The Long Conversations with Thu Huong Le

Quest Ventures’ managing partner on his early Carousell bet, Vietnam beyond unicorns, cutting through AI hype, and what makes founders stand out.

I didn’t plan for my third guest to have a connection to my second, but here we are: B2B ecommerce startup Kamereo is a Quest Ventures portfolio company. Perhaps that says something about how tight-knit this ecosystem really is.

I first met James Tan a few years ago in Da Nang, during one of his frequent trips to Vietnam. He struck me as unusually down to earth for a Singapore VC who has backed some of the region’s best-known startups, including Carousell, Carro, and ShopBack. When I later visited Quest’s office near Block71 in Singapore, the office felt more like a startup.

True to the spirit of this newsletter, there’s no big news here. Instead, we talked about how an early-stage VC thinks: the promise of applied AI, what makes a founder stand out, and why James remains optimistic about Vietnam, even if the country hasn’t produced enough unicorns.

This interview has been edited for brevity and clarity

The promise of applied AI

I’ve heard founders complain that VCs in Southeast Asia don’t really understand AI, and that if you’re serious about building an AI company, you have to go to the US. Is that a fair criticism?

And I’ll be the first to admit that I don’t understand every sector that comes along.

I would be lying to myself and everybody else if I said I was a master in logistics, then suddenly a master in fintech and payments, then an expert in Bitcoin and ICOs, cybersecurity, robotics, and now AI.

How is that possible? Even if I keep learning, there’s no way I can become a subject-matter expert in every new sector.

In our case, we do thematic funds. But every time a new sector comes along, I need people who understand it to help me evaluate it.

At the same time, startup founders need to be able to boil down what they’re doing into easy-to-understand terms. If I’m the only one VC who can’t understand it, then that’s my fault. But if a lot of VCs can’t understand it, then I think founders need to find a way to pitch it so that all of us can understand.

And along the way, they have to pitch to VCs at multiple stages too. I always say that if you really understand a subject well, you should be able to explain it very simply.

They want money from the best possible investors they can find. So if they think Southeast Asian VCs aren’t as good as, say, Silicon Valley VCs, then why aren’t they in Silicon Valley trying to raise?

My view is: go ahead and do it. We have a few companies that are doing exactly that. If they’re able to raise from investors who understand them better, that’s okay. Capital goes wherever there’s an opportunity.

I said this at the beginning of the AI boom: Southeast Asia is not the place where you can imagine another LLM being created. I just don’t think so.

That layer will be built by the Chinese, the Americans, and others. Once you have that starting point, it naturally leads to the next wave and the next phase of development. So a lot of that development will continue to happen in the US or China because they had that starting point.

The challenge for us is figuring out what role Southeast Asia plays in AI.

My thesis is that we should think of AI like the cloud. The applied AI layer is where we see more opportunities for Southeast Asia, rather than trying to build something very broad, like another foundational model.

So you don’t see Southeast Asian AI founders moving to the US as a problem, even if that means some of the region’s best talent is leaving?

First of all, even as I encourage them to go to Silicon Valley to raise funding, we also want to get our skin in the game as early as possible, perhaps by funding them before they go over, knowing there will be suitable investors for them at the next stage, or even during our stage.

If you’re a Silicon Valley investor looking at a company from Malaysia, you might also ask: are Malaysian or Southeast Asian investors already backing you? Having a Singapore-based investor like us on board provides some validation. That fits quite nicely with the early stage that we invest in. By the time they go to the US, they may already have some proof of concept and validation behind them.

Do I consider it a loss to our ecosystem when they leave? In the short term, yes, because they’re no longer here to bounce ideas off people. But who are they bouncing ideas off with anyway here?

I’d rather they go there and level up. Learn a lot of things. See how companies are run, not just the technical side, but the operations: how you raise the next round, hire people, fire people, and eventually go for a listing.

We don’t have that history here. We don’t have enough people who have gone through that painful process.

And one day, when they’re successful, I’m sure some of that capital, or the founders themselves, will flow back into our ecosystem. I’d rather they become very, very big and successful and then come back than come back halfway through the journey.

So I’m at peace with that.

Can you elaborate on why you think Southeast Asia’s opportunity is in applied AI?

Maybe I’ll cut it into two parts. First, it will be hard for us to compete on the foundational AI side. Remember I said the ecosystem builds upon itself.

You have people who resign from OpenAI and then join another AI company, and so on. Silicon Valley and China had that early start. They also had a massive pool of data and that initial base layer. We don’t have that.

A lot of our best talent has already gone to Silicon Valley – mostly Silicon Valley rather than China – to be part of that wave. So we don’t have that ecosystem of engineers leaving one AI company in Singapore, joining another company in Singapore, and staying there.

And I’m using Singapore rather than Malaysia, Vietnam or Thailand because there simply isn’t that much of an AI ecosystem elsewhere in Southeast Asia.

Then there’s market size. How do you get to OpenAI-level revenue if you’re serving just the Singapore market? You can’t. And the countries around you are already starting to build moats around themselves, saying: We also want to create our own sovereign AI capabilities and develop AI talent within our borders rather than collaborate.

So even as Southeast Asian countries become closer geographically and politically, there are still areas where countries want to say, “We’re very strong in semiconductors. We’re very strong in this particular area. We want to continue to be strong here, and you guys better not touch us.”

We’re seeing that happen in AI. So naturally, the ecosystem continues to concentrate in China and Silicon Valley. I don’t see that changing anytime soon. I don’t see these people coming back to Southeast Asia anytime soon to create the next major foundational AI company.

There also aren’t AI VCs here that have earned their chops to the point where a founder says, “Wow, we better go to Quest because they’ve backed so many successful AI companies.” There are none. The AI ecosystem is still over there.

But applied AI is different.

The easiest way I can explain it is this: Imagine you’re already doing credit analysis and giving out loans. Today, that can still be a very manual process. But you can apply AI to credit scoring, bringing in different sources of data, analyzing them and determining the probability that someone will default. That’s applied AI.

You can use that to create the next big company, and the resources needed to build those companies are much lower than they used to be. You don’t need armies of credit analysts going through thousands of Excel sheets. Now you can pump that information through a due diligence tool built on top of whichever AI engine you want.

That’s a hypothetical example, but it gives you an idea of what we’re looking for when we talk about applying AI to real-world problems.

And these companies still need to become big. If they remain small, we’re not happy. They still need the potential to become unicorns.

You talked about finding the next big companies in applied AI. Looking back at the last generation, companies like Carousell, Carro, and ShopBack are well-known names today, but when you first invested, they obviously weren’t. How did you find these founders so early, and what did you see in them that other people didn’t?

It’s our job to look for companies before they get funding from anywhere else.

Take Carousell as an example. They were very, very hungry. They were working out of the same coworking space as I was, Block 71, which at the time was run by NUS and other partners. They were always among the last to go home. Even when I wasn’t there, people would take photos and say, “Wow, these guys work very late.”

They also had a very good mix of people on the team. They had engineers and people who were stronger on the business side. So they could actually deliver what they said they wanted to build technologically, rather than outsourcing it to some third-party agency or getting an intern to do it.

Sui Rui also spoke with a level of maturity that you didn’t normally see in someone in their twenties. It was the same with Marcus. Then when we spoke to Lucas, the technical guy, he could explain very clearly why they had moved from one technology provider to another. He understood exactly why they were making those decisions. There was a lot of clarity.

But ultimately, all that clarity and all those late nights are useless unless the market you’re going after is also big.

Back then, we were still largely using laptops to buy and sell things online, and it was very clunky. Smartphones weren’t as prevalent. Android phones were just coming out and could be buggy and slow. Uploading photos and selling something online wasn’t nearly as easy as it is today.

Carousell had a very simple way of explaining the opportunity: everything you were doing on your laptop would eventually move to your phone. You would buy and sell on your phone.

And you could see the customer journey. Maybe someone starts by selling something standardized, like a textbook. Once you have that person on the platform, perhaps the next thing is a laptop, and then something else. You could envision that progression over time.

I was still surprised when they eventually started selling big-ticket items like houses and cars. I didn’t expect that to happen so quickly. But even back in the 2010 era, you could envision the journey beginning with something as simple as textbooks.

Spotting standout founders early

So for these early-stage companies, clarity matters a lot to you. If the founders can clearly explain where the company is heading, what else are you looking for at that stage?

Signals? I’ve always wanted something called ruthlessness. I think in our world, you cannot be number two or number three.

Okay, you can be number two or number three, but definitely not number five or number 10. You should try to be number one – or at least in the top three. And to get there, it’s not easy. You’ve either got to wait for your competitors to die, or you make sure they die.

You make sure they die by hiring the best people, or by sucking up the funding – which Travis Kalanick of Uber epitomized so well. If the VC funds me, that means the same VC will not fund my competitor, right?

So you need to be ruthless that way. I think I can see that in Carousell. I can definitely see that in the ShopBack guys as well – Henry and Joel.

Lucas from Carousell recently left, and he’s part of a broader wave of founders from Southeast Asia’s consumer tech companies moving into AI. Does that tell you that these companies – or even consumer tech in Southeast Asia more broadly – have become less exciting?

I can’t speak with certainty on this, but I’m going to assume that when Lucas left, there was some sadness. But it’s not like they were all sitting in one office and suddenly there was one empty seat and only two of them were left. Every one of them would already have their own team to run and their own KPIs to meet.

From an investor’s perspective, isn’t it good that there’s a replacement and there’s no loss? The company doesn’t suddenly fall off a revenue cliff. It runs as normal. So, good.

As for founders joining the AI wave, we are getting a lot of me-too ideas – the same ideas you see in Silicon Valley and elsewhere. But I totally get why some people don’t want to miss out.

If you were in 2017 and said, “Wow, everything must be on the blockchain,” I must admit that back then I had some skepticism. Although now you’re seeing it deployed at scale in banks and elsewhere.

Someone has to take the early steps. And as early-stage investors, if you’re not willing to believe in some of these things, then I’m not sure you should be doing early-stage investing.

So I’m okay with them riding any wave. It really goes back to timing. Today, if you say you’re going to start another ecommerce company, no. But if you had talked to me 15 years ago, two decades ago, then okay, no problem. Everybody was going into ecommerce, right? Everybody was trying to do another ride-sharing company.

But what’s the trend now? That matters, because when there’s a trend, there will also be VCs raising money for Series B, C, D and so on from LPs who believe that some of these companies will grow.

Which deal did you miss that you regret the most?

Grab. I can’t remember exactly, but I remember the stage. I got the deal referral at Series A. And back then, my mindset was: I’m still hardcore focused on my business in China, and I’m doing angel investments in Southeast Asia. Angel means first check, right?

The business made sense because I was traveling quite often to Silicon Valley and using Uber. And here in Singapore, it was very obvious that one of these companies would become a winner. I just didn’t know whether it would be Grab or one of its competitors at the time, but it was definitely something I wanted to look at.

Don’t count unicorns

I have to ask you about Vietnam, obviously. Why are you so optimistic about the country when Vietnam still hasn’t produced many tech unicorns?

I feel very optimistic when I talk to young people here. Every trip I make, I make sure I have lunches with young people – people who are in their 20s, who have just graduated.

I ask them, “Why are you working here when you studied in Australia or somewhere else?” They came back because they see a lot of opportunities here, even though they’re paid less than they would be in Australia, for example. The pace of life there is very structured – nine to five, very assured. Here, every day is different.

But first of all, I think the way we measure success is not correct. We shouldn’t count the number of unicorns as the measure of whether an ecosystem is successful. Every ministry does that, yes, but can we change the narrative?

Unicorns tend to be created in very large markets. That’s why China has so many. Vietnam has 100 million people, but that’s still small compared with China. I spent nine years in China, so you can assume that almost everything I see in Singapore, the Philippines or Vietnam looks small to me. China has roughly 13 times the population.

So why do you expect a US$1 billion company to be created from a much smaller population? If I create the equivalent of a US$1 billion company in China, the equivalent here might be a US$100 million company.

So rather than measuring success by the number of unicorns, one question we should ask is: How many of our Series B and C companies are using that capital to expand overseas? No matter how big they grow in Vietnam, many of them will find it difficult to become unicorns if they remain in one market. Why aren’t they using the money they raise to expand into the Philippines, Thailand and elsewhere?

So I’m saying we should change the way we measure success. That doesn’t mean giving the ecosystem an easy way out. I’m not saying, “Let’s just count how many companies raise Series B.” I’m saying there are better ways to measure success because the unicorn is a Silicon Valley invention.

There’s another issue. We don’t have enough late-stage investors here. Once a company wants to raise Series B, where does it go? Often, Singapore. Singapore is the VC hub for Southeast Asia.

And once you’re there, you may become a Singapore-registered company so you can raise capital. Does that suddenly make you a Singapore company rather than a Vietnamese company?

To me, it’s still a Vietnamese company. The business is still here. It’s still run by Vietnamese. And Vietnamese should be proud that these founders have been able to raise more and more funding, even if that funding comes from overseas.

A lot of VCs are now calling themselves “AI-native,” using AI for everything from deal sourcing to due diligence. What do you make of that? And how are you actually using AI at Quest?

We just apply whatever we think are the best tools. Right now, we’re using AI internally for due diligence. I’ll feed the entire data room into our due diligence tool and use it to build a picture of where the company is at that point in time.

Then we have what we call a capstone tool. The question becomes: from where the company is today, how do we get it to 9x?

Why 9x? Every round that a startup raises, we hope the valuation will go up by 3x, at minimum. So if we can help them achieve that twice, that’s 9x.

By then, we’ve probably taken care of them for two to three years, and I think we’ve done our job. New investors will come in, they’ll have different board seats, and they’ll want to have their own fair share of opinions.

But no matter how much AI we use, whether it’s in due diligence or filtering, I still need the human factor. I still need to meet the person, visit the factory, and see how he interacts with the people below him, above him, and his peers.

I need to see whether he’s comfortable sitting with people, having a beer. Those things are still important, because AI isn’t going to run the company for him.

Quest has backed companies like Carousell, Carro, ShopBack, and 99.co, but the firm itself has remained relatively low profile. Is that deliberate?

Not really. We still have to put ourselves out there, but what matters is being visible to the right people.

If I’m only known to Series B founders, it’s already too late for me. I need to be talking to founders who haven’t raised a single round yet. Even Series A can be quite late for us.

So we try to meet founders very early – when they might have just an idea, when they’re at a demo day, or even when they’re still in a university lab.

That’s why I’d rather spend my time reaching these founders through universities and the early-stage ecosystem. By the time a company is prominent enough that I’m discovering it through Tech in Asia or the Financial Times, it’s probably already too late for us.

More