The numbers Brief Episode map Timeline Takeaways Charts Quotes Clips Glossary Go next Transcript

Episode 160 · Deep tech · 47 min

The valley between the lab and the line

India has never lacked invention. It lacks the vendor, the yield and the rulebook that turn an invention into a production line. Three deep-tech operators make the case in numbers: an Indian solar cell burns 120 mg of silver where a Chinese one burns 90, and the country's first ocean farm had to be designed in India and dropped in Indonesian water because no Indian policy lets a company lease sea surface.

Deep tech & hardware Manufacturing Climate & energy

Sri Sailaja Nori, Avishek Kumar and Kedar Medhi

Founders and a corporate venture investor, Sea6 Energy, VFlowTech and Philips Ventures · with Vishal Krishna

47:03

Said in this episode

▶ 21:45

85%

Of India's crude oil is imported

Sea6's founding arithmetic: replacing that with biomass would take close to a billion tonnes, roughly India's entire agricultural productivity.

▶ 36:39

120 vs 90 mg

Silver per solar cell, India vs China

Stated on air as the consumption gap that decides cost competitiveness; the caption garbles the Indian figure, so read it as approximately 120 mg.

▶ 36:30

~60%

Yield on India's announced lithium gigafactories

Cited for the 10–20 GW battery plans announced by large Indian groups. Enough to build the plant, not enough to be price-competitive.

▶ 24:43

100 ha

Sea6's demonstration farm off Lombok

Commissioned February 2024 with Indonesian government support; all technology designed and manufactured in India and shipped out only for deployment.

▶ 25:46

11,000 km

India's coastline, with no sea-lease policy

There is no route for a private company to access sea surface in India, which is why the first ocean farm went to Indonesia.

▶ 37:05

100 MWh

NTPC's first storage tender

Described as roughly a $50 million project and, more importantly, the first demand signal large enough to pull component suppliers into India.

The brief

The argument in sixty seconds

The panel's shared claim is that India's deep-tech bottleneck was never invention. It is the short, brutal distance between a lab bench and a line. Avishek Kumar, who abandoned a solar PhD to help build the first giga factory for REC and now runs the vanadium flow-battery company VFlowTech, argues the binding constraint is vendors and yield rather than capital: an Indian solar cell consumes 120 mg of silver where a Chinese cell consumes 90, announced lithium gigafactories are landing near 60% yield, and if no supplier ecosystem exists there is nothing to leverage, so your first order has to be big enough to create a vendor out of nothing. Sri Sailaja Nori makes the policy version of the same argument. Sea6 Energy spent sixteen years inventing ocean agriculture from zero, from custom boats and anchoring without divers to seeding and harvesting machines and drone monitoring, all designed and built in India, then commissioned its first 100-hectare farm off Lombok, because Indonesia will lease sea surface to a company and India, with 11,000 km of coastline, has no policy that does. Kedar supplies the regulatory clock: in medical devices, intended use fixes the risk class, the risk class fixes the evidence burden, and a CDSCO manufacturing licence runs from a couple of quarters to more than a year, so the quality system has to be built before the approval is needed. The stakes are 85% imported crude, a billion tonnes of biomass needed to displace it, and a first NTPC storage tender of exactly 100 MWh against gigawatt-scale announcements.

Worth your time if you are

Hardware founders about to sign a first capex cheque

Deep-science teams stuck between a lab result and a line

Policy people who treat sandbox and sea-lease rules as back-office work

Corporate venture teams sourcing from Indian deep tech

Students weighing a PhD against a factory floor

Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

Deep tech & hardwareManufacturingClimate & energyRegulation & policySupply chain & agriHealthcare

01Invention was never the problem 0:00 The moderator frames the session around the valley of death between lab and line: fragmented manufacturing pockets in Bangalore, Pune and Hosur, a capital round that took one panelist 25 investors, thin material-science and precision-manufacturing talent and missing supply chains, set against PLI, the semiconductor mission, contract manufacturing, China plus one and returning talent. 02300 IPs, three startups, one plane 5:03 Introductions: a corporate venture investor with 28 years and over 300 IPs reviewed, a serial founder of three deep-tech companies across IoT, solar and storage who calls himself a resident of a plane, and Sea6 Energy's co-founder and chief scientist, in seaweed since 2010. 03Intended use sets the whole clock 7:06 Kedar walks through Indian medical-device regulation: intended use determines risk class, risk class determines the clinical-evidence burden, CDSCO splits investigational and manufacturing-licence tracks that can run from two quarters to over a year, and it is still less predictable than the FDA's 510(k). 04There are no rules for your new thing 12:10 Avishek reframes deep tech as solving a critical science problem that immediately creates a new one, such as a safer, longer-life battery with no regulatory category to sit in, and argues for regulatory sandboxes, citing a shrimp-vaccine biotech whose founder cannot validate a product he is not allowed to sell. 05From $4 a watt to a giga factory 14:14 Solar at $4 per watt peak sent him into a thin-film silicon PhD; China's wafer build-out taught him that abundance plus demand collapses cost, so he jumped to manufacturing and set up the first giga factory for REC, sand to silicon to wafer to module, while India made 4 GW of low-efficiency panels for water pumps. 06Solar is unsustainable without batteries 16:40 A 2017 attempt at AI in manufacturing proved too early and became a 20 GW services business; then the realisation that cheap, uncontrollable solar needs grid-scale storage, a Singapore vantage point on how China sequences demand and supply chain, and a co-founder met in 2018. 07A stable company and a new itch 18:50 With the solar services company profitable and run by a CEO, he went after a technology India needs built on a 100% local supply chain, and a co-founder who lifted non-lithium AC-side efficiency roughly 10 points above the usual 60–65% through IP. 08A billion tonnes of biomass, from the sea 20:58 Sea6 Energy started on the IIT Madras campus in 2010 chasing a bio-based feedstock big enough to replace crude: 85% of India's oil is imported, and matching it means roughly a billion tonnes of biomass, the country's entire farm output, so the team went to water that needs no land, fresh water or fertiliser. 09Why the first farm is off Lombok 23:40 Seaweed farming exists as knee-deep cottage industry in Indonesia, the Philippines and Malaysia; Sea6 automated it and commissioned a 100-hectare farm off Lombok in February 2024. Everything was designed and built in India, deployed abroad because Indonesia leases sea surface to companies and India's 11,000 km coastline has no such policy. 10Build the boats, then the drones 26:00 Being at the intersection of biology and engineering made funding the first hurdle, and every piece had to be invented in-house, from seeding and harvesting machines to custom boats, anchoring without sending anyone underwater and drone monitoring with data analysis, en route to a fully operational square kilometre and an HPCL MOU on aviation fuel. 11What a corporate VC actually checks 29:12 Kedar explains corporate venture logic, de-risking a new space or expanding a business line faster than organic effort, then the diligence list: quality systems, clinical collaborations, key-opinion-leader depth, journal publications and a clear joint proposition, illustrated by a Korean cath-lab AI company and the alternative marketplace model. 12Yield, silver and the vendor that isn't there 34:30 Batteries and solar are commodities that must clear an ROI benchmark, so new technology faces no demand, heavy capex, twenty finicky components and yields near 60%; India's cells burn 120 mg of silver to China's 90, vendors only appear if orders are large enough, and NTPC's first 100 MWh tender is the demand signal. 13Next: thirty to fifty square kilometres 38:20 Sea6's first square kilometre completes next financial year, after which downstream product demand dictates a jump to 30–50 square kilometres and, potentially, $50–100 million of Indian revenue over five years. 14Do universities actually collaborate? 39:22 An audience member who studied in Chennai, the UK and France asks why Indian university-industry collaboration lags; answers cover an IIT Kanpur incubation that yielded 60-plus IPs from free lab access, IIT Delhi work on vanadium from petroleum waste, a tier-two energy university that unlocked a state deployment, and healthcare's innovator-institute-hospital triangle. 15Rare earths, vanadium and anchor customers 43:20 On China's export bans: vanadium is the 15th most abundant element and recoverable from low-quality crude at Indian refineries, which is why local production was the thesis, and the closing advice to a fellow storage founder is to find an anchor user among NTPC, HPCL, BHEL or the IPPs, and to lobby for a broader PLI.

Timeline

The arc, briefly

Two hardware companies, and the distance between a working lab and a working line.

2010Sea6 starts on the IIT Madras campus. Straight out of graduation, chasing a bio-based feedstock big enough to replace crude. India imports about 85 per cent of its oil and matching that would take close to a billion tonnes of biomass, roughly the country's entire farm output, so the team went to water, which needs no land, fresh water or fertiliser.

2017Too early for AI on the factory floor. Avishek set up SunConnect to put an AI process into Indian solar manufacturing, found the market was not ready for it, and built a services business instead. It has since supported 20 GW of solar and wind, about 30 per cent of India's renewable capacity. Solar was already cheaper than coal that year, and, he decided, unsustainable without storage.

2018A co-founder, in the wrong country. He met Dr Arjuna in Singapore, which he calls the wrong choice because there is no market there. The technology is a vanadium redox flow battery, picked because vanadium is the fifteenth most abundant element on earth and can be recovered from the low-quality crude Indian refineries already run.

Feb 2024The first ocean farm goes in off Lombok. 100 hectares, commissioned with Indonesian government support. The boats, the seeding and harvesting machines, the anchoring that sends nobody underwater and the drone monitoring were all designed and built in India and shipped out only for deployment, because Indonesia will lease sea surface to a company and India, with more than 11,000 km of coastline, has no policy that does.

TodayNTPC puts out its first storage tender. 100 MWh, roughly a $50 million project, and the first demand signal large enough to pull component suppliers into the country. It has to be, because an Indian solar cell burns about 120 mg of silver against China's 90, and the announced Indian lithium gigafactories are landing near 60 per cent yield. Raising roughly $40 million took one panellist a very long time and 25 investors.

Next FYThe first square kilometre is finished. After that the jump is to 30 or 50 square kilometres, sized by what the downstream verticals need, with an HPCL memorandum pointing the feedstock at aviation fuel in three to four years.

+5 years$50 to 100 million of Indian revenue. Sailaja's number for what the ocean farms could carry if the downstream products hold up.

Takeaways

Ideas to carry out of this hour

The missing piece is a vendor, not an idea

A battery or a solar module is twenty-odd components, each needing a specific shape and grade, and India simply does not have the suppliers who make them. The consequence is circular: you cannot buy quality inputs until a vendor exists, and no vendor develops until someone places an order large enough to justify the tooling. Avishek's reading of how American semiconductors and Chinese manufacturing both scaled is that firms leveraged each other, so the Indian founder's real skill is mastering the art of leveraging government relationships, demand and whatever supply chain exists, because with no vendors there is nothing to leverage at all.

▶ 35:45 ▶ 36:45 ▶ 37:32

Yield, not capex, decides whether a line survives

The capital to build a plant is the easy half; the hard half is what comes off it. Announced Indian lithium-ion gigafactories in the 10–20 GW class are landing around 60% yield, which is not competitive at commodity prices, and even in solar, where India sources roughly 90% of raw materials from China, an Indian cell consumes about 120 mg of silver against China's 90. On a commodity with a fixed ROI benchmark, that gap alone decides the race.

▶ 36:15 ▶ 36:30 ▶ 36:39

When no regulation exists, ask for a sandbox

Deep tech by definition creates a product category regulators have not written rules for, such as a non-lithium battery with no fire risk or a vaccine for shrimp, and the absence of a rule is more paralysing than a strict one, because the delay is open-ended and outside the founder's control. Avishek's ask is a formal sandboxing process across government agencies: let the company deploy and sell under supervision, because a product you cannot sell is a product you cannot validate or scale. The data that comes out is what lets regulators write the next generation of rules.

▶ 12:59 ▶ 13:15 ▶ 13:31

Intended use sets the entire regulatory clock

In Indian medical devices everything cascades from one definition. Intended use fixes the risk classification, from surgical sutures at the low end to pacemakers at the high, and the classification fixes how much clinical and effectiveness evidence must be generated. CDSCO then runs two separate tracks, investigational studies and manufacturing licences, that take anywhere from a couple of quarters to over a year, still less predictable than the FDA's 510(k), so the runway has to be planned around it and an ISO 13485 quality system built before it is asked for.

▶ 7:40 ▶ 8:28 ▶ 11:45

The ocean is the only place left to grow a billion tonnes

India imports almost 85% of its crude, and the panel's arithmetic is that replacing the fuels, chemicals and materials it becomes would need close to a billion tonnes of biomass, roughly the entire agricultural output of the country. There is no more land and no more fresh water, but 70% of the planet's surface is ocean and the plants there need none of it. Sea6's bet is that aviation fuel and chemicals, the uses electrification cannot reach, get their feedstock from water.

▶ 21:43 ▶ 22:14 ▶ 22:28

India exported its first sea farm because of a missing policy

Sea6 designed, prototyped and manufactured the whole system in India, then shipped it to Indonesia purely for deployment, commissioning 100 hectares off Lombok in February 2024. The reason is regulatory, not technical: Indonesia treats the ocean as a resource and will lease 100 or 1,000 hectares of sea surface to a company for as long as that company exists, while India has no mechanism for private access to sea surface at all. A country with more than 11,000 km of coastline still thinks of itself as a land nation.

▶ 24:43 ▶ 25:12 ▶ 25:46

A corporate VC buys a joint proposition, not a demo

Corporate venture starts from the parent's inorganic growth strategy: de-risk a new category before entering it, or expand an existing line faster than internal effort can. What it then examines in a startup is whether the house is in order: the quality management system, who is orchestrating the clinical collaborations, the depth of key-opinion-leader engagement, publications in reputable journals, and above all whether the founders can articulate what your solution plus the corporate's solution looks like in market.

▶ 30:30 ▶ 31:51 ▶ 32:19

Deep-tech hardware is a game of patience, and the board has to agree

There is no short-term solution in deep-tech hardware and manufacturing: the first 100 MWh will not be profitable, and pretending otherwise breaks the plan. The moderator's own data point is that one panelist's roughly $40 million took a very long time and 25 investors to assemble. The obligation is to hold the roadmap, and to align the board to it in advance rather than mid-scale-up.

▶ 2:19 ▶ 38:00 ▶ 38:14

The numbers, drawn

What the episode measures

Every figure below was said on air, with timestamps included and caveats kept.

What an Indian solar cell pays extra for

mg of silver per cell

As stated in conversation: Indian cells consume around 120 mg of silver against China's 90. The auto-caption stumbles on the Indian figure, so treat it as approximate. Silver is the single costliest input after the wafer.▶ 36:39

The import bill Sea6 was founded against

% of India's crude oil

Imported85%

Domestic15%

Only the 85% import share was stated on air; the domestic slice is the arithmetic remainder. Displacing it with biomass would need roughly a billion tonnes, about India's entire farm output.▶ 21:45

The vendor loop

why a first order has to be bigger than the demand

Avishek's account of scaling a battery or a module made of twenty-odd components, each needing a specific shape and grade. The NTPC tender is described on air as roughly a $50 million project, and matters less as revenue than as the first demand signal big enough to pull suppliers into India.▶ 37:32

Built in India, dropped off Lombok

everything Sea6 had to invent, and where it ended up

Sailaja's account of sixteen years spent inventing ocean agriculture from nothing. Indonesia will lease 100 or 1,000 hectares of sea surface to a company for as long as that company exists; India has no mechanism for private access to sea surface at all.▶ 25:46

Worth keeping

Lines that stay

If there is abundance of something in nature and there is demand, the cost of anything can come down.

— Avishek Kumar ▶ 15:31

This is like the stone age of land agriculture. So we said, what if we use our engineering capabilities to intervene and bring in automation?

— Sri Sailaja Nori ▶ 24:12

We have more than 11,000 km of coastline as a country, and we never think of ourselves as a water nation.

— Sri Sailaja Nori ▶ 25:46

When you scale up, you have to leverage. If there are no vendors, no supplier — what will you leverage?

— Avishek Kumar ▶ 37:32

There is no short-term solution if you are in deep tech hardware and manufacturing. You need to play a game of patience.

— Avishek Kumar ▶ 38:00

Clips that travel

Short on time? Start here

Founders whose product has no regulatory category

The sandbox a new chemistry needs

The clearest statement of the deep-tech regulatory trap, that you cannot validate what you are not allowed to sell, and the sandbox fix.

12:26 → 14:14 · 2 min ▶ Watch clip

Researchers wondering whether to leave the lab

From $4 a watt to a giga factory

A PhD abandoned for a sand-to-module giga factory, and the lesson that abundance plus demand collapses any cost curve.

14:57 → 17:26 · 2 min ▶ Watch clip

Policy people who think regulation is back-office work

Why the first sea farm is Indonesian

Stone-age cottage farming automated into a 100-hectare farm, built in India and deployed off Lombok because India has no sea-lease policy.

24:09 → 26:00 · 2 min ▶ Watch clip

Health-tech founders courting a strategic investor

What a corporate VC actually checks

The diligence list behind a corporate cheque, plus a real example: Korean cath-lab AI slotted into an existing Philips channel.

31:31 → 34:30 · 3 min ▶ Watch clip

Hardware founders costing a first production line

Yield, silver and the missing vendor

The densest run of numbers in the session: 60% yields, 120 mg of silver against 90, and why the first big order has to create its own supplier.

34:56 → 37:42 · 3 min ▶ Watch clip

Glossary

The jargon, unpacked

Valley of death

The stretch between a result that works in the lab and a process that works on a production line, where the panel says most Indian deep-tech inventions stall.

CDSCO

India's Central Drugs Standard Control Organisation, which runs separate approval tracks for clinical investigation and for medical-device manufacturing licences, taking anywhere from two quarters to over a year.

510(k)

The US FDA clearance route the panel uses as the benchmark for predictability, a defined path India's process is still working toward.

ISO 13485 / QMS

The quality management system standard for medical devices; having it institutionalised early shortens regulatory approval and makes a startup credible to corporate partners.

Regulatory sandbox

A supervised space to deploy and sell a technology no rulebook covers yet, so that real data exists for regulators to write the next set of rules from.

Vanadium redox flow battery

VFlowTech's chemistry: a non-lithium, non-flammable, long-life storage technology built on vanadium, described as the 15th most abundant element and recoverable from low-quality crude at Indian refineries.

Yield

The share of units coming off a line that meet specification; at around 60%, the panel argues, an Indian battery or solar plant cannot compete on price no matter how good the capex.

Anchor customer

The first large buyer whose order justifies a factory and pulls suppliers into existence, here PSUs such as NTPC, HPCL and BHEL running 100–200 MWh storage pilots.

Connections

If this resonated, go here next

EP 129 · Deep tech

Full transcript

The whole conversation, searchable

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.