Speaking with TechGraph, Chetan Mehta, Founding Partner at Aum Ventures, outlined why deeptech remains one of the most underpriced opportunities in India’s startup ecosystem despite growing interest in sectors such as semiconductors, space technology, defence systems and advanced materials, and explained how the firm’s investment strategy focuses on identifying long-term technology companies that sit outside the traditional venture capital cycle.
Mehta also highlighted how Aum Ventures structures capital deployment around clearly defined milestones and helps founders validate commercial demand and resolve critical technical uncertainties as they build intellectual property capable of sustaining long-term competitive advantage.
Read the interview in detail:
TechGraph: Deeptech has become a strong narrative in India over the last few years, yet capital still tends to follow visibility and near-term outcomes. What made you build AUM Ventures around a thesis that goes against the natural flow of capital?
Chetan Mehta: The contrarian nature of it was deliberate, not incidental. When capital crowds around visibility, it creates persistent mispricing at the harder end of the technology spectrum, and that mispricing is where patient, informed capital can generate outsized, durable returns.
Deeptech companies building genuine IP operate on longer development arcs. That means they are systematically undervalued in the early stages, not because the opportunity is smaller, but because most capital is not structured to hold conviction through uncertainty. We built Aum Ventures specifically to occupy that space.
The broader context matters here too. We are at a once-in-a-generation inflection point where physical-world technologies, be it semiconductors, space, advanced materials, defence systems, are moving to the centre of global economic and geopolitical competition. India has structural advantages in this transition that are not yet fully understood or priced by the market. The early-stage window to back the companies that will define these categories is now and narrow.
TechGraph: There is increasing talk about India owning critical technologies, but sectors like semiconductors, space, and advanced manufacturing depend on long supply chains and institutional depth. Where do you see India genuinely building leverage today, and where is the gap still too wide?
Chetan Mehta: India has earned real leverage in three areas, and I would argue the depth of that leverage is underappreciated globally. In semiconductor design, India’s engineering talent pool is not just large; it is deeply experienced in complex, systems-level work that most countries cannot replicate quickly. The design heritage built through multinational R&D centres over two decades is now starting to flow into indigenous startups with genuine ambition.
In space systems, ISRO’s cost-efficiency record has created a credibility foundation that Indian new-space startups are building on, and international customers are paying attention. In defence electronics, the combination of a committed domestic customer, a forced indigenisation policy, and an engineering base that understands real-world constraints is producing companies of real quality.
Where the gap remains structurally wide is advanced manufacturing at scale. Capital intensity alone does not explain it. The challenge is institutional: supply chain ecosystems, metrology infrastructure, process engineering depth, and tolerance for long qualification cycles. These take decades to build. We should be honest about that timeline rather than pretending policy announcements close the gap.
Materials science and quantum computing are early-stage globally, which means India’s position is less asymmetric, and that is actually an opportunity if founders and capital move with urgency now.
TechGraph: A lot of early-stage deeptech companies look promising at the technology level, but struggle to move beyond that into repeatable business models. At the point of investment, how do you separate technical possibility from commercial reality?
Chetan Mehta: The question we keep coming back to is deceptively simple: is this a research project or is it a company? The answer lives in whether the team has genuinely grappled with who will pay, how much, why now, and what replaces them if the customer doesn’t show up.
Technical brilliance without a commercial hypothesis is interesting; it is not investable at our stage. We look for founding teams that are rigorous about both domains simultaneously. Not teams that have fully solved the commercial problem; that is unrealistic at early stages in deeptech, but teams that have mapped it honestly, tested their assumptions against real potential customers, and structured their next twelve to eighteen months of work around resolving specific uncertainties on the path to revenue.
Practically, we weight letters of intent and structured pilots far above pitch decks. We care about the quality of the customer conversations, whether the team has reached decision-makers with real procurement authority, or just enthusiastic engineers who have no budget and no mandate. We look for founders who can move between a technical whiteboard and a customer’s boardroom without losing credibility in either room. That combination is rare, and when we find it, we move quickly.
TechGraph: In this category, more capital does not always solve the problem, and in some cases, it can even accelerate the wrong outcomes. How do you think about capital discipline when backing companies that are still working through fundamental uncertainties?
Chetan Mehta: This is one of the most important questions in deeptech investing, and it is one that everyone in the broader venture community should think about rigorously enough.
Capital deployed before fundamental uncertainties are sequenced correctly does not accelerate progress; it accelerates the wrong things. It inflates headcount before the technical questions that justify it are resolved. It creates momentum pressure that makes honest reassessment harder. It raises the stakes on decisions that should still be in experimental mode. We have seen well-funded deeptech companies fail not because the technology did not work, but because capital forced a pace that destroyed the conditions needed for the technology to work.
Our discipline is structured around what each tranche of capital is actually buying in terms of de-risking. If a founding team cannot clearly articulate what the next twelve months of deployment will prove or disprove, not aspirationally but specifically, that is a serious signal to us, regardless of how compelling the technology looks on paper.
We prefer founders who are genuinely comfortable with uncertainty and honest about the boundary of their current knowledge. That intellectual honesty is not a weakness. In deeptech, it is the prerequisite for making good capital allocation decisions.
TechGraph: With AUM Ventures focusing on IP-led businesses, what does real ownership of technology mean to you in practice, and how do you test whether that ownership can hold up as competition builds?
Chetan Mehta: Patent filings are the beginning of the conversation, not the end of it. Real IP ownership means defensibility that compounds over time, and that requires something harder to manufacture than a patent portfolio.
What we actually look for is whether the underlying know-how is genuinely difficult to replicate, and whether that difficulty is distributed across the organisation rather than concentrated in one or two individuals. The most fragile IP positions we have seen are technically strong but institutionally thin: the knowledge lives in a founder’s head, the documentation is sparse, the team depth is shallow. A well-resourced competitor with patient capital can dismantle that position faster than most founders expect.
Real ownership, in our assessment, is a combination of patents, documented processes, institutionalised expertise, ongoing R&D that keeps the frontier moving, and a team culture that treats IP development as a continuous discipline rather than a one-time event. We also look carefully at whether the competitive moat widens as the company scales; network effects in deeptech are rare, but they do exist, particularly where proprietary data generation or systems integration creates lock-in that pure technology replication cannot overcome.
The stress test we run is straightforward: if the three most senior technical people left tomorrow, how much of the real IP would walk out the door? The answer to that question tells us more about durability than any patent filing ever will.
TechGraph: Many founders in this space face an early choice between building for global markets or aligning with domestic demand that may evolve more slowly. How do you see that decision shaping the trajectory of deeptech companies from India?
Chetan Mehta: It is one of the most consequential strategic decisions a deeptech founder can make, and the pressure to choose global positioning before honestly assessing it is one of the more common failure modes we see.
The framing of domestic versus global is itself slightly misleading. The real question is: what is the right first customer, and does that first customer create the proof point, the revenue base, and the reference case that enables the next move? For some companies, particularly in defence, critical infrastructure, and certain industrial applications, the domestic market is not a consolation prize. It is a strategically advantaged first customer with real procurement intent and a policy environment that actively prefers domestic supply chains.
Global positioning sounds more ambitious, but it comes with a reality check: longer sales cycles in unfamiliar institutional environments, capital requirements that scale with geography, and a higher burden of proof before international customers take a real bet on an Indian deeptech company they cannot easily reference-check. The founders who navigate this well are those who sequence it deliberately, building on a domestic anchor where it exists, and going global when they have the proof points, the capital, and the team to sustain the effort. Going global by default, without a credible plan for how you actually win in those markets, is a resource allocation error that deeptech timelines cannot absorb.
TechGraph: AI has brought significant attention back to core technology, but much of the value capture still sits at the application layer. Where do you see Indian companies realistically building depth in the stack over the next few years?
Chetan Mehta: The application layer will commoditise faster than most current valuations reflect. That is a structural reality of AI development, and founders building there need to be honest with themselves about it.
Where I see Indian companies with a realistic path to genuine stack depth is at the intersection of domain expertise and proprietary data; not AI as a category, but AI embedded deeply in specific verticals where the data, the regulatory context, and the operational understanding create structural advantages that a generalist global player cannot easily replicate from San Francisco.
Defence AI is the most significant near-term opportunity. The combination of a committed domestic customer, data sovereignty requirements that exclude foreign platforms by default, and an engineering base that understands the operational environment creates conditions for building systems of real depth and real defensibility. Agricultural intelligence and industrial automation are similar stories; local context and local data create moats that matter.
Semiconductor design for AI inference is the area where I have the most conviction at the infrastructure layer. India’s design talent is genuinely world-class in this domain, and the global demand for inference efficiency, driven by power constraints and deployment cost pressures, is creating a market large enough for multiple category-leading companies. India does not need to out-compete hyperscalers at the foundation model layer. That is not the right battlefield. The right battlefield is the one where India’s specific strengths create asymmetric advantage.
TechGraph: In sectors like defence, climate or advanced manufacturing, success often depends on external systems such as procurement cycles, policy direction, and institutional adoption. How do these realities change the way you evaluate risk at an early stage?
Chetan Mehta: They change the risk framework significantly, and I believe most early-stage investors systematically underweight this dimension in deeptech. A technology can be technically mature, commercially validated, and still fail because the institutional environment that needs to adopt it is not ready, or because a procurement cycle that was supposed to take two years takes seven.
We evaluate regulatory and institutional risk with the same rigour we apply to technology risk. This means understanding procurement timelines in realistic terms, not optimistic ones. It means mapping the specific institutional relationships the founding team brings, not relationships with researchers or junior officials, but with decision-makers who have mandate, budget, and a genuine motivation to move. It means understanding the policy trajectory well enough to take a view on whether the tailwinds are durable.
The sequencing question is critical. Can this company build a sustainable business in the first three to four years without depending entirely on government adoption moving at the pace they are projecting? Companies that have an answer to that question, a commercial path that does not require institutional readiness to arrive on schedule, are structurally more resilient. Companies that are betting entirely on procurement timelines are carrying a risk that no amount of technical excellence can hedge.
I would also add: in sectors like defence and critical infrastructure, the quality of the institutional relationships is itself a form of IP. It is hard to replicate, and it compounds over time.
TechGraph: Deeptech outcomes tend to be uneven, with long periods of limited visibility followed by sharp breakthroughs. From an investor’s standpoint, how do you recognise whether a company is progressing in the right direction during those quieter phases?
Chetan Mehta: The quieter phases are often where the most important work in a deeptech company actually happens, and where the most important signals are generated.
The signal we weight most heavily is the quality of what the team is learning, not the visibility of what they are announcing.
Are they asking progressively sharper questions? Are they adjusting their hypotheses in response to evidence, or defending their original assumptions against it? Are the people who understand the field at the deepest level, technical advisors, domain experts, early potential customers, still engaged and still intellectually energised by what the company is doing? Those signals tell us far more than a press release or a funding milestone.
The warning sign we take most seriously is a team that projects confidence in the narrative while staying vague about the underlying technical and commercial progress. Polished storytelling that outpaces actual learning is a pattern that tends to accelerate toward failure, not success.
We also look at internal culture during quiet phases. Teams that are honest internally about what is not working, that have the psychological safety to surface difficult information early, and that make resource allocation decisions based on evidence rather than momentum- these are the teams that tend to come out of quiet phases with a genuine breakthrough rather than a rationalised pivot. The culture a founding team builds during uncertainty is the one they will have when the stakes are highest.
TechGraph: Looking ahead, if India is to produce globally relevant deeptech companies, what will define the first set of breakout successes, and where is AUM Ventures placing its strongest conviction today?
Chetan Mehta: The first breakout successes will be defined by a specific combination that is rarer than it sounds: deep technical credibility, genuine commercial instinct, and the psychological resilience to build through long development cycles without losing either conviction or team. Skyroot Aerospace, where we were early investors and which recently crossed a unicorn valuation, is the template. It is a company that held a technically demanding thesis through years of invisible progress, built institutional knowledge that compounded quietly, and emerged with a position where they are considered as one of the most promising spacetech companies globally. That is the archetype we are backing.
On sector conviction: our strongest bets sit at the intersection of space technology and infrastructure, semiconductor design, and defence technology. These are not thematic preferences; they are areas where India has structural advantages that are not yet fully priced, where policy tailwinds are durable rather than cyclical, where the global market is large enough to reward genuine IP leadership, and where the window to establish category-defining positions is open right now.
The global context matters. The world is rewiring its critical technology supply chains in response to geopolitical realities that are not going away. India is positioned to be a significant beneficiary of that rewiring, not as a low-cost manufacturing alternative, but as a high-value engineering partner with genuine IP ownership. Skyroot is the first proof point that this is not a thesis about the future. It is already happening. The companies being built in the next three to five years will not just be India’s first deeptech breakout stories. They will be genuinely global technology leaders.
That is the opportunity Aum Ventures was built for. And the window- the right founders, the right backing, the right moment in the policy and market cycle- will not stay open indefinitely.
