Synopsis
In venture investing there are two ways to make money—buy low and sell high, or buy high and sell higher. In a bubble, everybody thinks buy high, sell higher works forever. But that only works inside the bubble. As soon as the bubble bursts, you go back to needing to buy low. That’s why I push discipline at B Capital, Morgan told ET in an exclusive interview.What were the early years like at First Round Capital?
We decided to treat it as an experiment. We raised a $7 million one-year fund and told investors, “Take a third of what you’d normally invest with us and try this fund, because we think we may do one-year funds for a few years and see if we like it.”
We raised three one-year funds, each slightly larger than the previous one, and then follow-on funds alongside them. Together those six vehicles became First Round Capital-I, which was about $115 million. They performed very well. By 2007 we realised this was a real business.
Was Uber the defining win for the fund?
Uber was in First Round Capital II, yes, but Uber was not the best investment in that fund. That fund returned about 15x and was an extraordinary vehicle. It had around 70 investments. Sixty-eight of them returned roughly 5x. One of them was Square. Uber returned about 22x. But Roblox actually returned more than Uber.
What’s the story behind the Roblox investment?
That came through a funny story. Chris Fralic, our partner, first pitched us a gaming deal and we said no. Then a month or two later he came back and brought his eight-year-old son, Max, to pitch us on Roblox. We did the deal. Because of how we invested and because we did less along the way, Roblox ended up returning more than Uber in that fund.
So yes, it was an amazing fund, but the lesson is that sometimes the thing everybody remembers is not actually the biggest winner.
How did B Capital happen?
That was when Raj Ganguly and Eduardo Saverin came to me. Raj came to New York to tell me what they were building next. I gave them half an hour and said, “I’d like to be a big LP.”
They said, “No, we want you to be chairman. We want you to help us avoid making the same mistakes you made building Renaissance, Idealab and First Round, so maybe we get a quicker start. And you can do it part-time.”
What convinced you to join?
What attracted me was that they were doing later-stage investing, they were global, First Round was not global, India was important from the beginning, they were mostly B2B, and they were partnered with Boston Consulting Group. That made it very interesting.
I spent a lot of time with Rich Lesser, who was then chairman of BCG, because I wanted to understand what he wanted from a venture fund and why BCG would want to do this. They had been burned badly in 2000, when BCG, IBM and J.P. Morgan put together a venture fund at the worst possible time and lost money. So there was a huge immune response inside BCG to the words “venture capital”.
But Rich had a view of how BCG could help, and it worked very well.
You’ve been a part of the venture capital industry for so long; where are we in the AI investing cycle?
I started investing formally in 1982, so that’s more than 40 years. What I’ve learned is that after every bubble ends, a few companies come out much stronger two or three years later. John Doerr (Kleiner Perkins) had said in 2001-2002, ‘The internet is underhyped’. That was after the crash. He said yes, there was a crash, yes, there were too many companies, but the eventual impact of the internet on society was still nowhere near fully understood. He was right. That’s true of AI too. In the first PC bubble, Jim Simons (founder, Renaissance Technologies) once said to me, ‘We have 300 PC companies and every one of them says it wants 1% of the market. That’s not going to work.’ And it didn’t. But then a few winners emerged.
But valuations are overheated…
Yes, prices are too high. In venture there are really two ways to make money—buy low and sell high, or buy high and sell higher. In a bubble, everybody thinks buy high, sell higher works forever. But that only works inside the bubble. As soon as the bubble bursts, you go back to needing to buy low. That’s why I push discipline at B Capital. One thing I enforce is vintage diversity. If we have a three-year fund, I won’t let them invest more than a third in any year.
What do you think of AI model companies and the amount of cash they are guzzling?
I do think OpenAI is probably overvalued, more overvalued than some others. Anthropic, for example, is somewhat better in terms of focus and valuation because it has been more B2B-oriented. To justify some of these prices, you need ten years of earnings, and that is very far out for venture capital.
How will venture capital make good of these AI bets?
What’s different now is the amount of money in venture. Venture didn’t get truly big until the late 1990s, and then too much money came in and pushed prices up. Starting around 2015, everyone wanted alternatives because they believed venture returned 20%, so huge capital flowed in. Naturally returns have to come down if entry prices go up that much.
Josh (Kopelman) once told our LPs (limited partners) at First Round that in 2005, the average pre-money valuation of companies we backed was about $2 million. By 2008, it was $6 million. Entry prices tripled, exit prices didn’t, so expected returns got cut sharply. That’s what’s happening again.
How do you read the geopolitical situation right now with the Iran war?
Venture investors are investing on a ten-year horizon. So, I have to assume any current war is not a ten-year war. What does matter is US-China geopolitics. That remains important. Full decoupling isn’t realistic because TSMC makes the chips everyone needs, while China controls a huge part of the supply chain for rare earths, rare metals and magnets.
Where does India figure in this?
It’s good to see India taking more manufacturing and companies like Apple coming here. But this is still mostly assembly, not yet the full supply chain. China’s advantage is density. In Shenzhen, if you need a new part, there are multiple suppliers within 20 minutes. India hasn’t yet built that kind of regional supplier density.
Where does India stack up for B Capital?
Morgan: We started in India in 2016 with Icertis and Blackbuck. We’ve had good companies, some misses too, like Bounce. Meesho and Blackbuck went public.
Historically, India-for-global had been more our focus than India-for-India. But on the consumer side, India is one of the biggest markets in the world, so we’re open to doing more here. AI may make B2B more competitive, which could make B2C more attractive as an investment area.
We’re also seeing deep tech emerge. In the US we’ve backed Figure, and in India we’re beginning to see more robotics and deep-tech possibilities. A model I like is technology built in India, but sales and marketing based in New York or other global markets. That’s worked well from Israel too.
Mohla: India is a key focus for us. Singapore only has so many companies, and China has become harder to invest in directly. Over the last few years, we also went fairly late into some Indian companies, like Meesho, but more recently we’ve started going earlier. That gives us a chance to participate in value creation sooner, including in consumer companies.
What’s the focus for India?
Morgan: In deep tech, there’s space, defence, robotics and anything which is sort of applied AI, physical AI. Also, there is advanced precision manufacturing. In Asia, India is our core market.
Mohla: There are some India-for-India pockets too, especially fintech. We did Mswipe and another bank-focused company. So, our India strategy is now more nuanced, and I think it fits our strengths much better than before.
What happens to jobs as AI disrupts IT services, especially in India?
India is particularly exposed given its service-heavy economy. Large teams in call centres and customer support are already shrinking as voice recognition and large language models handle first-level queries, leaving only complex cases for humans. Entry-level coding is also being impacted as AI tools take over routine programming tasks. Any function built on repetitive, detail-oriented work—from customer service to accounting—is vulnerable.
Companies are cautious about acknowledging this shift publicly, but the change is underway.
Where should India build in AI?
India’s opportunity lies in building applications on top of existing models, especially in consumer use cases where AI can aid decision-making. There is limited need to build foundational models, except for language-specific contexts.
In B2B, companies reliant on large support teams will need to rethink their models. The next phase of IT services could be AI-led consulting and implementation, particularly in sectors like financial services, insurance and healthcare, where domain expertise combined with data can create strong moats.
