This is a transcript of an AMA I did on 7 May ‘25, for the SeedToScale team (part of Accel India, runs seedtoscale.com). It was themed around unpacking the Indus Valley Report 2025 but it did go beyond the same. I have edited it mildly to make it more readable. If you see SPW, remember it stands for Sajith Pai’s Words. I have used SPW wherever I have added my comments on the piece, as distinct from the transcript of my speech.]

The AMA is here.

Intro to my AMA

Hey, this is Sajith Pai. I’m a VC at Blume Ventures, an early stage venture fund. Every year we publish the Indus Valley Annual Report. This year with my colleagues Anurag Pagaria, Nachammai Savithiri and Dhruv Trehan, we published the 2025 Indus Valley Report. 180 plus pages, lots of charts, lots of data, lots of insights and an overwhelming response. So this year, the wonderful SeedToScale team wanted us to kind of take the audience through some of the insights from this and they decided to do it through the format of an AMA, right? 

So what I’m going to do over the next 30 minutes is really respond to some of the questions that came following the announcement of the AMA. There are  about 40 plus questions, can’t take them all. So what we’re going to do is to take some of these questions, 7, 8, 10 questions, really over the next 30 minutes, try and give my perspective on the Indus Valley Annual Report and the questions. All right, let’s take the first question. 

Q1 / The AI investment dilemma

So the question is, with the pace of AI innovation, agents becoming no code in months and SaaS modes dissolving overnight, how do you assess investability in such a volatile landscape when billion dollar ideas can now be cloned by five member engineering teams? So what truly becomes defensible beyond just the founder? It’s a great question, right? When you can spin off, vibe code like an entire CRM into existence, right, overnight, then what really is a moat? And it’s a great question and we’ve been wrestling with this. And when we look at, for example, a system, there is creation, then there is, for example, maintenance and improvement, and then there is distribution. So in any system, when creation becomes very easy, the constraint point now moves to the other elements, right, which is maintenance and distribution.

So when creation becomes easy, when anyone can create, right, then you have to look at what is the effort needed in maintaining it? Who’s going to maintain it, right? And that’s where the energy goes. Who’s going to distribute it, right? So the person who has distribution power becomes central to this. So secondly, now when anyone can create overnight, the bar for creation goes up.

And I remember this wonderful anecdote from Chris Dixon in a podcast or so, where he said that a good CS student from Stanford or Berkeley or any of the top universities can spin up a self-driving car app in like, you know, which is 90% good in two days over a weekend, but 90% ain’t good enough for the streets, right? It’ll take the same student two years to get to 95%, still ain’t good enough. And it’ll take him 20 man years to get to 98%, which still won’t be allowed in the streets, right? So all the values now move to the edges. So when creation becomes easy, all the edge cases begin to matter now, more than ever.

So the defensibility comes from understanding the workflows of your customers, being able to understand what really are the pain points and map out what the challenges are and build great UI to kind of solve those edge cases. Then maintenance, okay? Someone needs a throat to choke and that is you as a SaaS vendor, right? That’s not going to go away. And then finally, distribution.

Can you kind of distribute this more than anyone can create? Okay, so that is sort of the way I think about this, and that is sort of the way other VCs are also thinking about this. If creation becomes easy, the value moves to edge cases, the bar in creation moves up, and the constraint points now move to maintenance and distribution. 

So the second question now.

Q2 / Quick Commerce

You highlight the rise of Quick Commerce in India. Given the low margin and high burn model, do you see this as a long-term shift or a temporary frenzy driven by convenience? 

In the Indus Valley report, we devoted a fair number of pages to demystify quick commerce and to look at Quick Commerce in India, its growth. And we also kind of questioned a few aspects, especially the bullishness around Quick Commerce and the belief that this is going to keep growing, growing, growing. This chapter was a particularly hard one to write because firstly, we wanted to get all the latest data, and while there was no lack of data, but really we had to parse it all together.

And then secondly, we had certain conclusions to present and we want to make sure that all of them are right. And those conclusions lay in the fact that while Quick Commerce is undeniably growing in India, right, if we take FY25 end, there were about 26 million non-unique users if you add up all the monthly transacting users across all the apps. But if you take only uniques, perhaps there were 20 million unique users.

That’s not small, it’s fairly large and that represents a majority of the India1 or consuming segment in India. So Quick Commerce over a period of just three years has rapidly made kind of inroads into all of consuming India. People love the convenience and Quick Commerce ain’t stopping at grocery, it’s gone into like, you know, apparel now, it’s gone into sports goods and what have you.

Where does it stop? And our view is that while Quick Commerce is solving a demand problem, it’s also solving a very interesting supply side problem for VCs. That is, VCs have cash and they want to deploy it. Where do they deploy it? Quick commerce and AI are the two areas where you’re seeing rapid deployment.

And it’s interesting that if you bring a Quick Commerce transaction to evaluate, there is interest from our IC, there is interest from all ICs, they want to kind of look at it more deeply. So it’s fundamentally solving, we also feel a deployment challenge for the VC; there’s so much money going into it, because it’s easy to deploy there.

It’s like the old saying that no one gets sacked for buying IBM. So some of that is also happening in Quick Commerce. Secondly, I think we are rapidly beginning to hit the, I would say the limits of consuming India now.

There are only 5% of PIN codes in India, which have more than five organized stores. And when you hit those 5% of PIN codes, which also have sufficient density, there are only about 900 PIN codes in India. And Quick Commerce is already present in a vast majority of them.

So I would say, while I continue to believe that Quick Commerce will grow, some of the pace will slacken. And especially if Zepto gets to IPO fast, they’ll have to be a little more careful on how much they can spend. So all of that is kind of going to pull down the growth velocity in quick commerce. 

Is it going to stop growing? No, definitely not. It’s going to stop growing a little slower, maybe. That’s how we see Quick Commerce.

[SPW: This piece was recorded on 7 May’25; since then we have already started seeing a mild slowdown in the pace of Dark Store openings from Zepto, as well as a push towards sounder unit economics from all the players.]

Q3 / On Superapps

All right. Now question three, multi-app companies versus deep moats. Are we witnessing a shift in India from building single deep tech unicorns to multi-app companies? Interesting contrast.

Is this driven by a realization of limited vertical depth in Bharat markets, or is it just a gold rush mentality? This one is a complex one. But let me kind of go back to the fundamental point it’s asking. The contrast is between multi-app versus perhaps single superapp. 

[SPW: I think I misunderstood this question when I read it out! It is probably asking for why a Zomato or Swiggy for instance had to expand beyond food services, and why a Urban Co had to expand into selling hardware. I read it instead as a super app vs single usecase app comparison.]

You know, the multi-app or kind of multiple use case in a single app is sort of the holy grail of consumer tech, right? Everybody wants to create like a Tencent or an Alibaba or, you know, kind of company. And when one app kind of solves multiple use cases. But I’m very skeptical when it comes to India, whether we’re going to see like a genuine multi-app.

If it had to happen, why hasn’t happened till now? Zomato, for instance, right? Could have been that. They had like a fairly highly frequent app, which is food ordering, where people were using it two, three times a week, perhaps. But when it came to Blinkit, which is even a higher frequency, they choose to kind of spin it out.

Then District is a separate app as well, right? Why did Zomato choose to do that? It’s an interesting one. Secondly, if anyone could have done that in India, I think it would have been PhonePe or perhaps WhatsApp. Because again, there are very high frequency use cases built in, they could have leveraged that to add other components.

But if anything, they aren’t following the script, right? Because PhonePe launched Pincode, which is a e-commerce app separately. So it’s a thought. I don’t know why we aren’t seeing specifically one multi-app, which is solving for multiple use cases.

Why aren’t we seeing a Swiss knife of consumer apps? I think it’s specifically to do with the fact that apps like these are heavily complex, right? And fundamentally, what you want to kind of do is keep the UI simple so that customers can kind of come and use the product and get what they want. And one of the things you’ve seen in the UI of Chinese apps, Japanese apps is they tend to have higher density of information or options packed into the screen, whereas Indian apps are not as highly dense. This could be something to do with it.

But it’s an interesting question. And as with all interesting questions, sometimes you don’t have a decisive answer, right? 

Q4 / India and manufacturing

And question four, this one’s on manufacturing. When do you think India’s manufacturing ecosystem will meaningfully take off? And can it ever become a larger story than India’s services sector? Let’s look at data.

Services is well over 50% of India’s GDP. Industry is about 30%. Of that, Manufacturing is about 17%. And interestingly, the other two are construction, which is about 9-10%. And mining and utilities, which are about 4-5%.

So, I very much doubt if manufacturing can overtake services. And if anything, the trend has been that services has grown faster and manufacturing’s share of GDP hasn’t. So, in the report, we actually dive deeper into why this is so.

Why is manufacturing not growing in India? I think our understanding was that fundamentally, it’s rate limited by the constraints of land, labor, and capital. Land is important. It’s very hard to get large packets of unencumbered land with good rights in India. And that fundamentally creates a challenge in setting up large plants. We’ve actually looked into India versus Thailand in the report where we’ve actually shown how it’s far cheaper to manufacture in Thailand.

Second, labor. Strangely, skilled labor is more expensive in India than in many other countries, right? And our peer countries when it comes to manufacturing. And that creates a challenge as well.

Now capital. Underdeveloped countries are always capital scarce, and that’s true of India too. So, if it were to change, what would need to change? I think we would need to see the constraints of land and labor disappear. Capital can still be managed. There’s enough capital that can come in from the venture funds, government, etc. So, that’s solvable.

What needs to be solved is land and labor. So, when would we see a manufacturing renaissance in India? Possibly, if you are able to look at micromanufacturing, which doesn’t need a very large surface area, which can be possibly one, two machines in maybe vertical things. Second could be robotics playing a role, where labor is not a constraint.

So, if some of these come in, we are likely to see manufacturing take off. Because if you look at it, every restaurant is a manufacturing unit, right? You can think of an Udupi as manufacturing dosas, like idlis, right? And I think when it comes to light manufacturing, India is able to kind of pull it off. But when it comes to heavy duty manufacturing, that’s where the constraint is.

So, when machines get smaller, we are able to pack more into less square footage. When robotics plays a bigger role, that’s when I think India is likely to see a manufacturing renaissance. So, the products are going to come out to be more personalized, more customized, perhaps, short runs.

And that is when we’ll start seeing interesting things happening. And there are companies like, for example, playing across like Zetwerk, that’s kind of shown that we can do large manufacturing at scale. That’s an impressive company. They’ve shown that we can aggregate small businesses who are manufacturers to create cloud factories, right? So, secondly, Ethereal Machines, for instance, like on the other contrast, has been able to create fairly cutting edge, high axial 5D printing machines, right? So, we can kind of innovate there as well. So, we’re likely to see more innovators emerge in the manufacturing space in India. But eventually, they’ll take off only when the constraints on land, labor and capital disappear.

Q5 / Building a scalable edtech venture

Question five is about the aspiring founder mindset. As a first time founder trying to build an edtech business, how do you scale to 100 crore revenue without following the typical growth hacks? Could you share a more visionary framework centered on mindset, strategic depth and intuition that you believe leads to meaningful and sustainable scale? Hmm. So, I do believe that the biggest mistake education founders did was to marry tech with it.

So, I know I should be careful, right? Edtech is fundamentally an oxymoron, because education and scale don’t go well together. And tech is fundamentally about driving scale. So, if you want to deliver quality, then education cannot scale without some quality breaking.

And this is kind of an axiom in education. And what tech should do is to be an ideal handmaiden to education by enabling certain things, for example, trying to kind of enable it to be available cheaply, kind of speed up, for example, some aspects around delivery. But if you try and build an edtech business, where the only purpose is to scale rapidly, and then quality kind of becomes a challenge and you cannot deliver quality.

So, I would say you certainly can build a sustainable education business, leveraging tech. And the way to do that is to fundamentally worry about the problem. Take any problem. Maybe it’s about teaching math to high school students. There’s such a shortage of high quality math teachers, because they’re all going to tuition centers and test prep centers, right? So, schools and parents are dying out for good math teachers. 

Or take the challenges around healthcare skilling, like one of our portfolio companies Virohan is doing. There are so many people who aspire to get jobs in healthcare, and there is so much demand for it. But there aren’t enough academies providing skilling and training. And that’s what they’re doing. So, just take these two problems, and you could build very large businesses around it. And similar to this, there are many such businesses where you can build with purpose and scale. So, scale will come.

But scale won’t come from day one. Scale will come when you put together a solution that’s affordable, that delivers quality, and eventually leads to successful outcomes. And if you do that year on year on year, you will build a large business.

In education, you can’t take shortcuts. And what has happened, sadly, is in the 2020-22, I think we’ve had some bad role models trying to grow unsustainably, on the back of a lot of capital. And I feel like some of those lessons need to be now taken to heart.

And we now need to kind of focus on building sustainable education businesses, not edtech businesses. 

Q6/ Scaling with purpose over playbooks

Question 6, you know, is sort of a continuation of question 5, except that it kind of goes beyond edtech. Scaling with purpose over playbooks.

For first-time founders who don’t want to follow the usual blitz scaling playbooks, how do you advise building long-term meaningful companies rooted in principles and not just metrics? Love the question. So, I think the fundamental mistake we do is to confuse venture and entrepreneurship. Venture is just one form of entrepreneurship.

Take Haldiram’s, take Airtel, take 100 such names. They were startups too, once upon a time. They were not venture-funded startups. But today, because of all of the excitement around venture and all of the articles and all of the podcasts, we think venture funding is the only way to, you know, start companies. It’s not. Fundamentally, companies are around solving a problem for customers.

And venture solves the funding problem. But venture is not the only way. For instance, you could build a company over 20, 25 years by reinvesting profits. Venture allows you to bend time, to compress time. The same company now can be built in 4 to 5 years by loss financing, because venture says make losses and we will finance you and you will grow in 4 years, where somebody else would have taken 20 years to grow. But not everybody needs to do that.

Not every business needs to go through that. Venture is ideal for businesses where there’s a cold start problem, like Uber and Airbnb, like marketplaces, where you need to bring both sides together. Venture is terrific for that.

Look at all the drivers, all the people letting out rooms and making extra money. Or for example, Zepto. For example, all the folks who are kind of employed through Zepto doing gig work. So venture can create great impact. But venture is not the only way. And I feel that founders of all stripes need to kind of figure that venture is just a financing mechanism.

You don’t necessarily have to blitzscale. Blitzscaling works for certain kind of startups. It works when you nailed the product, when you have what’s called product market fit. Then you can really throw in the money into growth and kind of blitzscale. But if you try to blitzscale before you’ve nailed product market fit, you can get into very murky, dangerous territory. You can wreck the company.

So I would say blitzscaling is great in the right context, the right stage for a certain kind of company. Not everybody needs to go through that. It’s only relevant for certain kind of venture companies. And venture is only relevant for certain kind of companies. 

Q7 / The chicken-egg problem of traction vs capital

So question number seven, traction versus capital chicken egg. Many early stage consumer tech founders face this paradox. VCs want traction before investing, but traction often requires upfront capital. What’s your advice for founders facing this? 

So the secret of venture is that there are two types of founders and there are two types of venture funding mechanisms that have evolved to cater to these. So if you’re a second time founder, you don’t need traction. You’ve proven yourself through previous company founding. You’ve done well. Or even if you haven’t done well, you’ve shown that you have the requisite experience.

So typically, VCs are happy to fund second time founders on paper plans alone. And there’s a lot of money that can kind of flow into these founders. But if you’re a first time founder, yes, you do need to show some traction.

And yeah, you’re right in saying that if you don’t get access to funds, how will you show traction? So there are two bits of advice I have for you. One is figure out what stage of VCs you’re talking to. If you are a first time founder, then try and take some money from friends and family or angel investors who are happy to kind of fund you.

For example, you’re working in a high growth startup and you can ask the founder of that high growth startup; if it’s a unicorn, then even better because they would have had some liquidity, some outcomes, and they’ll be able to kind of give you some starter capital. Through that starter capital, you can build the business to a certain stage, you can show growth, and then approach a VC who’s relevant for that stage, like a pre-seed VC or a seed VC. And then this is the capital that you get, which gets you to the next stage, you can approach a Series A VC.

So I would say stage your capital raising requirements. Don’t come to a series A VC if you’re a first time founder with a paper plan idea. Very, very unlikely that you’re going to get funded.

Second, let’s say that you have growth, you have traction, but VCs are passing on you saying that’s not enough. So what should you do? What you can do is to research out which VC has an interest or a thesis in your sector. Don’t approach every single VC. Go find out who those specific VCs are and build a relationship with them. Because they believe in your sector, in the problems that you’re attacking, and they have a thesis in that sector, they have a prepared mind, and they may be willing to take a risk. They may say that, hey, sure, we would have ideally liked this revenue, but, you know, willing to take the chance on you.

So you need one VC to say yes, right? So these are the two bits of advice I would tell you to solve the capital versus traction chicken egg problem. 

Endnote

Every year to write the Indus Valley report, my colleagues and I go into hibernation. It also coincides with Delhi’s winter. And December to Jan and early Feb is all about looking at hundreds of reports, you know, creating so many charts. And by Feb, you know, you’re just grateful that like the report is out. And now’s the time I get to kind of go and read all the books that are piled up, hear all the podcasts, see all the videos, etc, right? But this time, the Indus Valley report had an overwhelming response, so much interest in the report, so many people writing in, and it allowed us to kind of see that the questions, the problems, the challenges of the Indian startup ecosystem is facing, are now becoming mainstream, you know.

So this is India’s startup ecosystem’s moment to shine. And I hope that the report is able to create awareness of the challenges of the problem we’re all facing amongst the wider community. 

But this was exciting. It was great to kind of, you know, have the session, thanks to the SeedToScale team, do this AMA, the great questions, there’s so many I wish I could have covered, but time is a challenge. We still covered some fantastic questions. And I would love for you to keep, you know, the responses to this going, feel free to share this on social media, feel free to share your thoughts and criticism. Maybe I’ll respond there. Yeah, thank you so much.