I have shared this in private with a few people. Unfiltered thoughts below.

  • India SaaS was always hard. Compare the amount that has gone into India SaaS versus outcome for these SaaS companies. We still gave our best. Best there was a price arbitrage. A lot of the cost was GTM and people. You could set up a cold outbound center here, hire a big engineering team here, start selling in India and then eventually try to sell in the US. It worked till some point. But when the price point of software today is not the standard 20 dollars per seat but replacing people, suddenly relationship driven sales matter a lot. In that case, you need to do more in person sales. And now the India salary arbitrage does not work too. Indian salaries have gone up. Instead of fixed API cost, you have variable foundation model cost. So you can’t build another Freshdesk from India. See the revenue growth of Indian SaaS startups, how much funding they have got in the last 2 years and which startups are getting invested.

  • Oh but horizontal SaaS works, why don’t we try that? Actually, even in the US it is mostly AI SaaS. And the SaaS startups there are also getting gobbled by PE, listing at lower multiple, and most investments are now in AI. When every idea has 50 copycats, what matters is capital accumulation and narrative building. Say you do an AI design tool from India, raise 1 mil from some Indian VC. You build in Bengaluru. A startup with a similar idea. Similiar ability of the founding teams. But your US competitor is ex Figma, 2nd time YC founder. Who manages to raise 10 mil from a16z. Now a16z is adding them in their market maps, not you. Garry Tan is shilling them every week. Not you. Your Indian VC is panicking. First they will say grow fast. Then only you will get your next round. Then they will see you burning 3X of your revenue in Claude API costs. And they will panic and ask you to become profitable instead. The same schizophrenic behavior based on market cycles. But cycles are like monthly now. Based on narrative that you can’t control. Do you think Cursor would ever get funded in India? No. Windsurf got exit liquidity thanks to Google. Being in the US helps. Network helps. So no horizontal AI SaaS? What do you do?

  • If the main cost for AI SaaS companies is foundational model cost, then the only companies that can survive, not get gobbled up by foundational labs who will build the same use case in their app layer, or burn their runway and die, will be companies that can replace employees. Or cut opex. You can’t do seat based pricing. Where you charge 20$ but spend 60$ each month hoping you will make up thanks to your paid non users like it happens in the gym business model. But the power users in AI SaaS will always consume a lot and use the best model. Churn will be always high. Most of what we are seeing is experimental revenue anyway. Power law applies here. Usage based pricing? You are essentially reselling Claude credits. And again your burn is 2/3X your revenue. Power users always will use the best model. So no seat based pricing. No usage pricing. Just hey, now you can do the same work with 10 people instead of 50. So pay us 2K/ user/ month. I am sure 200$ Claude subscription will become 2000$ soon. And vertical SaaS is where most Indian builders and VCs will be focusing on. Financial copilot. AI doctor. AI nurse. AI accountant. Replacement of jobs. So that you can make 10k$/month and breakeven. But in the US. There is a reason Indian VCs are telling founders to move to the US immediately after their pre seed round.

  • VCs in India are mostly investing in vertical SaaS. For US. In India if you try to sell an AI QA tool, companies will laugh you out of the room. They will rather hire fresher QAs than pay more for an AI tool. Indians just throw bodies at problems. We are not supply constraint. So you can’t have replacement based pricing. Only option is moving to the US and selling there.

  • What works then? What should Indian founders build?

    • Vertical SaaS for the US. One founder should be based in the US. Hiring the majority of their team in India (for the cost arbitrage.) Sell AI copilots to US folks. Cheaper GTM from India. But of the Engineering team here.
    • Cost arbitrage companies. Find experts in India and do data labelling. Let people wear suits and collect data for robotics companies. And other arbitrage opportunities. There are plenty of these in AI.
    • AI services companies. No matter how much people shout on Twitter about AI being a great enabler, you still need people. Earlier services workers were average, but now with AI their productivity has become 10X. You can do software services but with far fewer people, supercharge with AI, far better margins, and make good money. Design agencies. Software MVP building services. VC scale company? Maybe not. But many micro Infosys 2.0 will be founded in the next few years.
    • Sovereign AI. I sincerely believe in this. The future is going to be more chaotic. More tariffs. More sanctions. More blocs being formed. More schizophrenic presidents and ‘let’s protect our people and fuck the rest’ will be the dominant narrative. So no country can rely completely on foreign companies for their AI needs. Imagine tomorrow India not getting the latest models like China does not get the latest GPUs. Anthropic has even stopped giving access to Claude in China. No matter which party is in the government, the future will be around securing sovereignty and every layer in the AI stack being reimagined. Someone was telling me recently that no one believed India could do a moon mission. No one believed we would have our own nuclear program. But we did. When you are put in a position where there is no way out but to push to remain independent, and not be bullied, policies will change, money will come, and the nation will do whatever it takes to never be in a position like this again. China did it the hard way. India will have to do the same.

Not AI, but I think a lot of money and energy will flow into deep tech too. It is not a question of if but when.

Context: I have worked/ met with at least 20/30 companies in the last 4 months since I left my job. And I can tell you the margin structure of most companies in AI. I have met enough VCs too. I think most would agree privately that this is an accurate representation of the situation here.