India is the place where I find this thesis hardest to evaluate honestly.
I want the country to win. I have watched scientific ambition, engineering talent and strategic urgency become more visible. I have also watched announcements, grants, challenges, prototypes, memoranda, trials and purchase orders collapse into one triumphant story about “the ecosystem.”
Proximity creates insight. It also creates confirmation bias.
The danger is that I begin with the conclusion—that India will become a deeptech power—and then collect whatever evidence makes the conclusion feel true. A new programme becomes proof of demand. A technical demonstration becomes a product. An order ceiling becomes revenue. A fabrication announcement becomes semiconductor capability. The story gets larger while the underlying company remains difficult to see.
I needed the framework in this book partly to resist that temptation.
India should not be the patriotic conclusion of the argument. It should be its most demanding test.
Three claims that should not be confused
When people say that India has an opportunity in a strategically important technology, they may be making three different claims.
The first is a national-capability claim: India should possess this technology because dependence creates security, economic or geopolitical risk.
The second is a market claim: Indian and global buyers will spend meaningfully on the capability.
The third is a company claim: a particular startup can own enough of the resulting value to become a durable, venture-scale business.
All three may be true. Frequently, only one or two are.
A capability can be essential to the country but structurally difficult for a startup to capture. The buyer may be concentrated, the development cycle long and the economics better suited to a public laboratory, prime contractor or project company. A large market can exist while an Indian company remains a low-margin engineering supplier. A startup can win grants and development contracts without creating a repeatable product.
Strategic importance is not a business model. Import dependence is not proof that a local supplier can qualify. Technical talent is not ownership. A government problem statement is not a purchase order.
The company-level question is narrower:
Why should this company, built from India, be able to earn and retain a position that another company cannot easily take?
That is what I mean by a right to win.
A right to win has to be earned
I currently look for six conditions.
A demanding first mission
The company needs access to a problem difficult enough to produce meaningful proof. India offers unusually demanding operating environments: contested airspace, variable infrastructure, industrial cost pressure, heat, dust, complex logistics, distributed energy needs and a large range of biological and agricultural conditions.
Solving a real Indian problem can be more valuable than demonstrating a technology in a forgiving environment.
But difficulty alone is not an advantage. The mission must be accessible. The company needs a buyer, test site, operating partner or procurement path capable of turning the problem into accepted evidence.
A knowledge base
India has scientific institutions, engineering talent, industrial operators, defence laboratories, software capability and large technical workforces. These create ingredients for company formation.
The test is whether a team can combine them around a product. Talent dispersed across research, services, suppliers and public institutions does not automatically become company capability.
An ownable layer
The startup must control something that remains valuable after the first deployment: architecture, process knowledge, qualification history, manufacturing, data, a critical interface, a proprietary component or the complete operating vehicle.
If the learning belongs to the customer, the design is controlled by a foreign supplier and the company performs integration by the hour, the country may gain capability while the startup captures little power.
A path through qualification and production
India can produce remarkable prototypes. The scarcer capability is often the path from prototype to frozen design, accepted test, stable yield, traceability, field support and repeat delivery.
The right to win strengthens when the first market allows the company to accumulate this history. It weakens when each deployment requires a new specification, new integration and new proof.
Capital matched to the journey
The company needs enough time and suitable instruments to reach buyer-legible proof. A seed round that funds a prototype but not qualification may leave the company stranded at the most expensive gate. A grant that excludes commercial work may generate knowledge without a route to adoption. Equity that finances every plant or installation may destroy returns even when the technology works.
The opportunity is credible only if the capital path is credible.
Portability beyond the first customer
The domestic mission should become a qualification ground, not a permanent custom-services contract.
The strongest company solves a hard local problem in a form that can travel. It adapts the product for the next Indian or global buyer; it does not reinvent the company each time.
This requires thinking early about standards, documentation, data rights, export controls, certification, interfaces and supply-chain resilience. Export readiness begins in product architecture, not with an overseas sales office added later.
The three patterns I see
Across defence, space, semiconductors, robotics, energy, advanced materials, biomanufacturing and frontier instruments, the technologies differ dramatically. The strategic positions repeat.
The complete challenger
The first company type integrates enough of the system to deliver the buyer's outcome.
In defence, this might be a counter-drone system that combines detection, command and control, effectors, field integration and service—not an isolated sensor shown in a laboratory.
In maritime systems, it might be an autonomous vessel that combines the platform, navigation, payload, communications, recovery and mission software.
In robotics, it might be a production workcell sold against throughput, uptime and intervention rather than a robot arm sold against a specification.
In energy, it might be a storage system that combines cells, power electronics, controls, warranty, service and the operating evidence needed for financing.
In biomanufacturing, it might be a qualified product joined to the process, quality system, capacity and repeat buyer—not a promising strain or molecule alone.
These companies are trying to be David. They absorb coordination because the buyer does not want to assemble the solution. Their opportunity is large because the complete outcome sits close to the purchase decision.
Their danger is equally large. Integration can become uncontrolled complexity. The company may own every operational burden without owning the highest-value layer. It may finance inventory, installation, service and customer delay before the product is repeatable. Vertical integration is not automatically strategy; sometimes it is an admission that the surrounding ecosystem is missing.
The complete challenger wins only when integration produces reusable learning and control.
The indispensable enabler
The second company type owns a layer that a credible challenger cannot easily replace.
Examples might include navigation that works when satellite signals are denied, qualified propulsion, a mature-node or radio-frequency chip designed into a platform, a photonic or sensing module with packaging and calibration, power electronics with field history, a critical enzyme or material, or a test instrument embedded in the customer's workflow.
The opportunity is not to be “a component company.” It is to become the control point inside a larger change.
That usually requires more than the core invention. A chip needs packaging, test, qualification and design-in. A sensor needs calibration, firmware, integration and service. A material needs standards, process compatibility, durability data and reliable supply. A propulsion technology needs qualification, integration and flight heritage.
The enabler must be narrow enough to dominate and complete enough to buy.
This path can suit India particularly well when the global system is being rebuilt and buyers need alternatives to concentrated suppliers. It can also trap companies in invisible subcontracting. If the prime owns the specification, interface, customer data and qualification record, the enabling startup may remain technically useful but strategically replaceable.
Arming David works only when the armourer retains power.
The orphan technology
The third pattern is the one I worry about most.
The technology is real. The demonstration may be impressive. The team may deserve admiration. But the company has not built the complete vehicle to reach the buyer and does not control an indispensable layer inside somebody else's vehicle.
A tape-out is not a design-in.
A defence prototype is not accepted procurement.
A robot demonstration is not a production cell running at contracted uptime.
A novel material is not a qualified input with repeat offtake.
A laboratory yield is not stable manufacturing yield.
These statements can sound unfair because they place commercial burdens on people doing genuinely difficult technical work. I do not mean to diminish the work. I mean to locate it accurately.
The orphan technology is dangerous precisely because it can attract attention for a long time. It may win programmes, pilots and press while remaining between architectures. The founder keeps waiting for a prime, manufacturer, distributor or customer to complete the system. That partner often captures the interface and economics when it finally arrives.
India will not build durable deeptech companies by producing more orphan technologies, however sophisticated they are.
The domestic market can teach or deform
A large first market is usually presented as an uncomplicated advantage. I do not think it is.
A demanding domestic buyer can help a startup cross the Difficult Middle. The buyer can define a consequential mission, provide the operating environment, validate performance and create the first reference.
The same buyer can deform the company.
Requirements may change. Trials may not lead to orders. Payment may arrive long after acceptance. The first programme may require extensive customisation that does not travel. Public procurement may reward compliance with a process before repeat product economics. A strategic customer may insist on rights that prevent the company from selling elsewhere.
The founder therefore needs two designs at once:
1. A product designed to win the first mission.
2. A company designed to survive and learn from winning it.
The distinction matters. A prestigious first customer can still be a bad first market if the work creates no transferable architecture, qualification or data.
Cost is a discipline, not the moat
India's cost structure can allow more experiments, larger engineering teams and products designed for constrained environments. Cost discipline can also create solutions that travel to markets ignored by expensive incumbents.
But “cheaper from India” is not a sufficient global strategy.
Labour advantages narrow. Supply chains change. Competitors copy. Buyers will not accept lower reliability in systems where failure is expensive. A company that wins only because engineering hours cost less may be outbid or retained as a vendor rather than trusted as a product owner.
The lasting advantage has to become something harder to reproduce: field history, qualification, process yield, proprietary operating data, a supplier network, a design-in, a service system or an architecture that becomes easier to extend with each deployment.
Cost can help the company enter the Difficult Middle. Accumulated proof has to help it leave.
What the ecosystem actually needs to produce
If this thesis is right, ecosystem health should not be measured mainly by the number of startups, programmes, incubators or funding announcements.
Those are inputs.
I would look instead for movement through the same company-building sequence:
→ complete product or indispensable layer
→ accepted qualification
→ first consequential buyer
→ repeat deployment
→ strategic or platform expansion
This changes what institutions should optimise.
Research institutions should make knowledge and talent transferable without stripping the company of its future. Test facilities should produce evidence recognised by buyers. Procurement should distinguish development activity from accepted delivery and create paths to follow-on orders. Investors should fund the next proof rather than a fashionable category. Later-stage and project capital should arrive when repeat physical deployment becomes the bottleneck. Global buyers should engage Indian firms as owners of products and systems, not only as engineering capacity.
The number I want is not how many companies entered the funnel.
It is how many emerged with a position they could defend.
What would change my mind
I would become less confident in India's right to win if several patterns persist through the decade:
• Indian firms continue to provide engineering effort while foreign companies own the product, interface and customer.
• Programmes and pilots multiply without repeat procurement.
• Critical imported inputs remain uncontrolled at the moment of scale.
• Each deployment remains bespoke and qualification does not travel.
• Companies reach technical success but repeatedly fail for lack of working capital, test access or suitable deployment finance.
• Buyers praise performance but do not publish, contract around or pay for the resulting economics.
• Global customers buy talent and services from India but not Indian-owned systems, components and processes.
These would not mean that India lacks technical ability. They would mean that technical ability is failing to become strategic company power.
That distinction is the whole point.
Questions I now ask
1. Is this a national-capability need, a market opportunity or a company-level opportunity?
2. What gives an India-built company a specific right to win?
3. Is the company a complete challenger, an indispensable enabler or an orphan technology?
4. Which layer will it own after the first deployment?
5. Who can provide the demanding first mission and accept the evidence?
6. Does the first customer create reusable qualification or permanent custom work?
7. Which critical components, rights or interfaces remain outside the company's control?
8. Can the product be adapted for the next buyer rather than rebuilt?
9. What capital carries it through qualification, production and payment delay?
10. What evidence would show that the apparent right to win is disappearing?
India may have one of the world's most interesting combinations of difficult problems, technical talent, strategic demand and cost pressure.
That gives founders raw material. It does not give them victory.
The country becomes a proving ground only when solving its hardest missions produces evidence that travels—and companies that retain what they learn.