GenAI Fund partner and co-founder Kai Yong goes on call to share his VC journey with us, from his time at AWS to joining Insignia Ventures Academy’s Cohort 3 amidst the pandemic to then starting GenAI Fund with fellow AWS employees and the many lessons he has learned since on what it takes to help AI startups scale and fundraise.

Watch all five parts of our conversation with Kai Yong on this playlist:

About Kai Yong Kang

Kai Yong is a Partner at GenAI Fund, bringing years of experience in business development and community building within the startup ecosystem across Southeast Asia. Prior to joining the fund, he led numerous AI and startup accelerator programs at Amazon Web Services (AWS), supporting thousands of startups across Southeast Asia and Pakistan. He also played a key role in launching AWS ASEAN’s first AI and Generative AI accelerator. With deep expertise in driving startup growth and ecosystem development, Kai is a prominent advocate for AI adoption and scaling in the region.

Directed by Paulo Joquiño

Produced by Paulo Joquiño

Follow us on LinkedIn for more updates

The content of this podcast is for informational purposes only, should not be taken as legal, tax, or business advice or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any ⁠⁠⁠⁠⁠⁠Insignia Ventures⁠⁠⁠⁠⁠⁠ fund. Any and all opinions shared in this episode are solely personal thoughts and reflections of the guest and the host.

Transcript

Part 1: Joining Insignia Ventures Academy and co-founding GenAI Fund with AWS colleagues 

Paulo: Hi, I am Kai from GenAI Fund. We are an AI-focused fund in Southeast Asia.

Thanks so much for coming on. It’s great to finally meet you in person. Thanks for braving the Filipino heat. I don’t know how this compares to Ho Chi Minh heat, but yeah, thanks for doing this with me. It was great to see you at TechShake’s Ignite event.

I think during your panel you also mentioned that the last time you were here was also for that event six years ago. We talked off the record about your experiences and culture shock coming to the Philippines, which is great. But for this podcast, I wanted to focus a little bit more on your experience coming into Insignia Ventures Academy as you’re an alumnus from one of the previous cohorts. 

So maybe we can start with that. How did you find out about Insignia Ventures Academy? I think at the time you were at AWS. What made you decide to join the program?

Kai Yong: Yeah, thanks Paulo, for having me here. It’s good to be back in Manila and shooting this with this very interesting weather. I’m sure we’ll survive that as hustlers in the startup ecosystem.

Yeah, so in terms of the journey with IVA, I think that’s a very interesting one because it was probably four or five years ago during COVID time when everyone had so much time to kill. I’m one of those guys, right? Like AWS work is not keeping me busy enough that I wanted to—I’m not sure AWS should hear this, but joking aside—your work is really busy, right? During COVID time, everyone has to do double or triple workload.

But it’s just that you still have a lot of free time and you wanted to really pick up new skills and things like that.

So that’s where I started thinking about what is the best way to upskill myself while keeping it relevant to the kind of industry and ecosystem I want to be more specialized in.

So actually, the starting point was that my ex-manager at AWS, Digi—he was actually in the very first cohort with IVA. So that’s how I knew about the program initially. Knowing him, I know he is pretty selective in terms of how and where he spends his time. This is where I started feeling that, okay, maybe this is something that I should definitely check out. 

So this is how I got inspired, and then I started talking to the program manager. I think back then it was Gail, and trying to understand the curriculum, I saw that a lot of the elements and components in the curriculum are very hands-on, which I don’t see in many other programs that I was trying to enroll in. So that’s why I decided that okay, let’s give it a try and do something more fun and hands-on.

Paulo: So you went for it and you joined. I think Digi was part of that first cohort, and since then we’ve had a lot of great folks from AWS like yourself join a little mafia, so to speak, there. And coming out of IVA, but specifically your experience during the 12 weeks, what was the most memorable part of that whole 12 weeks for you?

Kai Yong: I have to say that I was very lucky because during COVID time you are stuck at home and you can’t make a lot of new friends as well. And I was quite lucky to be put in the same team with four other teammates we still keep in touch with until today. 

And I think the most interesting thing is that every one of us treated this very seriously. Because being in a program like IVA, it’s up to you to maximize the value that you want to get out of this program. It’s adult learning outside working hours and everyone is busy as well. 

The very fun thing is that all of us in the team shared the same interest that, hey, you know we like startups. We want to learn more about the industry. And we want to see how we can eventually become an investor ourselves so that we can add more value to the startups.

That kind of environment, that kind of vibe, I think is a very fun learning process for me. On top of the practical experience, like for example, every week we came together behaving like a real VC in some way — “Hey, where’s the pipeline and what have you done and which are the startups that you guys want to pitch for the IC this week?” 

That kind of teamwork, that kind of real-life simulation where everyone treats it very seriously, I think that’s itself is a fun experience for me. And looking back, it made me feel like I did try my best to spend my time on really meaningful things during COVID time.

Paulo: I think I remember you guys as being one of the most active teams. And was there a particular takeaway from that whole experience? I’m sure there were a lot of sessions. There’s also that practical experience as well. What was a particular takeaway that sticks with you until today?

Kai Yong: I would say that the theme for me is still very much centered around the team. Because whether it’s IVA or whether it’s GenAI Fund for what I’m doing now, it’s really the people that make you feel that you want to do this and you want to do even more. 

Because it really comes down to a lot of self-discipline. So I would say that I think quite a bit of it is luck, but in the way we, as a team, we actually try to set up very tight cadences. Even beyond just the IVA curriculum, and for example we are so serious to the point that a few of us actually decided to co-invest in a few startups together, right?

And one of the startups actually got acquired in just a few months after we did the deal together. And then we also did another deal, which is a YC-backed startup. So I’d say that we couldn’t have happened without having the right teammates and committing to this process and eventually making it something real so that we really walk away with a real-life experience.

Paulo: And then I guess coming out of the IVA program you were still at AWS, but I remember you moved from KL to Vietnam. What was behind that position? And then maybe fast forward a little bit, how did you end up joining forces with your fellow AWS folks from around the region to start GenAI Fund?

Kai Yong: The move to Vietnam is both personal and business at the same time. Personally, I think it’s really that kind of COVID experience where you are stationed in one place for years without moving around. Yeah, they change the way that you want to experience life. And so I’ve been thinking about it during COVID time that I wanted to make a move. 

And I was thinking very hard about where it is going to be, because for me, I really want to stay in Southeast Asia, right? I know a lot of people want to go to the US, they want to go to Silicon Valley, they want to go to different more advanced economies where the startup ecosystem may be a bit more established. 

But ASEAN has always been where my heart is. So if it’s not Malaysia, then where should I go? And for me, the way I think about it is that I want to be on a rocket ship.

When I look at the different markets, I think Vietnam has that kind of potential that hasn’t been unleashed. And because I had to choose between Indonesia and Vietnam, oh I see. And for me, I felt like Indonesia is at a peak. And for me to bring certain value, I think it’s not going to be super meaningful in that sense.

Vietnam is still on the rise and there’s so much that I can do. So that’s why I decided that, okay let’s go to Vietnam. I like the people—they are so hustling and maybe they could help me to even find a better version of myself by pushing myself even harder, right? Yeah. So that’s the reasoning behind the move to Vietnam.

Then after three years in Vietnam, somehow there are a lot of AI initiatives that I have been piloting in Vietnam, mainly because the technical talent in Vietnam is just so quick to respond and take on the opportunity to ride on the AI wave. Even before ChatGPT, there were really a lot of very technical builders playing around with AI. 

Then from AWS perspective, obviously AI is important. I run different AI-focused programs for AWS starting from 2023. So even before GenAI Fund, I’ve been getting my hands dirty with running AI-focused programs in Southeast Asia, right? So it just comes to a point where I started feeling that this is such an opportunity that you have to build something out of it. It’s just great timing that so happened that the other two colleagues, ex-colleagues of mine, Laura and Danny, are also available, right? 

So I think the timing is just so nice that GenAI Fund just happened and there’s so much opportunity that’s untapped, and at the same time, I have two ex-colleagues who share so many similar working traits and culture that it is just so easy for us to work together and understand each other personally and professionally. And we just decided that, hey, it’s such a fun time. There’s a lot of uncertainty. Maybe we should build something together.

So it is very different from the usual kind of startup way where you find a problem and fix it. But for me, because it’s a lot focused on the team, you know it just so happened that a few great people with similar culture and synergy came together and decided that, hey, let’s do something and maybe a GenAI Fund, maybe it’s something else. But somehow we landed on the idea of GenAI Fund.

Part 2: GenAI Fund partner Kai Yong Kang on how 3 AWS employees founded an AI fund in Southeast Asia 

Paulo: You talked about GenAI Fund, but one interesting thing you mentioned is that you just wanted to work with these colleagues of yours, but you weren’t sure whether you were going to start a VC firm or fund, or what exactly you were going to do. Maybe you can walk us through that process of eventually deciding to start a fund.

Kai Yong: So we actually started with the idea of building a fund because we already have angel investors that are committed to investing into our fund. And we know that we want to do AI. But the thing is, we don’t know where to start. 

The thesis, the strategy of deployment, what we’re going to do about it, the exit strategy and so on. And that’s why we decided that, okay let’s just get into action and let’s write our own thesis. And somehow we decided to turn that into a report.

So that is why we launched the first Southeast Asia GenAI Startup Report last year together with AWS because we just felt that there’s a lack of understanding for us and also the market as well. 

If you talk about Gen AI, maybe one or two years ago, everyone’s talking about the US, China, Europe, maybe India as well, but there’s very little context and understanding of what is actually happening in Southeast Asia and what is the thinking around what should be done.

So that’s why we felt that, okay let’s do this—let’s do a public service, but at the same time, it helps us to build our conviction as well. So that’s how we started writing the report and then did a roadshow in six ASEAN countries trying to talk to not just startups, but also government stakeholders and enterprises.

Just trying to get a better view of where the opportunities are and where we should focus our investment and timing in.

In fact, I mean there are a lot of things that have been said in the report. It’s probably 90 plus pages. But I think that there are some things that we said that still remain very relevant until today, even though AI changed so, so fast. Yeah. But we’re glad that some of the predictions that we caught are still very much relevant today.

Paulo: Can you give us one and then maybe that as a teaser and then the rest they can check out in the report. 

Kai Yong: So I think that in the process of trying to understand the ecosystem, we started with talking to a lot of AI startups. We probably spoke to a hundred or more startups in true conversation, then the rest through a survey of a few hundred AI startups across the region.

Obviously we asked a lot of questions, right? But coming from AWS, trying to do the working backward stuff and being customer-obsessed, we asked a lot about what the startups are doing and at the same time what kind of help they need.

So interestingly, and not so surprisingly, 92% of AI startups are B2B focused. And I think that it’s still the same until today based on the startups that we’re tracking. So that’s a huge number. 

You and I, we spend quite a bit of time in the Southeast Asia ecosystem. It’s always been B2C for Southeast Asia, so the B2B shift is very interesting for us. And the number one help they need is actually go-to-market. It’s not even funding, even though we are going through a funding winter. 

I’m sure they need money but just in terms of the priority, go-to-market with big and large corporates is actually something that they really need help with.

So interestingly, as we do the roadshow in the six different countries, what we also noticed is that the people who downloaded our report, the people who came to our event, about 30% of these people are coming from enterprises, right? And they started asking questions like, “Hey, I saw this startup in your report. Can you introduce us to this startup?” Or, “Hey, tell me more about this startup in this industry. Yeah. I’m very curious.”

Coming from AWS and spending about six and a half years in the startup ecosystem in Southeast Asia that has not really happened before. Because the enterprises will never entertain early-stage startups. If you’re not Series B and above, they have no interest in talking to you. 

But things somehow changed when GenAI took off, where the enterprises realized that they don’t move as fast as the startups and because this is so new and they fear being left out, there’s a lot of FOMO. Being left out and missing the chance and being disrupted. And that kind of increasing appetite of working with AI startups in a very early stage spiked up.

And because of that, it insights—it actually determined a lot of the things that we are doing right now. And that is one of the reasons why we are still consistently doing a lot of the enterprise-startup matchmaking initiatives beyond investment. Why are we still doing all these things right now? It is because of this conviction, right? And it remains to be true until today.

Part 3: GenAI Fund founding partner Kai Yong Kang on building AI centers of excellence for enterprise 

Paulo: So you talk about corporates a lot and that being something you hadn’t seen a lot of before up until this point when it came to AI. Like how enterprises—obviously one shift is them looking more at early-stage startups, but how have their expectations and I guess their considerations when it comes to partnering with startups evolved because of AI over the last few years that you’ve been running?

Kai Yong: So maybe I can give you some context of what we have been doing so far. Yeah. And what we have learned. So after the report, we decided that let’s try it out by putting this into a structured program where we actually match the enterprises with AI startups. 

The whole goal of this match is to drive a POC opportunity, prove a concept where the enterprise can test and validate with the startups to see if this can drive certain tangible business outcomes for their business. 

So we started doing that late last year. And then fast forward to today, we’ve probably done 300 matches. Wow. Between 70 plus enterprises and 2,300 AI startups in our database and network. So that is the scale. And it’s really proven that the interest is there.

Along the way, there are many things that we learned from this process. It’s very clear that there are a lot of manual things that AI cannot do in this context. So I think one is that the enterprise AI interests continue to explode. 

That’s something that we are seeing because as we run these programs, the network effect continues to grow, where enterprises come to us and we have a lot of these pipeline requests to the extent that we have to build an AI platform out of this process. So you need to manage all of that. 

So actually we just did a beta release of this AI platform about two months ago, where it’s like Tinder, for enterprises and startups for AI adoption. So basically we make it so easy that the enterprises can join our platform for free. They can post the use cases that they have on our platform, and the moment they post it, they can click a button called “AI Match.” They click it and then instantly we’ll recommend at least 10 to 20 startups to choose from our database.

Then if there’s an interest for them to work with them, it’s literally like Tinder—they click “love” or they just swipe and then we will help facilitate the next steps for the enterprise and the startup. So I think that goes to show you that why we are doing so much is because the interest is real. So that’s one.

The second is that you can see that the enterprises are really trying hard to change the way they approach things. For example, we started seeing a lot of patience and co-creation for success, right? So in the past, for a startup to actually ask enterprises to provide anything like data or budget for experimentation, it would have been almost impossible. 

But because of AI and the need for them to work on something that’s going to make or break it for them, there are a lot of exceptions that are being made right now. So maybe overall from an organizational level, there are still a lot of the usual procurement process, the usual compliance that they have to go through. But you start seeing enterprises changing and tweaking and trying to come up with their own special mechanism, right?

Like for example, there are enterprises who started setting up their own innovation funding, right? Which is a pocket that they can tap into very quickly without very little process and policy involved so they can actually run experimentation very quickly with startups.

Then second is that you start seeing people talking about having dedicated AI task forces. Which brings me to something that we have been talking to a lot of enterprises about, which is the concept of an AI center of excellence.

So there is a rise of this need where the enterprises realize that to actually have a very good AI strategy, it cannot be a siloed effort. Especially when you look at the enterprises and the conglomerates in Southeast Asia, they have so many different businesses, right? Like in the Philippines, I don’t need to tell you about the families and the kind of different businesses that they own from banks to telco to very different industries, whatever industry they have. In fact, right now, a lot of the AI efforts are siloed. 

Just a few days ago when I was here in Manila, I spoke to one enterprise and they were telling us that we have 150 AI engineers for one business unit, but I don’t know about the rest, right? So everyone has their own effort and it’s very siloed.

But the thing about AI is that it has to come from data, right? It has to come from an infrastructure that is ready for AI to actually then provide the adjacent experience to automate whatever that you are looking for. And it cannot be having data somewhere here and there and then have a very different experience, and the potential will be limited. 

That is why the AI center of excellence, AI COE, has been very interesting to a lot of enterprises because they are looking to build a dedicated AI task force that is centralized at the group level where they will run experimentations with different groups they use, but eventually they’ll bring these best practices to all the different business units and eventually help everyone to innovate using AI.

Paulo: We are seeing that they could even be like a resource allocator for all of these, like AI needs in terms of managing the vendors and managing costs. Do you see that already happening? Because obviously I think one pain point is getting so many vendors and nobody knows which vendor they got or which vendor, and then they end up racking up all these AI costs, right?

Kai Yong: That is definitely part of the AI COE. Because when we look at how to build this AI COE, it comes into many different components, right? So you have the talent part where you have to upskill everyone and then you have the data and infrastructure part where you need to make sure that this is ready for scale.Then you have the solution. You have the partners that will be involved in this process. So it’s a very holistic approach.

Paulo: I’m curious to know, you mentioned the platform, you mentioned the COE. Where does the fund come into play in particular? The investing part for GenAI Fund?

Kai Yong: When we run the process of enterprise and startup matchmaking, I think maybe it’s important for me to break it down like how we have done it so far. It’s a repeating process for us already. So we always start with enterprises first because what we believe is that we do not want to just invest and wait for the opportunity. We want to create the market opportunity actively. So we always go to the enterprises first.

So though I don’t mean to brag, some people call us this way, that, hey, it sounds like you are doing a McKinsey for AI. Because there’s a lot of time. There’s a lot of handholding. So we need to start educating enterprises about what AI is. What can you do with AI? What is the ROI for AI? And helping them to craft out their AI strategy. 

Then from there, distill it down to what are the key use cases that they want to implement. And not to be too ambitious, but just for this quarter, how many use cases—like three, three use cases that you want to implement? What is the ROI for that?

So that’s where we always start first. Then once we have the use cases, that’s where they can post it on our platform. We will use AI to recommend startups. Startups can apply on the platform as well to solve the problems.

If there is an interest this way, we will facilitate matchmaking and usually that will be done in person, right? So we partner with a lot of different hyperscalers, like AWS, Google Cloud, and also Nvidia. 

The idea is that because virtual meetings don’t usually give you that very personal touch, right? We try to do a little bit of programming around it, so that we will invite the startups. Usually what we do is that we will focus on one particular market. 

So a few weeks ago, we just did it in Malaysia, right? So we have 10 enterprises that are looking for AI solutions, and what happened is that we have selected about 25 AI startups for them. These startups come from different places. Some of them are in Malaysia, some of them are outside of ASEAN, and some of them are in ASEAN. 

What we do is that we invite these startups flying to Malaysia and they’ll actually meet each other in person. So it’s just like what we are doing now, is obviously with an in-person kind of meeting, where the enterprise invites the different stakeholders and the startups will then present to the enterprise. “Hey, this is the proposal. If we were to go into a POC in the next 12 weeks, this is the plan, right?” So they go deep into it.

Even before the one-on-one meeting, we actually set up a lot of briefing calls, right? So the briefing call is meant to help the startups to understand the enterprise more and prepare for the one-on-one meeting because the one-on-one meeting is usually 20 minutes. It’s very short. 

We want to make sure that the startups and enterprises can really maximize the value out of it. So once they finish the matchmaking, and if the enterprise decides that, okay, I like this startup. I want to work with this startup. I want to go into a POC now, and by the way, the POC is paid—we don’t do free POCs. That’s going to kill startups.

So the moment the enterprise decides that, okay, I want to work with this startup, we will then onboard them into our AI program with Nvidia, right? The whole point of the AI program with Nvidia is that in 12 weeks time, this enterprise and startup will join our program.

They will launch their POC into production in 12 weeks time and then sign a contract. That is our goal, right? That’s really speeding up the whole enterprise buying process and the startup selling process for success. 

Where we come in from an investor perspective is that the moment the deal is sealed and they close the deal, we’ll invest. We are actually going to announce the cohort one of this AI program by next week. So we have selected about seven startups and all these seven startups are really working on enterprise deals. 

That they’re working on. From our perspective, that is our conviction, right? We want to see real traction, but at the same time we are not going to wait for it. We want to proactively help startups to get it, and if they prove themselves in that process, we will double down our effort by investing in them.

Part 4: GenAI Fund founding partner Kai Yong Kang on building a customer first venture capital firm 

Paulo: Thanks for bringing us through that whole process. I think it’s quite interesting. It’s not like your typical fund. It’s a lot more programmatic which allows for more startups to try their hand at it and a lot more opportunities for founders, especially those who are B2B and then plugging into this whole AI ecosystem as well.

I was curious to know because you’re working with so many enterprises, how has that impacted, I don’t know if you could share this, but like how has that impacted your LP strategy, right? Does this lock you out, because you don’t want to be too biased towards certain enterprises or because you’re trying to help as many as you can. 

How has that impacted your LP strategy or how you thought about it? 

Kai Yong: So I think that’s a very interesting question. In fact I think it’s very aligned with our LP strategy because in our report, we actually set up the three key things that we want to do. 

One is that we want to invest in a lot more AI startups in the region to drive the growth. The second thing we want to do is the go-to-market part, which we are already doing. The third part is actually the exit part. 

Because of the first two things that we’re doing, and especially the alignment of doing a lot of GTM for startups with enterprises, it actually is very aligned with the exit part where some of these enterprises, which is really happening, are looking at core investing into the startups that are working with them.

Reason being, a lot of these AI startups are dealing with a lot of data that the enterprises have, right. The enterprise is also thinking that, “I’m giving so much of our valuable data to these startups to fine-tune, to build our workflows and we should definitely make them part of our ecosystem for the long run for strategic reasons.” Actually that’s exactly what we want to see. 

We never intentionally asked the enterprises about, “Hey, do you want to invest in the startups?” Because we understand that it’s just too early for them. Because it’s always a show-and-tell. They have to see AI that’s really working. It’s not just like using PowerPoint decks or something like that.

So a lot of them really need to see that this AI can help certain workflows that are end-to-end. But it’s not a surface-level kind of GPT. And so our thinking has always been that, let’s prove that this works. Let’s prove that an AI startup is the right strategy for you. 

Because from an enterprise perspective, they can totally build using their own AI team. They can outsource to their IT vendor. Hyperscalers always provide free engineers to enterprises to build prototypes. They have a lot of options.

So when we work with them, we make it very clear that, hey, we are not going to convince you to stop everything that you’re doing. But all we ask is that be open-minded to working with AI startups as one of the ways to supercharge your innovation. 

We’re starting to see this taking off. That they are warming up with the idea that, “Maybe we should do a strategic investment in the startups.” And eventually they have a strategy to acquire these startups as well. So that’s very aligned with what we’re thinking.

I don’t think we worry so much about, “If we pick one side and then it’s going to make it very hard for us,” but it’s more I think it’s a FOMO right now. Whoever takes the first step we will be very open to structuring this kind of engagement with them and work with them to double down on the AI strategy. 

Paulo: I’m wondering what your view is on what will define, because obviously right now you’re trying to focus on maximizing the players in the space, especially from the startup point of view, and maximizing the amount of connection and productivity with the enterprise and startups. 

But obviously, I guess from a venture mindset, you want there to be like a winner so to speak. What do you think will define that winner? And obviously like with AI, like pooling or apps, there’s a lot of overlap sometimes. One HR app starts with this feature, but then all of them eventually will overlap at some point. So what do you think will define a winner in this space?

Kai Yong: I think that we don’t think too much about who is going to be the ultimate winner because to go to your point, eventually there’ll be overlaps in some way or another. In fact there’s a lot of synergy between the different AI startups. 

Because we always start with the demand and we just go where the demand is. Eventually, what we believe where we’ll land is that we will start building a suite of AI solutions that kind of have certain synergies, like it could be we probably have five AI startups that are just very focused on banking and there’ll be one that focuses a lot on say the investment part. There’ll be one that’s very focused on the customer experience part. 

So I think regardless of the overlaps, they’ll have their own specialization. In fact, it’s something that we are also trying to put together now that we want to build that consortium of AI solutions.

So that the toolkit—when we go to an enterprise, we already have this package deal, we already have this package of consortium that, “Hey, these are all the different AI startups and you can work with them right in different areas. And these are the specializations.”

Coming from AWS and the cloud background gives us the opportunity to think about how to make that work as well. Because in AWS we also have this kind of partnerships program where we partner with the B2B SaaS solutions to sell to the AWS customers. 

So I think we are trying to take a bit of a leaf out of that playbook. How can we do the same thing so that we can continue to snowball these efforts that we are doing. We are not too worried about who’s going to win because one is just too early to tell right now. Second is that the demand—there’s so much. There’s so much demand that we need more. In fact, we need more AI solutions.

Paulo: Quite interesting how, I guess maybe this effect of coming from AWS but having that customer-first mindset I guess as opposed to, obviously you’re still thinking about LP, you’re still thinking about the founders, but have, but starting with the customer, the end customer first has impacted the way you’ve set up this fund and set up its operations and all that.

Part 5: GenAI Fund founding partner Kai Yong Kang on rewriting the venture playbook in Southeast Asia 

Paulo: To tie it back to IVA, having gone through the 12 weeks and obviously the program is—maybe we don’t claim to be like the be-all and end-all of VC knowledge, right? It’s just our Insignia view. Like how is that kind of program influenced or impacted as well, the way you’ve done this GenAI Fund?

Kai Yong: I think that I just love the “do it” mentality. Like when we join the program and it’s just a lot less bullshit. Like obviously there will be the usual workshops and sharing, which I think is still really helpful. But ultimately it really comes down to the action, and having someone like Yinglan who is going to sit on your review and IC, pushing you and challenging you. And that kind of stress—it’s not enjoyable, but you also feel the heat that yeah, you have to perform.

You do not want to sound stupid in front of all the other people who are also very smart as well. So I think that kind of “do it” mentality, that kind of environment is still it still stays with me until today that for a lot of the things that we do at GenAI Fund now, as painful as it is, we just always go with it and say, “Hey, let’s just do it and make it happen.”

For example, there are so many things that we do not necessarily need to do. There are so many things that are outside of our scope, for example spending a lot of time with enterprises trying to educate them about AI startups in so many ways that we could have avoided and just choose to look for startups and do the usual VC thing. “Hey, talk to a hundred AI startups and see where that lands us.” 

But I can tell you that 30% of my time is spent with enterprises. But we just think that it’s very important. With that kind of relationship, with that kind of synergy that we build with enterprises, we bring so much value to the startups, to the extent that one of the startups, who used to be an ex-founder who had exited in the US, and this is coming from the founder. So he was saying that actually the value that you bring to us is more helpful than YC. So that’s a very strong validation for us.

They give us a lot of motivation to do all this work with the enterprises, making sure that we can bring the startup closer to the deal. So like I said I think it’s a lot of unusual work in the context of VCs. 

But I think that in Southeast Asia and to be frank, none of the VC playbooks have really worked out in Southeast Asia, and we have to try different things. In a way we are like a startup as well, trying to prove that AI works in Southeast Asia, they are good AI startups in Southeast Asia. There’s something proprietary here that belongs to Southeast Asia and so there are a lot of things to prove. So we are trying to be a bit more unconventional. So maybe we get a different result or maybe we fail, but we tried.

Paulo: I was also curious to know since we talked a lot about how AI itself has impacted the way we’ve built this firm, and obviously zooming out, it’s impacting the way VCs are thinking about funding. 

We read a lot of founder insights about how there may be we don’t need to raise as much anymore. Maybe because of AI, we’re not that dependent anymore on VC funding or the milestones are a little bit larger right? Or faster cycles. How do you see GenAI impacting the way startups fundraise or think about cost structures and profitability?

Kai Yong: I think there are three things that are interesting here. A lot of this is based on our observation and learning from the report as well. One is that when we did the survey and when we were keeping track of the AI startups, we also looked at what stage they’re in and what kind of business stage they’re in as well. And almost 40 to 50% of the AI startups are bootstrapped. And if you zoom in on that, a lot of them are second-time or exit founders. 

So in fact in our current cohort, the first cohort of our AI program, most of the founders are probably second-time founders. So that kind of tells you that yes, money is important. But probably they have enough to start, right? And they have enough to bootstrap for quite a while to see certain tangible outcomes before they decide, let’s raise a big round.

I think that from that perspective, that’s also why we have better conviction in the market that we are seeing better founders that are trying to do something with AI. And then the other two things that I will mention is that these are probably more coming from the challenges of AI startups, especially in Southeast Asia. 

One is that coming to the compute and GPU costs, it’s yes and no. When it comes to it, is it really expensive? Is it going to be burning their pockets? It’s yes, because when it comes to scaling, that’s where the cost will be exponentially high. 

No, because in the very early beginning you don’t need as much. There’s no need for a lot of GPU costs, especially for POCs and testing. And most of the startups here are at the application layers. They don’t really need a lot of crazy GPUs. 

One thing that actually we try to help, especially for the scaling part, is that you’ll notice that a lot of the programs that we run, especially the enterprise and startup matchmaking, they’re all done in partnership with hyperscalers. 

That is why we can really help the startups to lower down a lot of their costs. Because coming from a hyperscaler before myself, I completely know their strategy and why and how and what makes them tick. 

So there are a lot of ways that we can have partnerships with the hyperscalers to actually bring down the cost of innovation for startups and help them to scale from there. So I think I would say that the partnership with hyperscalers is really important.

The second part is that because a lot of them are B2B focused, that means that the sales cycle is going to be longer. So like it or not, as much as the enterprises want to innovate, they are still slow. There are still a lot of processes to get through and that’s where we exist. 

As GenAI Fund, we want to be positioned as a go-to-market partner with these startups so that when they work with enterprises, maybe it was supposed to be six months, but if we can get it down to two or three months, I would say that’s a win for us to get them to get a foot into the door and at least you’ll secure something as a quick win.

Then continue building a long-term collaboration with enterprises. Obviously it may not be us. There might be other distribution channels that can really reduce a lot of the friction for the startups who go to market, especially in this age right now. I think there are a lot of distribution channels and some startups are so good at finding them. That’s where I think that they can significantly lower down their burn and get to some tangible results very quickly.

Paulo: I wanted to wrap up with one more question and like any advice for—we’re coming to our 10th cohort for the academy. It’s been quite a while since you joined the program. Any advice for folks who I guess were in your position prior to joining the program? How could they best take advantage of IVA for their own personal career? 

Kai Yong: I’ll say that investing in the startup industry is very different from investing in general. At least for me, I think because startups are meant to build something that is very disruptive, that five years from now or even in the case of AI, two years from now, you’ll see something dramatically changed because of this particular company that seems so small in the very early beginning. 

So having said that means building a startup is very tough. And the founders, the team they hustle so much. They just hustle so much. So what I would like to see more is investors who are more empathetic, right? That they truly understand what it takes to build a startup. Maybe most of the investors don’t come from an operator background or startup background. Yeah. But I think that empathy is just so important right now. 

Paulo: How do you build that empathy if you don’t have the experience necessarily?

Kai Yong: That’s a very good question. So I would say that you should treat yourself as a startup as well. When you think about VCs, when you think about investment firms, maybe if you’re not Sequoia or the very large firms, most of the time it’s very small, probably with 10, 20 people in a firm. 

It’s exactly the size of a startup where there’s a lot of room for innovation. There’s a lot of room for experimentation. And why limit yourself to doing the same thing again and again, where you know to talk to a hundred startups every month and then try to go into events and network.

I do think that investors have the responsibility to be innovative themselves, right? And throughout the process and journey, then they’ll realize that, oh, actually, this is how it felt. I built something, I tried something. I ran an experiment. You fail and also, this is maybe how it felt like for the startup. 

So investors have to be builders in some way. And especially right now, you can buy code anything. You can do so much with AI and there’s no excuse for you to not build anything and then claim that you understand startups, but you don’t.