Investment Thesis

    The Non-Disruptible Thesis

    "AI makes building faster. Atoms make replication slower."

    Bits / Software

    Intelligence is Becoming Abundant

    Hyperscalers like Google and Meta are making intelligence a commodity. What once required significant teams, time, and capital can now be built in weeks.

    If everyone can build, what truly lasts?

    Build Speed: Fast
    Replication: Easy
    Capital Bar: Low
    Moat: Weakening
    Atoms / Physical

    The Physical World is Becoming More Valuable

    The world is being forced to rebuild its physical foundations. AI infrastructure needs, Industry 4.0, and urgency around climate are driving innovation in energy, manufacturing, and infrastructure.

    Unlike software, physical systems are constrained by time, capital, and domain expertise.

    Build Speed: Slow
    Replication: Hard
    Capital Bar: High
    Moat: Strengthening
    The Implication

    In a world where anyone can build, anyone can be disrupted too. The most valuable companies will not just be disruptive. They will be resilient to disruption themselves.

    For the past decade, value was created by building faster. But as intelligence becomes a commodity, building software becomes cheaper and simpler. Speed is no longer the advantage.

    Non-disruptible companies have structural moats: IP that cannot be shortcut, regulatory embedment that cannot be bypassed, physical-AI convergence that cannot be replicated, and execution compounding that only deepens with time.

    What We Back

    Four Structural Advantages

    IP-led

    Deep Science and Engineering

    IP accrues over years of domain-specific work that no amount of capital can shortcut.

    R&D timeline cannot be compressed with money.

    Government-tailwinds-led

    Regulatory and Infrastructure Embedment

    Embedded in national systems like defence, power grids and space infrastructure where switching cost is prohibitive by design.

    Institutional dependency takes years to unwind.

    Intersection-led

    Physical-AI Convergence

    Convergence across hardware and intelligence creates a moat that neither a pure AI nor a pure hardware company can replicate.

    Neither side alone can replicate the other's half.

    Execution-led

    Long-term Compounding Advantages

    A decade of proprietary supply chains, enterprise relationships and domain datasets compounds into an advantage that newcomers cannot buy.

    A decade of context is not a job spec requirement.

    Where We Invest

    Six Investment Categories

    Hardware

    Semiconductors and Advanced Materials

    The hardware substrate of AI India. Companies building chips, sensors, packaging and advanced materials that power the physical layer of digital infrastructure.

    Structurally Defensible

    Focus Areas

    Custom silicon for AI inference at the edge

    Advanced materials with multi-year R&D moats

    Packaging and heterogeneous integration

    Sensor technologies for physical-world AI

    India-native supply chain independence