Boundless Ventures

what if india's biggest ai opportunity is in products that understand how indians actually shop, learn, and live? 950m internet users. mobile-first habits. growing ai comfort. the ingredients are here. the question is who'll build it. our gp Natasha Malpani shared some thoughts with Inc42 Media on why the breakout happen at the intersection of consumer and infrastructure products and what we're backing at boundless. Ankush Das

we’re hosting a small private breakfast in bangalore on June 4th. founders, researchers, operators. people building across AI, robotics, biotech, and frontier tech. the format: we invite ten outliers. each brings one friend doing groundbreaking work. details below Natasha Malpani Prashant Kataria Rishabh Singh

AI-designed drugs are passing Phase 1 at 80 to 90 percent. the narrative says this is proof the approach works. Phase 1 tests whether the molecule kills people. the 90 percent failure rate in drug development sits in Phase 2 and 3, where drugs fail on efficacy. AI has not moved that needle. one drug cleared Phase 3 in December 2025. one. after decades of promises. the field should be asking precisely why that method worked rather than assuming it validates everything else. South Asians are 25% of the world’s population and 0.8% of genomic study participants. the drugs being designed right now are optimised against the wrong population genetics for two billion people. the model companies will commoditise. the data infrastructure companies will compound. and the most valuable data infrastructure in biology is the one that does not exist yet.

AlphaFold compressed fifty years of structural biology into months. most people read that as an AI story. it is a digitisation story. the unlock was never the model. it was the search space going digital. that distinction changes everything about where value goes when biology finally digitises. we mapped it. four parts. 

everyone is using the term "world models." nobody is describing the same thing. that is not a semantic problem. it is concealing three different bets with three different timelines. the generative bet produces convincing visuals. the infrastructural bet owns the simulation supply chain. the cognitive bet, what lecun is actually building toward, requires architecture nobody has yet built. for a research paper, the confusion is tolerable. for capital allocation, it is not. the foundation layer consolidates to three or four players. everyone else becomes an application on top of someone else's model. three definitions. three timelines. one winner per layer.

twenty five founders and operators in one room. one question. what gets built in orbit and who builds it from india. uplink brought the kind of conversation india's space ecosystem needs more of. we brought together four founders already building this layer from India for a very honest fireside chat: Gaurav Seth (PierSight) is building a maritime intelligence constellation using SAR satellites. Jay Panchal (Aule Space) is building autonomous docking and servicing for satellites in orbit. Shreyaans Jain (QOSMIC) is building the optical ground station infrastructure that replaces RF. Yogeshwaran Jayaraman(Astrogate Labs) is building the laser communication terminals that connect satellites in orbit. the conversation got specific. all four spoke about how deliberate the choice to build from india has been. the ecosystem support is real, the talent coming out of ISRO is increasingly finding its way into startups, and that is a meaningful shift for the sector. the collective view in the room: the infrastructure layer is only going to get bigger. there is a significant amount that will get built from india and the foundations are already being laid. but the honest part: the momentum is there. the commercialisation is not yet. getting to global markets, positioning indian space products for international customers, that remains the harder problem. thank you to Gaurav, Jay, Shreyaans, and Yogeshwaran for the honest perspectives. to Prashant Kataria for moderating a conversation that needed no steering. and to everyone who showed up for uplink.

when machines can finally understand natural language, the barrier between human intent and machine action collapses. the mistake most businesses are making right now is using generative AI to do the same things faster and cheaper. streamline the workflow. cut the headcount. reduce the cost. that is the floor. the ceiling is entirely new products, markets, and experiences that could not have existed before. a personalised AI tutor that adapts to exactly how your child learns. a hospital that redesigns primary care from the ground up. a satellite constellation that makes the ocean visible in real time. the businesses that will matter in ten years are not the ones that used AI to reduce costs. they are the ones that used AI to build something that was previously impossible. our founder Natasha Malpani joined a panel at the et maharashtra business summit with Nitish Mittersain (ceo at Nazara Technologies Limited), Shrini Viswanath (founder at Upstox), Ankur Sharma (founder at Rebel Foods), and Gaurav Sachdeva (ceo at JSW One Platforms Ltd.) unpacking exactly what that means for indian business. ETEntrepreneur

the infra layer and the application layer aren't separate anymore. the winners are building both. intelligence alone isn't enough. it has to be reliable, trustworthy, deployable in the real world. that's the actual problem. consumer AI has no clear vertical breakouts yet. that's not a warning. that's the window. india has the users, the willingness to pay, and the technical depth to build what's next. the evidence is already there. thank you Outlook Business for featuring our GP Natasha Malpani on exactly this. PierSight, Armatrix, Alter, Shram, glide, superhealth Shashank Bhatt, Outlook Publishing (India) Pvt. Ltd.

we are overwhelmed by the response to uplink. the timing could not be more right. last week, GalaxEye launched Mission Drishti, the world’s first OptoSAR satellite and india’s largest privately built earth observation satellite. Pixxel announced Pathfinder, india’s first orbital data centre satellite, built with Sarvam. the in-orbit economy is not coming. it is here. and that raises the questions that actually matter now. how do you connect satellites once they are up? what does it take to build systems that let them work together in orbit? and where does india sit in shaping this infrastructure layer before it is defined by someone else? if you are a founder or operator in indian spacetech, an institutional voice, or an ecosystem stakeholder shaping the sector, you should be in this room.

35 consumer AI teams, selected from over 200 applications for the boundless x anthropic consumer AI breakout program. travelling in from hyderabad, mumbai, delhi and chhattisgarh. the energy was electric. we had an incredible mix of founders building across health, fintech, entertainment, productivity and education. the Anthropic team shared how consumer AI is shaping differently in different markets and why india is uniquely positioned. the founders of Round1 AI by Grapevine, Supernova and Dashverse got on the hot seat. they shared honestly what actually worked, what broke, and what building for indian consumers really demands of you. the working groups on consumer insight, distribution, retention and monetisation took the conversation even further. grounded, specific and driven by people in the middle of figuring it out. what made the room work: founder-to-founder honesty. no pitches. real lessons from people building in the mess. thoughtful curation meant every conversation added value. we are early in india’s consumer AI story, but the founders writing it are already here. june 8th is next. a closed technical deep dive with anthropic’s applied AI team for shortlisted teams. thank you to Yeop Lee, Aditya Chauhan, Sriranjani Hadigal from Anthropic for making the trip and bringing real depth to the room. Saumil Tripathi, Maharishi R B, and Soumyadeep Mukherjee for the honest hot seats and Natasha Malpani, Rishabh Singh for the curation and hosting.