers deeply. We’re back to the good old days of young founders shipping under their desks—not a lot of ceremony.
Andreas Deptolla (28:23.726)
You said AI and new tooling let you build solutions faster at lower cost. How does that change the game? Does it mean founders need less capital? Does it get more competitive? What consequences do you see?
Jeffrey (28:56.225)
Yes. Teams that design, build, and ship along one line don’t need that much capital. Younger founders also need less, because burn can be very low. If they land quickly in a country that knows how to buy from startups, revenue ramps much faster than before. In the U.S., it’s not unusual for a 3–5-person team to reach $3–12 million ARR with maybe half a million raised. We’re seeing that here too. We’re usually about a year behind U.S. adoption cycles. My sense is: in this season you’ll see many more companies in Southeast Asia that raise modest amounts and execute really well—and realize they don’t actually need many people.
Andreas Deptolla (29:29.998)
Mhm. Wow.
Jeffrey (29:54.976)
And they’ll earn good money—and realize they don’t need large headcount.
Andreas Deptolla (30:03.79)
When you say the region is a year behind—why is that? Culture? Different procurement cadences you mentioned? What’s the distance between Southeast Asia and the U.S. in startup terms?
Jeffrey (30:32.896)
I’d say there’s a decade of compound experience in the U.S. on what to launch and how.
Jeffrey (30:58.015)
The top 10% of stealth companies—people in the scene often know who they are and even what they’re building. In the U.S., many know what’s being built early. That creates a baseline of momentum. Elsewhere, founders start on similar ideas—but they might be 9 to 12 months late.
Andreas Deptolla (31:02.222)
Hmm.
Jeffrey (31:28.448)
So…
Andreas Deptolla (31:29.102)
Hm?
Jeffrey (31:51.871)
That’s why it’s hard to build truly massive startups if you don’t have strong insight into what to build and any unique advantage. There’s big demand for technical founders building for the global frontier. They just need more confidence before jumping in.
Andreas Deptolla (32:20.366)
You mentioned connecting with top founders who are early on the most innovative things. How do you find them? How do you connect and build with them?
Jeffrey (32:36.19)
Today they’re on Twitter. Gen Z—27-year-olds—spend hours there. But in Southeast Asia, people don’t really use Twitter. I don’t know about Germany—but most folks here aren’t on Twitter. So we miss a lot. Some have good LinkedIn games, but Twitter is where the frontier conversation lives. Step one: follow them on Twitter, interact, DM them. There’s nothing wrong with cold-reaching. In the Bay Area the culture is very open—people are there to help. It’s not a zero-sum game. If you need help and they have time, they’ll help.
It’s easy for us to reach out, meet people, ask the right questions. If you attend conferences—virtual or in person—that helps. If you travel, do more of that. But honestly, younger founders are very open to Zoom coffees. Some prefer Zoom to in-person. It’s perfect for us in Singapore or around the region.
Jeffrey (34:07.242)
We can explain who we are, and it doesn’t matter we’re not in the same city. I think people are a bit more open now than before.
Andreas Deptolla (34:24.334)
So the recommendation is: build networks, travel, go to conferences, and increase peer density—put yourself where the best ideas are.
Jeffrey (34:45.978)
There’s that saying: you’re the average of your five closest people. The question for me is: how do I make the world’s best AI founders my friends? If they become my friends…
Andreas Deptolla (34:58.062)
Mhm.
Jeffrey (35:13.069)
…then the probability rises that I’ll be among the best in the game too.
Andreas Deptolla (35:27.118)
Does the same logic apply to angel investing—curating the right peer group of co-investors as part of your edge?
Jeffrey (35:57.121)
Yes—but you still need your own niche. I’d say the best investors are still in the U.S.—really, the Bay Area. I follow highly successful investors like Elad Gil and Jeff Dean—hit rates are unparalleled, almost impossible to match. I’m still learning.
I’ll never be done learning.
Andreas Deptolla (36:28.138)
What separates the most effective angel investors in AI?
Jeffrey (36:37.212)
Experience and background. If your career was steeped in STEM, tech, CS—and then you took risk, maybe joined a startup that could die—that’s already a rare personality profile. Those people attract similar people. If you came from a top-tier school doing leading AI research, your peers are the same type. When they start companies, they start serious companies. When they join startups, it’s serious startups. Over time, you either stay in that orbit—or you drift out into lifestyle businesses, real estate, whatever. Then you’re out of the flow. But if you stay in, you’re like the PayPal Mafia—you stay in the current. You could say the only real criterion is “be lucky.” If you study success, there’s always luck.
You’re born into certain circumstances, personalities, parents—so luck matters. But also the environment you choose. The best in the world are consistently trying; they seek problems; they attempt solutions. Smart people ship solutions. They learn and iterate. They invest, make mistakes, invest again. Over time they build a good “trader’s account.” They do right by founders. Founders respect them highly. Then signal compounds.
If you invest, that’s a statement. There’s a halo effect. VCs notice. That’s the nature of it. In every new sector—Internet, Web3, AI—the top 1% keeps playing.
Andreas Deptolla (39:29.185)
Hm?
Jeffrey (39:33.755)
…and the community knows who they are and gravitates toward them.
Andreas Deptolla (39:39.118)
Who’s in that top percentile right now—the people you’d recommend following and studying? Who comes to mind?
Jeffrey (39:51.675)
Top few: Elad Gil, Jeff Dean, and Sarah Guo from Conviction. Newer ones: there are several new AI funds under $30 million—many of them are very good.
Andreas Deptolla (40:18.798)
And on the VC firm side—who do you look up to? Firms that are really doing it right—maybe not comfortable to admit, but worth studying to get better?
Jeffrey (40:41.338)
By performance, the brand-name firms are still very strong at what they do. The top dozen in the Bay Area are excellent. The newer platforms—YC has resurged since Garry came back. New groups like South Park Commons, Neo, the new AI hacker houses and residencies—they’re on an interesting trajectory. Established AI firms like Radical (Canada) are good. Niche, smaller funds like Outside Capital too.
Many come from strong technical research backgrounds. There are lots. I probably follow 40 of them. It’s hard to list them all.
Andreas Deptolla (41:58.383)
I have to ask—among those 40 funds and individuals, on the European side—angels or funds—do any fit your “top percentile” category?
Jeffrey (42:12.538)
Yes—I’d mention Index, Balderton, Harry from 20VC—Harry’s different but he’s doing well now. I’d say, in German-speaking countries too, but overall I’d still point to North America first.
Andreas Deptolla (42:37.55)
You mentioned fund size a few times. In your view, is there a correlation between fund size and expected returns? Can a fund be too big or too small? How do you see it?
Jeffrey (42:58.449)
Fund size ties to the power-law distribution at the bottom. If you pick a geography with few outliers, you won’t see big winners often. In the U.S., you might see ~30–35 billion-dollar companies per year—probably higher now. In Southeast Asia, maybe one every seven years.
That’s the nature of it. Meanwhile, valuations here can be 30–40% higher than the U.S., while the U.S. market size is ~40x larger than Southeast Asia—and ~4x larger than Europe. The math isn’t “fair.”
Andreas Deptolla (43:44.888)
Ouch.
Jeffrey (43:58.649)
So fund size matters a lot in geographies that only produce a few outliers. Remember: if you can only find a few hits in a given year, and your deployment window is four years, you either get into the one—or you don’t. And if there are 40 funds chasing the same geography, it’s just mathematically tough if you don’t know what your fund size should be. Outside the Bay Area—really, outside that core—you have to size funds very carefully.
Andreas Deptolla (44:52.382)
Interesting—the correlation between market opportunity and fund size; too big or too small becomes a problem.
Jeffrey (45:05.625)
Yes—and generally smaller funds are historically… more reasonable by default.
Jeffrey (45:21.272)
But in other countries—especially Europe—it’s particularly important.
Andreas Deptolla (45:32.878)
On your personal investing—outside the fund—how do you allocate? Public markets, real estate, crypto? What are you into?
Jeffrey (45:47.436)
Puzzled?
Andreas Deptolla (45:48.852)
Yes—your personal side.
Jeffrey (45:51.312)
I keep a super-safe bucket. Then I have a sleeve for things I know—if AI is booming, I’ll lean there. If I don’t understand it, I put money into ETFs or very safe products.
Andreas Deptolla (46:18.03)
It looks like the public markets are increasingly driven by a few giants—the new Microsofts—and the S&P is dominated by those. Are we heading for a change? Where are we in the cycle?
Jeffrey (46:49.687)
Markets are cyclical—every 12–18 months you can feel a turn. It’s sector- and company-specific. If you’re not timing in and out perfectly, you’ll churn a lot. I tend to go long. There are drawdowns, but being long on quality is the important thing.
Andreas Deptolla (47:16.992)
If you look at valuation drivers in public companies—AI is often pitched as cost reduction, job displacement, efficiency. Think call centers and so on. Is the bigger opportunity on revenue—or where do you see the main value?
Jeffrey (47:46.455)
I’d say efficiency—and some disruption. I think true disruption isn’t fully understood yet. A lot is still under the hood—infrastructure, hardware—still noisy. IT budgets are there. On the infra side, things are still in flux.
You’ll see more deployable models, more deployable startups, more interesting things. Those are always moving. But overall, optimism is still very strong.
Andreas Deptolla (48:32.718)
When you say adoption isn’t quite there—can you quantify it? Do you see that companies buy AI solutions but employees don’t use them day-to-day? What exactly is the adoption problem?
Jeffrey (48:55.255)
Right now it’s more about production concerns—how governable and explainable systems are. Whether outputs are consistent with the same prompts. For bigger enterprises it’s compliance and on-prem requirements. Those solutions take time.
Andreas Deptolla (49:01.326)
Then…
Jeffrey (49:24.854)
IT departments and CIOs need to understand, adopt, and approve. There’s a lot of security and compliance work. All of that moves at its own speed. Some regions move faster, some slower, some wait and see. The solutions aren’t fully there yet. We’re still a bit in the cowboy phase—but I think that’ll settle soon.
Andreas Deptolla (49:58.145)
What you describe—compliance, security, the right tooling—sounds like opportunity for startups. How can founders attack those problems and sell to enterprises?
Jeffrey (50:15.094)
Yeah—think of the EU AI Act, for example. Then you need startups along the chain—from testing to actual compliance, process and workflow.
Andreas Deptolla (50:19.918)
Mhm.
Andreas Deptolla (50:32.974)
A couple of personal questions: in your work, what gives you the most satisfaction?
Jeffrey (50:49.845)
For me, it’s the technology and the problems founders are solving. If I find founders working on that—and I can support them—that’s it for me. If I were a builder, that’d be exciting too. I’m a VC because I’m passionate and curious—I can’t focus on just one thing. I want to learn. We invest, so we have to study again and again—it’s like going back to school. And it’s exciting. You get to do a lot. You keep feeling useful.
I think most of us are problem solvers at heart. That makes it feel very natural.
Andreas Deptolla (51:56.047)
If you could give your 25-year-old self one piece of advice, what would it be?
Jeffrey (52:11.444)
The same thing most people say: take more risk. But also surround yourself with the right people—or the people you admire. I had no idea back then…
Andreas Deptolla (52:20.01)
Interesting.
Jeffrey (52:37.332)
I thought I was taking a lot of risk in the U.S.—but there’s always more. For example, for 20 years it was a very risk-on environment. If you started a company that didn’t work, it was still fairly forgiving. Sometimes we think we’re being bold—feels uncomfortable—but there’s room to go further.
Andreas Deptolla (53:07.072)
So: find the right people, take more risk—especially early.
Jeffrey (53:14.468)
And maybe know yourself. Sometimes you don’t know what you don’t know. If you’re with the right people, they’ll give you feedback and ideas that change your trajectory. But at the same time, all advice is a bit sad—it’s easy to say and hard to live.
Jeffrey (53:45.483)
I can give you these tips—but you might do it for two months and stop. Then your life is your life. Sometimes everything’s timing. You might listen to this podcast and miss what I said. You might hear it but not internalize it—then remember in ten years and only then apply it. Generally, yes—take more risk.
Andreas Deptolla (54:23.982)
Yeah.
Andreas Deptolla (54:29.454)
Take more risk, especially early on when the opportunity cost is much lower than later in your career.
Jeffrey (54:34.548)
Yes.
Jeffrey (54:44.404)
But at that age you don’t realize it—when you’re 25, you think you’re already old.
Andreas Deptolla (54:46.926)
Yeah.
Andreas Deptolla (54:55.065)
Jeff, if you could recommend our next guest, who comes to mind? Who should we invite?
Jeffrey (55:14.076)
I think your audience would benefit from… I could give you names, but I’d say: invite a new AI-focused fund in the U.S., probably under $30 million—maybe a solo-GP fund. You should speak with one of them.
Andreas Deptolla (55:30.83)
Mhm.
Jeffrey (55:42.195)
And it doesn’t have to be only one—there are several who are quite good. The vintage for them—investing in the last two years—is perfect. These folks are forward-leaning. They’ve seen a lot. If you get even one who has a few really good companies in the portfolio, they’ll have a lot to say—and a lot to teach.
Andreas Deptolla (55:54.606)
Mhm.
Jeffrey (56:12.396)
And founders in your program would learn from them as well.
Andreas Deptolla (56:15.786)
Great—yeah, Jeff, we’d be happy to follow up and get your recommendations. Thank you so much for taking the time today, outlining what made you and Golden Gate Ventures successful. We’re looking forward to what’s next for you.
Jeffrey (56:46.515)
Thank you.