Inside Girish Mathrubootham’s quest to build India’s global AI giants

Was there a tipping point that made you decide to focus on Together Fund and AI?

One of my life lessons has been about riding a wave. In 1998–99, I ran a Java training business and found it easy to scale. But I had to shut down the Java training institute in 2001 after coming back from the US. What mattered was not that I was a better trainer in 2001, but that the wave was gone. I asked the same question before starting Freshworks. I was happy and well-paid at Zoho, but when I saw the SaaS wave coming, I knew I should be riding that.

I think this is something very similar; it’s almost like a déjà vu moment. At Freshworks, we were not afraid of any competition, but AI was the most important thing that could change the trajectory of the business. We saw the massive disruptive potential after the ChatGPT demo in November 2022. As a product person, I got very excited. I’ve spent 15 years operating one company; now, I look at my role as ‘how can I enable hundreds of companies?’.

SaaS was the catalyst for the dream of India as a product nation, and now AI is the accelerator.

Was it a bittersweet moment to step away from Freshworks, or had you prepared for it?

It was a well-thought-out transition plan. It was not a single moment. We hired Dennis Woodside in September 2022, and we practiced operating as ‘two in the box’ for 18 months. I then stayed on for another year-and-a-half to finish the transition, write the AI strategy for the company, and hire the next leadership team. The transition happened in my mind the minute I gave the CEO title.

What is Together Fund’s AI philosophy now, and how do you view startups in this space?

Since India currently lacks its own foundational models (like OpenAI or China’s DeepSeek), the opportunity for Together Fund is predominantly in the enterprise application space and developer tools. We have the experience to understand the workflows and customer problems in enterprise applications, allowing us to use available AI technology effectively.

We see tremendous opportunity in enterprise applications, both horizontal and vertical in several industries. From developer tools to defense tech. A key area is consumer AI.

Until now, consumer AI has always been in the hands of the companies. Like, Netflix uses AI to recommend movies for you. Amazon uses AI to recommend products for you. How about you having your own AI to give you personalised recommendations? We have not built that yet. We have partnered with Prosus and we are going to incubate a company.

From an investment strategy standpoint, we have three pillars or three vectors along which we operate. This is predominantly based on the founder archetypes that we work with.

Netflix uses AI to recommend movies. Amazon uses AI to recommend products. How about having your own AI for personalised recommendations?

On one end, we have the Swarm Space Studio, an initiative that comes with AI, access to AI research, AI lab, and community. Swarm Space is for mostly younger founders who are typically either in college or just out of college. They are very good in tech, but they have never built companies. They don’t really understand customer problems.

We have a 24-year-old founder called Akshat (Tyagi), who is the founder of Metaforms. The company will probably hit $5 million now. Akshat has never worked in any company. So he does not understand business. Now he has entered the ‘market research’ market and he has a lot of Fortune 500 companies that are using his products.

Then we have the regular VC deals that we do. Emergent (AI coding platform) is a good example of that, where the CTO of Dunzo (former chief technology officer, Mukund Jha) wanted to start up with his twin brother. So he came and met me and we decided to invest.

When I say a typical VC deal, we aim for 15–20% ownership. We may write a $2–5 million check. In Swarm Space Studio, we write the checks usually up to $1 million… and aim for 8–10% (ownership) of the company.

On the other end, we have an incubation strategy, where we have the idea, we are coming up with what kind of company we want and which space we want. Like I told you, the consumer AI space, where we have partnered with Prosus and committed a total of $15 million. We have 50% coming from Together Fund and 50% coming from Prosus.

Now, we are finding the initial founding team who will believe in the idea and who have the capability to execute on that idea. That’s the third model, where our ownerships will be higher, because we are like a co-founder.

The internet is open and treated as a public good. That’s not the case with AI. Countries are using it as a geopolitical tool. In your view, what is the risk that India faces? How should India approach it?

Definitely, AI is a geopolitical tool. But so is tech itself. Like see, Amazon and Google and Facebook, they are all powers, software powers. They can decide who can have access to [their platforms] if they want.

India’s soft power is in IT and tech. So we have to build our own capabilities. That’s the short answer.

Everybody has to collaborate. The government has to collaborate with industry and academia, and really have a focused effort where we have to cut our reliance on those [global] models, because every country is starting to go towards sovereignty.

It is going to be a theme that we have to live with, not that we want it. In our generation, we have been beneficiaries of globalization. We are starting to see nationalism and sovereignty play a bigger role. So we’ll have to prepare for that. And I’m sure the government is working on making some of these things happen.

Now that you are involved in AI, and the government has its AI Mission, will we see you working with the government in an advisory or a more active capacity?

I think we will probably work with the startups building these solutions. We will also establish tie-ups with academia. Personally, I don’t think I’m qualified to be an AI advisor to the government. There are deep technical experts who will probably play a better role there. But we will definitely work with these young startups that are building deep tech.

Considering AI is still evolving, what are you telling startups about building a moat?

One advantage that startups have is they can move fast and adapt. A good example is Composio, which is one of our portfolio companies. When they started, they started off with trying to build agentic integrations. But when the Model Context Protocol (MCP) was launched by Anthropic, they were one of the first ones to adopt it and quickly build the entire MCP infrastructure, where today Composio is widely loved and used.

More than moat, I would say speed of execution creates the opportunity to serve the market when the need is hot. Then, adoption becomes the moat—having more and more customers.

On the other hand, on the enterprise application side, I’ve talked about the music and gadget analogy. AI by itself is a gadget. What the customer values is the music. We can throw away the gadget, but we hold on to the music.

AI by itself is a gadget. What the customer values is the music. We can throw away the gadget, but we hold on to the music.

Can you really solve customer problems well? If you really look at enterprises, they really don’t care about whether you’re using OpenAI or Claude or DeepSeek. What they care about is, “Hey, if you’re, let’s say, automating the call centre, can you do it without sacrificing my customer experience? And can you give me 70% automation, 50% automation? And if my customers are happy with the quality of the support and if my CFO is happy with the cost, then I will buy these tools”. So, our advice to founders has always been to focus on the music.

But sometimes big moves happen, especially in the developer tool space. For example, OpenAI recently launched its agent toolkit. What OpenAI is trying to do is like a phenomenal play to try and become the operating system of the future… That is the big play. Companies like Microsoft and Google should be worried.

[OpenAI] built an agent toolkit that allows anybody to build connectors to their apps. That puts startups that were doing that at risk. Things are changing fast. This is all part of the game. There is such a gold rush that some people will be successful, [while] some people will be falling behind and failing. But the ones who will win are the ones who are making the picks and shovels, as in any gold rush. That is our job, to find enough people who are making the picks and shovels.

What are you looking for when you assess startups to invest in?

We are an early-stage [investment] firm. Many times we are pre-revenue, pre-product. The most important bet is on the founders: what makes them unique, or what is their X factor, or what is their super power. And we then look at are they trying to solve a hard problem? The problem has to be hard, otherwise 100 people will do it. Then we also look at what gives them the right to actually win and to be able to crack that hard problem.

For me, successful founders are the ones who have spent enough time worrying about the problem domain. If technology is available, anybody can build it. But a deep understanding of the problem domain is what is required. That is where we spend our time.

We also look at: do they have the ability to convince people who are as smart as them or smarter than them to join them? This is a long journey. The whole game is talent.

And then we also look at craftsmanship. Whatever they are building, nobody is going to buy it because it is cheap. The India low-cost play has ended. You have to really have a world-class product. And then, obviously, the market. Is this a big enough market where a large company can be built?

There is an expectation from AI startups of rapid revenue growth. Has this changed?

I think over the last six months, I would say our learnings are also evolving with certain nuances. There is one set of AI-native companies that I wrote about as ‘cheetah companies’. These companies are really scaling fast, like Emergent. $10 million in two months, $50 million in three months. When they went to $10 million, they had like 10 members. Now they are probably 15, 20.

But the point is that this is not a model that will work for every AI startup. Today our nuanced understanding is, for one set of companies, specifically around completely new areas like vibe coding, you have these ‘cheetah companies’ possible.

But it is quite okay to have maybe another AI-native startup, like we have Unstract, which is selling to larger enterprises. These are $200,000–300,000 deals that go through the enterprise sales cycle. These are deep technical products which have to be integrated into the data engineering teams, and they have to evaluate it and put it in production. The sales cycles are longer. But that’s fine because the customers still love the product.

The enterprise adoption is proving the product market fit. They may not actually scale to $10−15 million in two months, it is fine. But when they raise their follow-on funding and hire a GTM (go-to-market) team, they can get some speed and velocity.

It’s been four years of Together Fund. How would you rate its performance?

What we have managed to accomplish as a brand, we are really proud [of] as a team, because today Together is being compared with the top ones in the country—Accel or Lightspeed or Peak XV. That, I think, is a phenomenal accomplishment because these are brands that have been there and these are global brands. We take a lot of pride that we have done good work to earn that reputation as a brand.

We made the call to pivot to AI much before everybody else and we made some really good AI investments.

We are also fortunate that we made some right decisions, where the fund’s performance is doing really well and that is being appreciated by investors. We made the call to pivot to AI much before everybody else and we made some really good AI investments, which are paying off really well.

You look at companies like Emergent or Composio or Confido Health or Spry, or Presentations AI. All these companies are doing well. We have many companies that have scaled past $5 million, which is a rarity in the startup world. When you are investing in early-stage [startups], I think having so many companies that have crossed $5 million and $8 million is really a good positive sign.

The Freshworks IPO was the metric to define your success to a certain extent as a founder. When you look at Together Fund, what will be that success metric for you?

That’s very clear in my mind. If we are able to create the next 10 global AI companies from India, and if Together is part of that, that’s success.