Healthcare Deep Tech: Building India’s Next Generation of Global Healthcare Leaders

The investment case rests on structural demand, a new generation of enabling technologies and India’s ability to engineer affordable, clinically useful products - provided companies build evidence and regulatory discipline early.

By Kaustubh Savant

**EXECUTIVE SUMMARY**

Build globally credible healthcare companies by solving hard problems, not by adding software to existing workflows

| **INVESTMENT PRINCIPLE** Back founders who can demonstrate a defensible technical edge, a consequential clinical problem, a credible path to adoption and economics that improve as the company scales. |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

## WHY THIS SECTOR

Healthcare combines non-discretionary human need with complex delivery systems. The global population aged 60 and over is projected to increase from 1.0 billion in 2020 to 1.4 billion in 2030 and 2.1 billion by 2050. Ageing is not the only demand driver: chronic disease, earlier diagnosis, long-term monitoring and more complex multimorbidity all increase the need for reliable, affordable care.

India’s problem is not simply that healthcare must become larger; it must become more productive. In 2022-23, out-of-pocket spending still represented 43.4% of total health expenditure. In 2023, noncommunicable diseases represented 71.1% of deaths in WHO’s India country profile. Those figures point to the importance of early detection, continuity of care and productive use of scarce clinical capacity.

## WHY NOW

The toolkit has improved across AI, sensors, imaging, sequencing, protein science, automation and biomanufacturing. The US FDA reported more than 1,600 AI-enabled medical devices authorised for marketing by September 2026, showing that AI-enabled devices are moving through regulated pathways at scale. At the same time, global digital-health funding reached US\$22.3 billion in 2025, up 19%, while deal count fell 9% to 1,474; US\$100 million-plus mega-rounds captured 44% of funding. The signal is not “AI wins”; it is that capital is concentrating around companies able to prove quality, adoption and durable differentiation.

## WHERE WE PLAY

Prioritise AI-enabled clinical workflows, precision and point-of-care diagnostics, specialised devices and robotics, and research or biomanufacturing tools. These areas can produce defensible products with measurable performance and a reachable buyer. Treat therapeutics selectively: scientific upside can be exceptional, but technical, clinical, manufacturing and financing risks compound over long timelines. Generic consumer-health applications without proprietary technology, defensible distribution or evidence should not receive deep-tech priority.

## WHAT WE LOOK FOR

The scorecard assesses technical differentiation, clinical problem, clinical evidence, market opportunity, founder and team, unit economics, global scalability and capital efficiency. Quantitative scores organise debate; they do not replace mandatory vetoes for patient safety, weak reproducibility, unclear intellectual-property ownership, an implausible regulatory pathway or no credible route to paying customers. Technical claims must survive external validation and comparison with current standard-of-care alternatives.

## HOW WE SUPPORT FOUNDERS

The fund’s practical contribution is to make the next proof point achievable: design clinical and technical validation, establish a fit-for-purpose quality system, select an appropriate regulatory pathway, convert pilots into contracted procurement, and choose first export markets on clinical fit and reimbursement rather than headline GDP. Capital tranches should reduce a named risk, with the next cheque linked to an agreed milestone instead of time alone.

| **CALL TO FOUNDERS ·** Start with a serious clinical problem. State what is scientifically different, show the evidence that would disprove your thesis, and build for global quality and regulatory expectations from day one. |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

*Scope and interpretation - This report is an investor thesis, not a clinical guideline, regulatory opinion or market-sizing report*.

**WHY THIS SECTOR · STRUCTURAL NEED**

Healthcare need is certain, but health systems cannot meet it efficiently

Demographic pressure and chronic disease make prevention, diagnosis and treatment more important; workforce, affordability and access constraints make productivity-enhancing technology valuable.

![Figure 1](healthcare-figure-1.png)

Exhibit 01. Global population aged 60+: historical baseline and WHO projections. \[1\]

| **SIGNAL** | **LATEST EVIDENCE** | **INVESTMENT IMPLICATION** |
|-------------------|-------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------------|
| **Ageing** | Global population aged 60+ rises from 1.0B (2020) to 1.4B (2030) and 2.1B (2050). \[1\] | Demand broadens for diagnostics, monitoring, chronic-care workflows, mobility and assistive technologies. |
| **Disease mix** | Noncommunicable diseases accounted for 71.1% of deaths in India in 2023 in WHO country data. \[3\] | Earlier risk detection and reliable longitudinal care may matter more than episodic, app-only engagement. |
| **Affordability** | Out-of-pocket expenditure was 43.4% of total health expenditure in India in 2022-23. Government health expenditure was 43.7%. \[4\] | Products must demonstrate total-cost impact and fit actual payer, provider and patient budgets. |
| **Workforce** | WHO projects a global shortfall of 11 million health workers by 2030, concentrated in lower-income settings. \[2\] | Automation and decision support should augment trained staff and improve throughput, not simply add another screen. |

## The bottleneck is often the pathway, not the absence of a clever product

Healthcare failures accumulate between the first symptom and the right intervention. A patient may be screened late, an abnormal result may not be followed up, specialist review may be unavailable, and a treatment plan may be fragmented across facilities. Each hand-off creates delay, cost and loss of information. Deep technology is valuable when it measurably changes one or more of these failure points: a test that performs accurately closer to the point of need; software that safely prioritises cases; a device that makes a difficult procedure reproducible; or a manufacturing tool that improves the consistency and cost of biologic production.

The most investable problems are not necessarily the most dramatic. A device that reduces repeat testing, a diagnostic that shortens time-to-treatment, or a clinical workflow tool that improves completion of follow-up can create substantial system value if the effect is measurable and the buyer captures part of it. The relevant unit of analysis is therefore not the feature or the model; it is the end-to-end clinical and economic pathway.

## India makes the constraint especially visible

India combines large and varied patient populations with considerable differences in specialist access, facility capability, affordability and infrastructure. This creates a demanding design environment: products must work across heterogeneous sites, users and operating conditions, often with limited training time and service capacity. A product that is both clinically reliable and operationally robust in this context may have a strong export proposition, but only if it also meets the target market’s regulatory, evidence, documentation and purchasing expectations.

Public financing is improving over the longer term, but the latest National Health Accounts estimate still shows material household exposure to healthcare costs. That makes affordability an engineering and business-model requirement rather than a branding line. Investors should ask which party pays, how the purchase is budgeted, whether savings accrue to the purchaser or another stakeholder, and how the company will prove the savings without overclaiming. \[4\]

## What structural demand does not prove

Demographics do not automatically create venture returns. Need may be served by public programmes, incumbent products, low-cost manual workarounds or solutions with difficult reimbursement. The investment filter must therefore combine burden with technical feasibility, a clear user and buyer, a credible evidence plan and a path to repeatable sales. Population statistics establish relevance; company-level evidence establishes investability.

| **SO WHAT** The best opportunities remove a consequential clinical or delivery bottleneck and demonstrate who benefits, who pays and why the improvement persists after the pilot ends. |
|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

*Source / method: The demographic series is a WHO global estimate/projection. India health financing and causes-of-death figures refer to the years explicitly stated, rather than a single common reporting year. \[1-4\]*

**WHY THIS SECTOR · DEEP TECH VS. SERVICES**

Deep tech can compound value because differentiation is embedded in technology and evidence

India’s cost and engineering advantages matter most when they support a product whose performance, IP, workflow integration or manufacturing process is difficult to reproduce.

Service-led healthcare and digital health are important parts of the system, but they often scale through more clinicians, beds, locations, account managers or user acquisition. Product-led deep tech has a different potential: once a product works, is validated and is supported by a repeatable quality and service model, each additional deployment can reuse much of the underlying technology. That operating leverage is real only after the company earns trust, completes validation and makes the product fit clinical practice.

| **DIMENSION** | **SERVICE-LED HEALTHCARE** | **GENERIC DIGITAL HEALTH** | **HEALTHCARE DEEP TECH** | **INDIA STARTING ADVANTAGE** |
|-----------------------------|--------------------------------------------------------------------|------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------|
| **Value source** | Clinical time, facilities and care delivery. | Aggregation, interface, workflow and distribution. | Proprietary IP, engineered performance, validated algorithms, materials, assays or platforms. | Engineering talent, large and diverse use cases, process know-how and cost-sensitive design. |
| **Differentiation** | Local reputation, clinician relationships and operating execution. | UX, acquisition, convenience and integrations. | Measured clinical performance, hard-to-reproduce design, regulatory evidence, know-how and quality systems. | Pharma, diagnostics, manufacturing and software experience can be combined into global product teams. |
| **Scaling constraint** | Recruitment, physical capacity and site-level consistency. | Customer acquisition, retention, workflow fit and platform dependence. | Validation, manufacturing yield, regulatory approval/clearance, service and evidence generation. | Design-to-cost and ability to test across varied contexts, if quality is maintained. |
| **Core diligence question** | Can care delivery expand without losing quality? | Can retention and monetisation support acquisition costs? | Does the technical edge produce safer, better or lower-cost clinical outcomes at scale? | Can the product meet global quality standards and sell outside its home market? |

## The moat must be specific and observable

“Deep tech” is not a synonym for complexity, a patent count or a research-heavy pitch deck. A moat should be expressed in observable assets and performance: a validated assay with a reproducible workflow; a sensor architecture with superior signal quality under real-world conditions; a model with robust external validation and a governed update process; a robotic system with safe, repeatable precision; or a bioprocess that produces consistent quality at a commercially relevant yield and cost. The company must explain what competitors would need to reproduce, what it would cost them, and how long replication would take.

A patent may support that argument, but patents are not sufficient. Freedom to operate, trade secrets, manufacturing process knowledge, clinical data rights, distribution relationships, installed-base learning and regulatory history can all contribute. The weight of each differs by category. Software may be defensible when embedded in a regulated workflow, supported by proprietary data rights and validated across sites; hardware may be weakly defensible if it relies on commoditised components and easily copied assembly.

## Design-to-cost is not “cheap at any cost”

India’s cost environment can encourage architecture-level simplification: fewer parts, easier calibration, lower consumables burden, modular hardware, automated quality checks and reduced dependence on highly specialised operators. These are sources of advantage only when clinical performance, reliability and service life remain acceptable. A lower price that increases false negatives, device downtime or repeat testing is not a system-level saving.

## Decentralisation is attractive when the use case supports it

Point-of-care screening can reduce travel and specialist bottlenecks, but moving a test closer to the patient shifts responsibilities to training, sample handling, calibration, quality assurance and referral. The investment case must specify which steps are decentralised, which stay in a laboratory or specialist centre, and how discordant results are resolved. A decentralised test is an end-to-end pathway, not just a smaller instrument.

## The right comparison is value per clinically meaningful outcome

Service and product models should not be treated as mutually exclusive. A service-led company may be the right route to learn a workflow or establish distribution; a product company may still require installation, training, maintenance and clinical support. The fund should favour architectures in which service revenue enables repeatable product deployment rather than hiding a labour-intensive business under a software multiple. Evidence of gross-margin improvement, reduced implementation time and repeat purchases is more persuasive than a claim of “scalability.”

| **SO WHAT** Commercially attractive deep tech solves a costly or consequential constraint and protects the benefit through verified performance, know-how, quality and a business model that scales. |
|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

*Source / method: This comparison is an investment framework, not a claim that every service-led or software-led model lacks defensibility. India’s National Medical Devices Policy explicitly positions global competitiveness, innovation, quality and affordability as sector goals. \[9\]*

**WHY NOW · CONVERGENCE & CAPITAL SELECTIVITY**

Technology is advancing, but funding increasingly rewards proof over promise

AI, biology and precision engineering are becoming useful in regulated settings; the investable company still needs clinical validation, trusted deployment and commercial pull.

## Exhibit 02 · 2025 digital-health funding shifts

| **INDICATOR** | **2025 RESULT** | **WHAT IT SUGGESTS** |
|-----------------------------------|-----------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------|
| **Global digital-health funding** | US\$22.3B; +19% year on year. \[5\] | Capital was available, but aggregate volume alone does not identify the winning subsegments. |
| **Deal count** | 1,474; down 9% year on year. \[5\] | Funding was spread across fewer transactions, consistent with more selective deployment. |
| **US\$100M+ mega-rounds** | 44% of total funding. \[5\] | Large rounds absorbed a major share of capital; earlier companies must show a credible path to differentiated outcomes. |
| **AI-enabled devices** | FDA reported more than 1,600 authorised devices as of September 2026. \[6\] | AI is present in multiple regulated device categories; authorisation count is not a substitute for post-market performance or adoption. |

## Three forces are converging

### 1. Clinical-grade AI is becoming more operational

Clinical AI has moved beyond general demonstrations into defined intended uses such as imaging analysis, signal interpretation, triage and selected workflow support. The FDA’s AI-enabled device list provides a public record of devices authorised for marketing in the United States and is updated periodically. Its value for investors is as a signal that regulated pathways exist, not proof that a new product will clear review or improve outcomes. The list is also not comprehensive. \[6\]

The diligence question is therefore specific: what is the intended use, what is the comparator, which subgroup and setting are represented in the evidence, how will the model behave after deployment, and how will changes be controlled? In healthcare, data drift, external validity, explainability appropriate to risk, human factors, cybersecurity and monitoring are product requirements, not later-stage governance details. India’s ICMR has published ethical guidelines for AI in biomedical research and healthcare covering principles, ethics review, governance and informed consent. \[11\]

### 2. Biological toolkits and biomanufacturing are expanding the design space

Sequencing, protein-structure methods, high-throughput experimentation, improved assays and computational biology can reduce the time and cost of selected research cycles. But a faster experiment is not equivalent to a successful drug. The translation gap still includes target validity, reproducibility, safety, manufacturability, clinical benefit and reimbursement. Investment should favour tools that remove a repeated bottleneck across multiple customers or programmes, or therapeutics that have a tightly defined scientific and clinical milestone path.

India’s BioE3 policy supports innovation and entrepreneurship in high-performance biomanufacturing and envisages Bio-AI hubs, biomanufacturing hubs and biofoundry capacity. This policy direction may improve the shared-infrastructure ecosystem, but it does not establish that every startup can access a specific facility on commercial terms or that a new process is ready for scale. Each company must evidence facility access, process transfer, batch consistency, quality controls and unit cost at relevant scale. \[13\]

### 3. Sensors, optics, robotics and manufacturing precision are improving

Smaller sensors, better optics, improved components and more capable embedded compute can support portable diagnostics, monitoring, imaging, assistive devices and automation. The investment advantage appears when the product removes workflow friction or yields clinically meaningful performance in realistic use. Hardware diligence must include calibration, durability, cleaning, consumables, installation, cybersecurity where relevant, service coverage and replacement cycles. A promising laboratory prototype is not a validated deployable system.

## India’s digital public infrastructure is a potential enabler - with limits

As of 5 August 2025, the Ministry of Health and Family Welfare reported approximately 79.9 crore Ayushman Bharat Health Accounts (ABHAs), 4.18 lakh registered health facilities, 6.79 lakh registered healthcare professionals and 67.19 crore health records linked with ABHA. These figures illustrate the scale of digital-health infrastructure. They do not, by themselves, demonstrate that all records are complete, interoperable, clinically useful, consented for a startup’s intended use or available for model training. Product teams still need lawful access, data quality, representativeness and careful integration design. \[10\]

The WHO Global Strategy on Digital Health was extended through 2027, reinforcing the importance of aligning digital tools with health-system capacity, finance, workforce and governance rather than treating technology as a standalone solution. \[12\]

## Capital selectivity: how to interpret the signal

The data do not show that investors have stopped funding digital health. They show funding increasing while the number of deals declined and capital concentrated in mega-rounds. For seed and early-stage investors, the practical implication is a higher bar for differentiation and evidence: prove a real user and buyer, establish why the technology is hard to copy, demonstrate a repeatable path to validation and show how the next round will be earned. “AI-enabled” is a category label, not an investment thesis.

| **SO WHAT** Back verifiable clinical and technical moats, but underwrite the route from prototype to safe deployment, procurement and post-market learning. Technological convergence improves what can be built; it does not remove healthcare’s proof burden. |
|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

*Source / method: Digital-health investment data are CB Insights’ global private-market classification for 2025. FDA device count refers to US marketing authorisations listed by the agency, not all AI tools or a universal measure of clinical benefit. \[5, 6\]*

**WHERE WE PLAY · OPPORTUNITY MATRIX**

Five themes combine technical differentiation, global relevance and an investable path to proof

Prioritise opportunities that can show measurable improvement within a manageable development cycle; retain selective exposure to longer-duration biotech where the science and capital plan justify it.

| **OPPORTUNITY SEGMENT** | **DIFFERENTIATION\*** | **GLOBAL SCALE\*** | **DEVELOPMENT BURDEN** | **FUND PRIORITY** |
|---------------------------------------------|-----------------------|--------------------|------------------------|-------------------------------|
| **AI clinical workflows** | 4/5 | 5/5 | Medium–High | High priority |
| **Precision and point-of-care diagnostics** | 4/5 | 4/5 | Medium–High | High priority |
| **Devices and specialised robotics** | 4/5 | 4/5 | High | High priority |
| **Research tools and biomanufacturing** | 4/5 | 4/5 | Medium–High | High priority |
| **Biotech and therapeutics** | 5/5 | 5/5 | Very high | Selective |
| **Generic apps / consumer health** | 2/5 | 4/5 | Low–Medium | Deprioritized, if moat absent |

*Source / method: \* Directional investment hypotheses from the framework, not externally validated rankings or allocations. Re-score by clinical indication, evidence burden, buyer, regulatory class and company stage.*

## 1. AI-enabled clinical workflows

Most attractive: products that integrate into a defined clinical workflow and change a measurable action - for example, identifying time-sensitive findings, improving documentation quality or reducing avoidable administrative work. Favour products with a narrow intended use, robust local and external validation, clear human oversight, auditable performance and a buyer that captures the productivity benefit. Avoid generic copilots whose differentiation is limited to a foundation-model wrapper or whose business case assumes clinicians will change behaviour without implementation support.

Proof points: prospective or well-designed retrospective validation appropriate to the intended use; comparison with current practice; subgroup performance; workflow time and error metrics; data rights; integration burden; adoption and renewal; documented change control; cybersecurity and privacy posture. The revenue model should align with hospital budgets, usage, measured value or a defined per-site deployment rather than pricing alone.

## 2. Precision and point-of-care diagnostics

Attractive opportunities include assays, readers and integrated systems that reduce time-to-answer, improve analytical performance, enable earlier intervention or make testing viable in settings where laboratory access is limited. The product’s total system includes sample collection, preparation, operator training, quality control, calibration, consumables, interpretation, confirmation and referral. Unit economics should include repeat tests, wastage, service and field performance - not just the reagent or instrument cost.

Proof points: analytical sensitivity and specificity where applicable; reproducibility across lots and sites; clinical sensitivity and specificity in the target population; comparison with an accepted reference; invalid-test rates; operational performance under intended conditions; clinical utility and willingness to pay. A strong benchmark is not “better than no test”; it is better than the best realistic alternative in the intended market.

## 3. Devices and specialised robotics

Look for equipment and robotics that materially improve precision, safety, access, throughput or repeatability. Strong product designs anticipate procurement, maintenance, training, consumables, cleaning, uptime, sterilisation where applicable and field-support economics. Engineering excellence alone is insufficient if installation is onerous or the product depends on scarce specialists. The best teams demonstrate manufacturability and serviceability while the design is still evolving.

Proof points: bench and simulated-use testing, human-factors evidence, safety and performance studies appropriate to risk, design controls, manufacturing yield, reliability, service-response time and a credible regulatory strategy. Identify the predicate or novelty profile early; the US FDA may route devices through 510(k), De Novo or PMA depending on classification and risk, while India uses risk-based medical-device rules. \[14, 15\]

## 4. Research tools and biomanufacturing

Enabling technologies can sell to multiple pharma, biotech, diagnostics or academic customers and may diversify risk across programmes. Attractive categories include tools that improve target identification, measurement, sample throughput, process control, assay development, analytical characterisation, automation, scale-up or manufacturing consistency. A tools company can be more capital-efficient than a therapeutics company only when software or service revenue are repeatable.

Proof points: benchmarked performance against established methods; reproducibility across operators and sites; evidence that customers will change a validated workflow; instrument utilisation; consumables or service attach rates; field reliability; and economics at the customer’s real process scale. For biomanufacturing, ask what happens to yield, quality attributes, contamination risk, cycle time and cost when moving from bench to pilot and commercial scale.

## 5. Selective biotech and therapeutics

Therapeutics offer potentially exceptional intellectual-property and clinical value, but the capital and time needed to establish safety, efficacy and manufacturing readiness can be substantial. Invest selectively where biology is compelling, the target product profile is explicit, the team understands translational risk and the next experiments can change a high-value decision. A patent around an unvalidated target or an elegant preclinical result should not be treated as clinical validation.

Proof points: target and mechanism rationale, reproducible preclinical data, relevant disease models, translational biomarkers, CMC feasibility, safety plan, IP and freedom-to-operate analysis, regulatory engagement, trial design and financing to the next decisive milestone. Use milestone-linked tranches and pre-agreed stop criteria. Avoid financing a broad platform narrative without evidence for a lead programme and a credible path to partner, license or commercialise.

## Why generic consumer health is not automatically deep tech

A large addressable audience does not equal a defensible company. Consumer health can be attractive where it owns a trusted distribution channel, proves outcomes, creates a regulated product or possesses distinctive data rights and sustained retention. Without those assets, functionality is easily copied and customer-acquisition costs can absorb gross margin. Such businesses should be assessed on their own merits rather than receiving deep-tech priority because they use AI or health data.

| **SELECTION PRINCIPLE** Prioritise measurable differentiation and a buyer who values it. Prefer categories where the next milestone can create credible evidence and improve commercial learning before the company commits to a long, expensive development path. |
|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

| **THERAPEUTICS DISCIPLINE** Scientific upside can be high while the probability of success, timing and financing requirement remain uncertain. Stage capital around specific evidence gates, not broad ambition. |
|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

**WHERE WE PLAY · TRADE-OFF ARCHITECTURE**

Lighter development burden can buy faster proof; therapeutics earn selectivity, not a free pass

Milestone timing, capital intensity and revenue velocity are directional planning variables. They vary sharply with indication, product risk, evidence design, regulatory route, CMC and geography.

| **THEME** | **ILLUSTRATIVE MILESTONE CYCLE\*** | **INDICATIVE CAPITAL ENVELOPE\*** | **REVENUE VELOCITY** | **PRIMARY DOWNSIDE** |
|----------------------------|-----------------------------------------------------------------------------------|-----------------------------------|--------------------------------------------------------|--------------------------------------------------------------------|
| **AI clinical workflows** | 12–18 months to meaningful validation / deployment evidence | US\$3–8M | Potentially faster after integration and procurement | No workflow change; poor external validity; unclear buyer ROI. |
| **Precision diagnostics** | 18–36 months to robust validation / regulatory milestone | US\$5–15M | Moderate; affected by test volume and distribution | High cost per valid result; quality drift; consumables economics. |
| **Devices and robotics** | 24–48 months to relevant design validation / market milestone | US\$10–25M | Moderate; hardware and service cycles matter | Manufacturing, maintenance, clinical adoption and working capital. |
| **Biotech / therapeutics** | 48–84 months to a major early clinical value inflection; approval may take longer | US\$25–60M+ | Slow; often no product revenue during core development | Clinical failure, CMC risk, long funding gaps and dilution. |

*Source / method: \* Indicative ranges retained and elaborated from the original thesis. They are not verified market averages or total cost to product approval.*

## India’s bioeconomy provides breadth, not a healthcare-only market size

| **INDICATOR** | **DATA** | **INTERPRETATION** |
|----------------------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **India bioeconomy** | US\$195.3B in 2025, compared with US\$165.7B in 2024; +18% year on year. \[7\] | Broad aggregate across health and non-health sectors. Do not use it as the total addressable market for a healthcare startup. |
| **Medical-device exports** | ₹42,360 crore in FY2024-25. \[8\] | Evidence of export activity and manufacturing capability across the category, not proof that every product segment is globally competitive. |
| **Medical-device imports** | ₹1,37,088 crore in FY2024-25. \[8\] | Imports were about 3.2 times exports; the implied gap was ₹94,728 crore. This points to areas for investigation but is not a directly addressable startup market. |
| **Policy direction** | National Medical Devices Policy 2023 targets an innovative, globally competitive sector and sets an aspirational ambition for a 10–12% share of the expanding global market over 25 years. \[9\] | A policy aspiration is not a forecast. Winning companies still require quality, performance, pricing, regulatory access and distribution. |

![Figure 2](healthcare-figure-2.png)

Exhibit 03. Indian medical-device imports and exports, FY2024-25. Source: DGCIS data reported by the Department of Pharmaceuticals / PIB. \[8\]

## Capital should follow risk reduction, not calendar time

For a product company, the relevant question is what uncertainty each rupee removes. A technical tranche may prove performance in a controlled environment; the next may validate reproducibility across operators or sites; a later tranche may establish clinical utility, quality-system readiness and procurement economics. For therapeutics, the gates will be biological and translational before they become clinical and CMC milestones. The tranche size should reflect the cost of the next proof point, the probability that it changes the investment case and the likelihood of securing a credible financing partner afterward.

Do not interpret a faster validation cycle as guaranteed revenue. Hospitals may require budget approval, integration, clinical governance and procurement cycles after technical validation. Diagnostics require channel access, quality controls and recurring consumables. Devices need service and replacement plans. A therapeutic’s value may increase sharply after a positive data readout, but failure can erase most programme value. Portfolio design should account for these different risk shapes rather than compare companies solely on headline time to first revenue.

## Portfolio construction implications

Shorter-cycle categories can produce operating feedback earlier and may support a portfolio learning loop across regulatory planning, clinical partnerships and procurement. The fund should still avoid spreading capital too thinly across unrelated categories. Common playbooks are useful only where the underlying milestones and buyers are genuinely comparable. Reserve follow-on capital for companies that meet evidence and commercial thresholds, and be willing to stop funding a thesis when evidence contradicts it.

| **SO WHAT** Use short-cycle businesses to learn quickly and scale validated products; reserve long-duration therapeutic capital for programmes where the science, milestones and financing syndicate justify the risk. |
|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

**WHAT WE LOOK FOR · SCORECARD & GATES**

An explicit scorecard makes trade-offs visible; non-negotiable gates protect scientific and clinical integrity

The score is a structured discussion aid. Patient safety, reproducibility, IP ownership, regulatory feasibility and a credible buyer are mandatory gates, not items that can be compensated for by a high total score.

## Exhibit 04 · Eight-dimension diligence scorecard

| **DIMENSION** | **WEIGHT** | **QUESTIONS TO ANSWER** | **WHAT STRONG EVIDENCE LOOKS LIKE** |
|-------------------------------|------------|---------------------------------------------------------------------|---------------------------------------------------------------------------------------------------------------------|
| **Technical differentiation** | 20% | What is difficult to reproduce? Which benchmark proves superiority? | External benchmark, relevant IP / know-how, defensible data or process, reproducibility. |
| **Clinical problem** | 15% | How severe, frequent and costly is the problem? What happens today? | Clinician and patient evidence, current pathway, unmet need and quantified consequence. |
| **Clinical evidence** | 15% | Does performance generalise to intended users and settings? | Appropriate comparator, pre-specified endpoints, credible sample, representative sites and transparent limitations. |
| **Market opportunity** | 15% | Who is the user, buyer and payer? Can the company reach them? | Bottom-up market model, named buyer segments, procurement route, alternatives and willingness-to-pay evidence. |
| **Founder and team** | 15% | Can this team bridge science, product, regulation and sales? | Relevant technical depth, clinical/regulatory leadership, execution track record and coachability. |
| **Unit economics** | 10% | Can contribution margin improve with scale? | Costed bill of materials / assay, gross margin bridge, service and channel costs, credible utilisation assumptions. |
| **Global scalability** | 5% | What must change by market? Is the use case exportable? | Regulatory map, cross-market need, localisation plan, distribution path and competitive benchmarks. |
| **Capital efficiency** | 5% | What proof point does the next round buy? | Milestone-linked budget, runway, downside plan, financing milestones and explicit stop criteria. |

## Non-negotiable vetoes

- Patient-safety risk that the design cannot reasonably mitigate, an intended use that is not responsibly framed, or a team unwilling to surface adverse data.

- Results that cannot be reproduced, a weak comparator, data leakage, cherry-picked subgroups or conclusions that exceed the study design.

- Unclear ownership of background IP, employee/academic assignment gaps, unresolved third-party rights or no credible freedom-to-operate workstream.

- No defined regulatory route, no plan for quality management and post-market responsibilities, or a business plan that treats the “pilot” as proof of clinical efficacy.

- A top-down TAM derived from total health expenditure, a product with no identified buyer/payer, or a sales model that assumes unpaid pilots will convert automatically.

## Stage gates: what must be true before more capital is committed

| **GATE** | **EVIDENCE REQUIRED** | **DECISION** |
|--------------------------------|------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------|
| **1. Discovery** | Clinical problem verified; alternatives mapped; intended use defined; basic scientific premise plausible. | Proceed only when the problem and technical hypothesis are specific. |
| **2. Prototype** | Bench performance meets pre-agreed criteria; key technical risks listed; IP and data rights assessed. | Fund focused development, not feature sprawl. |
| **3. Validation** | Appropriate external or multi-site evidence, relevant comparators, defined limitations and a regulatory plan. | Proceed if evidence supports the intended claim and risk is manageable. |
| **4. Early commercialisation** | Paid or contractually committed deployments; training, service, quality and procurement workflows tested. | Scale only when usage and value persist beyond founder-led pilots. |
| **5. Scale-up** | Repeatable sales, reliable manufacturing or deployment, gross-margin path, quality controls and working capital. | Release expansion capital based on demonstrated repeatability. |
| **6. Global expansion** | Target-market regulatory plan, local clinical fit, reimbursement/buyer logic, service partner and distribution. | Enter markets where evidence and operating model transfer credibly. |

## Diligence red flags that deserve deeper testing

A company’s strongest slide is often the one that deserves the most scrutiny. “First” or “only” claims need a search of prior art and commercial alternatives. Benchmark results should use active comparators and an appropriate ground truth. Model accuracy needs context: prevalence, predictive values, calibration, subgroup performance and real workflow consequences may alter clinical meaning. A prototype’s manufacturing cost should include yield, rejected units, installation, maintenance, consumables, regulatory testing and warranty. A large TAM should be rebuilt from the number of reachable buyers, annual use, pricing and likely adoption constraints.

| **SO WHAT** The scorecard structures debate; hard gates control the decision. A strong technical narrative cannot offset weak evidence, unclear rights, an unviable buyer or a capital plan that cannot reach the next proof point. |
|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

**HOW WE SUPPORT · TRANCHED PARTNERSHIP**

Milestone-linked support and capital; choose the first market on fit, not headline size

The objective is to de-risk technical, clinical, regulatory and commercial uncertainty in a sequence that preserves optionality and builds institutional credibility.

## Hands-on support framework

| **SUPPORT WORKSTREAM** | **WHAT THE COMPANY NEEDS** | **WHAT GOOD EXECUTION LOOKS LIKE** |
|---------------------------------------|-------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **Clinical and technical validation** | A protocol that answers the investment-critical question, uses appropriate comparators and measures the right endpoint. | Pre-specified analysis, justified sample size, representative sites and users, transparent limitations, external review and documented data integrity. |
| **Regulatory roadmapping** | A clear intended use, classification rationale, evidence plan, QMS requirements and post-market obligations. | A pathway-specific plan for CDSCO, US FDA and/or the target jurisdiction; design controls and quality-system work begin early, not after the prototype is frozen. |
| **Commercial procurement** | A clinical champion and a budget owner; a pilot design linked to a purchase decision. | Written success criteria, paid or strategically justified pilot, procurement timeline, implementation cost and a defined conversion decision. |
| **Global market selection** | Evidence that the product fits a specific clinical pathway and buyer, and can be supported locally. | Priority market selected by clinical urgency, regulatory fit, reimbursement, distribution, service logistics and competitive alternatives. |
| **Manufacturing and service** | Reliable production, quality checks, field maintenance, user training and supply continuity. | Documented design transfer, supplier quality, yields, calibration, service turnaround, installed-base tracking and working-capital plan. |
| **Partnership and financing** | Milestones that external partners or the next investors will underwrite. | A concise data room, milestone-linked budget, well-defined IP, clinical/regulatory evidence and transparent downside scenarios. |

## Build the regulatory roadmap around intended use

Regulatory strategy starts with what the company claims the product does, for whom, and under what conditions. A diagnostic, clinical decision-support tool, autonomous image-analysis function, surgical device and therapeutic product have different evidence and regulatory requirements. The US FDA’s device pathways include 510(k), De Novo and PMA, selected based on device classification, risk and predicate availability. India’s Medical Devices Rules use risk-based classes A through D, and the European Union’s Medical Device Regulation establishes obligations for conformity assessment, clinical evaluation, quality management and post-market surveillance. These routes are not interchangeable; a company needs jurisdiction-specific advice and a product-specific plan. \[14, 15, 16\]

For AI-enabled medical devices, change control deserves special attention. If model performance changes with new data or updates, the company needs a documented approach to versioning, validation, monitoring and deciding when changes trigger a new submission or assessment. Regulation is not merely a launch cost: designed well, quality and evidence become trust infrastructure that supports hospital adoption, enterprise contracting and entry into demanding markets. \[6, 11, 16\]

## A practical milestone-linked financing architecture

| **TRANCHE / PHASE** | **CAPITAL SHOULD BUY** | **RELEASE CRITERIA** | **STOP / REDESIGN TRIGGER** |
|-------------------------------------|-----------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------------|
| **A. Feasibility** | Proof that the central technical premise is plausible. | Bench performance meets pre-agreed criteria and major failure modes are understood. | Results not reproducible, no material advantage over existing alternatives. |
| **B. Validation** | Evidence in relevant samples, sites or intended users. | Comparator, endpoint, protocol and analysis are credible; external expert review is satisfactory. | Performance falls below target, intended use must narrow materially, or safety concerns cannot be mitigated. |
| **C. Regulatory and deployment** | Quality systems, submission readiness, implementation and service model. | Regulatory route is confirmed; design controls and documentation are on track; pilots meet defined criteria. | Regulatory pathway changes the economics, or implementation burden overwhelms customer value. |
| **D. Repeatable commercialisation** | A repeatable route to paid deployments and positive contribution economics. | Paid conversion, retention/reorder, validated ROI, predictable procurement and service. | Pilots remain unpaid, usage declines, buyers do not own the budget or gross-margin path fails. |
| **E. Expansion** | Capacity to scale quality and reach additional markets. | Repeatable manufacturing/implementation, local regulatory plan, channel and service capacity. | Quality deteriorates, working-capital needs exceed plan or market-entry economics do not hold. |

## Choose the first market on clinical fit over size

- Clinical fit: target a setting where the disease burden, workflow problem and performance benefit are clear enough to support adoption.

- Predictable payment: establish the actual payer and procurement route. Confirm whether payment is fee-for-service, a bundled episode, a hospital capital budget, a lab budget, a payer contract or direct patient payment.

- Service and logistics: ensure training, calibration, spare parts, consumables, maintenance and response times are deliverable with startup resources.

- Regulatory portability: prioritise geographies where the product’s evidence package and intended use can credibly transfer; map local clinical standards and data-protection obligations before committing to launch.

- Reference value: select early customers who can produce credible evidence and references, but do not confuse a prestigious logo with repeated commercial demand.

## Turn pilots into procurement

A pilot needs a written decision contract: baseline, primary outcomes, minimum performance threshold, duration, implementation responsibilities, data governance, price after pilot and the person who signs the purchase order. Paid pilots are stronger evidence of demand than free pilots, though a free pilot can sometimes be justified if it unlocks indispensable validation and has explicit conversion criteria. In either case, “the clinician likes it” is insufficient without a budget and procurement path.

| **SO WHAT** Each round should reduce an agreed, measurable axis of technical or clinical risk. The fund’s value is not only capital; it is helping founders sequence evidence, regulatory readiness and buyer adoption so that the next financing decision is earned. |
|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|

**CALL TO FOUNDERS · THE BAR FOR LEADERSHIP**

Show what is broken, what is different and why it will scale

Great healthcare founders combine technical ambition with intellectual honesty about what remains unproven. The global thesis must be present at the architecture stage, not appended after domestic traction.

## 1. What is broken?

Describe the clinical problem in the context of current care. Who experiences it, how frequently does it occur, what is the cost of failure, and what do clinicians or patients do today? Distinguish a visible inconvenience from a clinically meaningful problem. Quantify the consequences where possible: missed or delayed diagnosis, repeat testing, time to treatment, avoidable procedure risk, poor follow-up, staff burden or manufacturing inconsistency. Explain which stakeholder feels the pain and which stakeholder has a reason to pay for its resolution.

## 2. What is different?

Explain the technical insight in language a clinical buyer and a scientific reviewer can both test. Identify the benchmark, the current best alternative and the performance gap. State what is validated, what is only simulated or preclinical, which populations or settings have not been tested, and what result would disprove the claim. Describe the defensibility: patents, know-how, proprietary and lawfully controlled data, manufacturing process, regulatory evidence, distribution or embedded workflow - without assuming any one of these automatically creates a moat.

## 3. Why will it scale?

Show how the product can be manufactured, installed, maintained and sold repeatedly. Name the user, clinical champion, budget holder, payer or procurement owner. Explain pricing and the buyer’s economic return. Map the relevant regulatory pathway, reimbursement or payment mechanism, quality system, clinical evidence and post-market responsibilities. Explain what changes when moving from India to another market, and why the core product remains valuable after those changes.

## Readiness checklist

| **CHECK** | **QUESTIONS THE DECK SHOULD ANSWER** | **EVIDENCE TO INCLUDE** |
|-------------------------|------------------------------------------------------------------|------------------------------------------------------------------------------------------|
| **Clinical problem** | What is broken, for whom and how consequential is it? | Workflow map, current alternatives, clinician/patient insight and quantified burden. |
| **Alternatives** | What does the customer use today, including manual workarounds? | Competitive matrix, active comparators, incumbent and substitute analysis. |
| **Differentiation** | What is technically new and difficult to copy? | Benchmark data, IP position, reproducibility and clear limitation statements. |
| **Validation plan** | What proof is needed before the next capital milestone? | Protocol, endpoints, sample/site logic, comparator and independent reviewers. |
| **Regulatory strategy** | What claims are intended and what pathway applies? | Classification rationale, quality plan, data governance and post-market approach. |
| **Commercial model** | Who buys, who pays and why now? | Named buyer roles, procurement route, price, ROI and paid-conversion evidence. |
| **Global thesis** | Which international markets fit first and why? | Country prioritisation, local pathway, evidence transfer, distribution and service plan. |
| **Capital plan** | What does each tranche prove and what if the result is negative? | Milestone budget, runway, stop criteria and next-financing logic. |

## The founding mindset investors reward

A credible founder does not present uncertainty as a weakness. They separate demonstrated fact from hypothesis, identify the experiments that can resolve uncertainty, and disclose results that do not fit the story. They know when to narrow the intended use, redesign the product, seek a clinical partner or stop a programme. In healthcare, trust is built by disciplined evidence and transparent limitations as much as by ambition.

India’s opportunity is not to make a less expensive copy of an established product by accepting lower standards. It is to build products whose architecture, validation and cost structure solve real problems in India and remain compelling in global markets. That requires scientific differentiation, rigorous quality, regulatory fluency, thoughtful commercialisation and teams able to execute across disciplines.

| **SO WHAT** India’s next healthcare leaders will be defined by the problems they solve and the standards they meet - not by how novel the pitch deck sounds. |
|--------------------------------------------------------------------------------------------------------------------------------------------------------------|

**REFERENCES & SOURCE NOTES**

Primary sources and current evidence used in this report

Sources checked on October 2026. Links lead to the publisher or official government source wherever available.

\[1\] World Health Organization. “Ageing and health.” Fact sheet, 1 October 2025. Demographic estimates and projections for populations aged 60 and over. [<u>Open source</u>](https://www.who.int/en/news-room/fact-sheets/detail/ageing-and-health)

\[2\] World Health Organization. “Health workforce.” Overview; projection of an 11 million global health-worker shortfall by 2030. [<u>Open source</u>](https://www.who.int/health-topics/health-workforce)

\[3\] World Health Organization. “India - WHO Data.” Country profile and cause-of-death distribution; 2023 data shown on the profile. [<u>Open source</u>](https://data.who.int/countries/356)

\[4\] Ministry of Health and Family Welfare, Government of India / Press Information Bureau. “Union Health Ministry Releases the National Health Accounts Estimates for India 2022-23.” 27 May 2026. Government health expenditure and out-of-pocket expenditure as shares of total health expenditure. [<u>Open source</u>](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2265816&lang=1&reg=3)

\[5\] CB Insights. “State of Digital Health 2025 Report.” 22 January 2026. Funding, deal-count and mega-round metrics for global digital health in 2025. [<u>Open source</u>](https://www.cbinsights.com/research/report/digital-health-trends-2025/)

\[6\] US Food and Drug Administration. “Artificial Intelligence-Enabled Medical Devices.” FDA list and overview; more than 1,600 devices authorised for marketing as of September 2026. The FDA notes that the list is not comprehensive. [<u>Open source</u>](https://www.fda.gov/medical-devices/digital-health-center-excellence/artificial-intelligence-enabled-medical-devices)

\[7\] Biotechnology Industry Research Assistance Council (BIRAC), Government of India. “India Bioeconomy Report 2026.” 2025 estimate of US\$195.3 billion; broad bioeconomy scope across healthcare and other sectors. [<u>Open source</u>](https://www.birac.nic.in/webcontent/IBER_2026.pdf)

\[8\] Department of Pharmaceuticals, Government of India / Press Information Bureau. “Import and Export of Medical Devices.” 13 March 2026. DGCIS data for FY2024-25 and previous years. [<u>Open source</u>](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2239563&lang=1&reg=3)

\[9\] Department of Pharmaceuticals, Government of India. “National Medical Devices Policy 2023 - Strategy Document.” Policy aims, sector priorities and long-horizon ambition. Policy goals are not market forecasts. [<u>Open source</u>](https://www.pharma-dept.gov.in/sites/default/files/Strategy%20Document%20on%20NMDP%202023_0.pdf)

\[10\] Ministry of Health and Family Welfare, Government of India / Press Information Bureau. “Update on Ayushman Bharat Digital Mission.” 12 August 2025. ABHA, facility, professional and linked-record counts as of 5 August 2025. [<u>Open source</u>](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2155449&lang=2&reg=48)

\[11\] Indian Council of Medical Research. “Ethical Guidelines for Application of Artificial Intelligence in Biomedical Research and Healthcare.” 2023. Ethical, governance, informed-consent and review framework. [<u>Open source</u>](https://www.icmr.gov.in/ethical-guidelines-for-application-of-artificial-intelligence-in-biomedical-research-and-healthcare?q=ICS)

\[12\] World Health Organization. “Global Strategy on Digital Health 2020-2027.” Published 1 December 2025 following extension through 2027. [<u>Open source</u>](https://www.who.int/publications/i/item/9789240116870)

\[13\] Press Information Bureau, Government of India. “Cabinet approves BioE3 Policy for Fostering High Performance Biomanufacturing.” 24 August 2024. Official summary of BioE3 policy and planned hubs / biofoundry approach. [<u>Open source</u>](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2048569&lang=2&reg=3)

\[14\] US Food and Drug Administration. “Medical Device Safety and the 510(k) Clearance Process.” Overview of 510(k), De Novo and PMA pathways and risk-based distinctions. [<u>Open source</u>](https://www.fda.gov/medical-devices/510k-clearances/medical-device-safety-and-510k-clearance-process)

\[15\] Central Drugs Standard Control Organization, Government of India. “Medical Devices Rules, 2017: Frequently Asked Questions.” Risk-based classification classes A-D and associated regulatory framework. [<u>Open source</u>](https://www.cdsco.gov.in/opencms/export/sites/CDSCO_WEB/Pdf-documents/MDFAQ1324.pdf)

\[16\] European Parliament and Council. Regulation (EU) 2017/745 on medical devices (EU MDR), EUR-Lex. Includes quality management, clinical evaluation and post-market surveillance obligations. [<u>Open source</u>](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32017R0745)

\[17\] European Commission. “EUDAMED - Overview.” Update notes that four EUDAMED modules became mandatory to use from 28 May 2026. [<u>Open source</u>](https://health.ec.europa.eu/medical-devices-eudamed/overview_en)

## Interpretation and verification notes

- Official datasets use different dates, scope definitions and update cycles. The report labels year and scope next to material numbers rather than blending them into a synthetic “current market” figure.

- The India bioeconomy figure is a broad cross-sector estimate. It should not be treated as healthcare revenue or a startup’s total addressable market. \[7\]

- Medical-device import and export figures are national trade values by the source’s device definition. The trade gap is a directional signal, not a direct estimate of addressable domestic production. \[8\]

- The digital-health funding data refer to CB Insights’ category and methodology and may differ from other private-market datasets. \[5\]

- The opportunity matrix, scorecard weights and financing envelopes are strategic portfolio hypotheses. They should be recalibrated against each company’s indication, device class, evidence needs, customer economics and capital plan.

Back to Research