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

Episode 86 · Deep tech · 38 min

Deep tech has no playbook, only silos

Speciale Invest's argument is that the one thing which made Indian venture repeatable, a playbook you can carry from company to company, is exactly what deep tech denies you. Quantum, space and green hydrogen are separate markets on separate clocks, and the numbers move fast enough to prove it: the Indian space opportunity was sized near $8 billion four years ago and $100 billion in an analyst report published a fortnight before this taping.

Deep tech & hardware Venture capital India macro

Vishesh Rajaram

Managing Partner, Speciale Invest · with Vishal Krishna

38:04

Said in this episode

▶ 14:00

$100bn

Indian space market, latest sizing

An analyst report a fortnight before taping put India's space opportunity at $100 billion; four years earlier the two-year-forward number was about $8 billion, and Speciale invested in space back in 2018.

▶ 11:55

50–70%

Gross margin that justifies the wait

Revenues may not arrive for two years, but when they do the margin has to be in the 50–70% band rather than the 10–15% of an incremental business. Otherwise the fight isn't worthwhile.

▶ 8:51

30–40 hrs

Time spent on the founders alone

Before the technology gets scrutinised, Speciale spends thirty to forty hours on founder background. In a long-gestation asset class, tenacity is what is actually being underwritten.

▶ 25:37

8–10 yrs

Speciale's fund life

Not the fifteen-year vehicles some US deep-tech funds run, which is why the GP has to translate phased technical validation into progress an LP can read.

▶ 15:48

$2–3bn

Deep-tech deployment India can't absorb yet

As stated on air, and said quickly: not in 2023-24, but five years out absolutely, provided the ecosystem does a host of groundwork in between.

The brief

The argument in sixty seconds

Rajaram's claim is that deep tech fails the test venture capital normally applies: replicability. There is no playbook that carries from an EV company to a battery company to a quantum-security company, only pages with a huge standard deviation between them, so founders and investors alike have to underwrite each silo on its own physics and its own clock. His filter is exponential rather than incremental: an efficiency gain in the tens of percent is an engineering project, a few-hundred-percent change in the problem statement is deep tech, and the gross margin should eventually land at 50–70% rather than 10–15% to make the wait worthwhile. Speciale is a seed fund comfortable entering at lab scale and underwriting the scale-up risk, which is why 30 to 40 hours go into the founders' background before the technology does. Around that sit his other arguments: capital chased fintech and e-commerce first because those markets were obvious and deep tech's are not. Indian space was pegged near $8 billion four years ago and $100 billion two weeks before taping. India's technical universities spent two decades moving from pure research to corporate IP transfer and are only now producing professor-CEOs. The country's real edge is inherited: liberalisation-era chemistry and pharma, ISRO's guard rails, an OEM manufacturing base. His three forward bets are material science, synthetic biology and manufacturing, deliberately excluding space, robotics, quantum and semiconductors because those are already happening.

Worth your time if you are

Deep-tech founders writing their first risk timeline

Seed investors underwriting lab-to-factory scale-up

Global CVCs scoping an India office

Professors weighing a company against a lab

LPs who expect a SaaS-shaped return curve

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 & hardwareVenture capitalIndia macroRegulation & policyFounder journeyFundraising

01Deep tech as convergence, not a label 0:00 Host Siddharth Mehta opens TDK Ventures' India deep tech series and Rajaram defines the abused term as multiple streams of science (AI, vision, material science, a new form of manufacturing) converging into a moat whose timing had not previously come. 02Exponential impact, or it is engineering 1:38 The line between a lab problem and an engineering problem is drawn by impact, an efficiency gain in the tens of percent versus a few-hundred-percent change in the problem statement. As a seed fund Speciale is unfazed by lab scale because scale-up is the risk it has learned to underwrite. 03Two decades from papers to companies 3:10 IIT Bombay, Delhi and Madras and IISc Bangalore spent twenty years moving from pure academic research into corporate IP sharing and transition, and only lately into building organisations, while a second generation of founders, an Indian echo of the PayPal and Flipkart mafias, leaves scaled startups to build again. 04Can a professor run the company? 6:47 The intuitive bias says academics make good CTOs and doubtful CEOs; Rajaram's experience says they have already built centres, raised large grants and run organisations, and an IIT Bombay EV motor company from eight or nine years ago inspired a wave of professors to look beyond academia. 05Four questions before a cheque 8:51 Thirty to forty hours go into the founders alone, then the frame: what do you know that others don't, why is the timing better now than ever, is there white space around the IP, and will this translate into 50–70% gross margins rather than 10–15%. 06Why capital chased the obvious first 12:40 Capital finds large, legible markets first, and internet and mobile penetration made fintech and e-commerce obvious. Deep tech's markets refuse to be sized in advance, as when the Indian space opportunity went from about $8 billion to $100 billion in four years. 07Every silo is its own market 15:02 Quantum, space and green hydrogen cannot be broad-brushed as one sector and there is very little replicability of a playbook between EV, battery, robotics and quantum-security companies, which is also why India cannot yet absorb billions of deep-tech deployment, though five years out it can. 08Where India already has guard rails 16:20 Liberalisation-era strength in petrochemicals, chemistry, pharma and biotech sets up the next wave in green chemistry, battery materials and synthetic biology, while the biotechnology department, ISRO's opening of private space and a manufacturing OEM base supply the policy, talent and precedent that let sectors flourish first. 09A policy with the words deep tech 19:55 India's deep-tech startup policy landed roughly a day before this recording, and the self-described trained optimist argues that the government using the phrase at all, plus reported plans for a fund, is the signal worth reading, whatever the awkward clauses. 10You have to see it to believe it 21:30 To global funds who barely have India on the radar: the country has always led on adoption, SemiCon India drew international participation, geopolitics makes India a trusted partner for sensitive technology, and belief will follow a couple of SpaceX-like or two-wheeler-Tesla-like companies actually appearing. 11Break the company into risks 23:35 With no shared playbook, founders must map risk ground up and show which phase each cheque retires, investors need a nuanced rather than merely long-term lens, and Speciale, an eight-to-ten-year fund rather than a fifteen-year one, has to translate phased technical validation into LP-legible progress. 12Labs, test rigs and a GaN line 26:40 He refuses the MIT and Stanford comparison as an unequal race between a three-decade and a six-decade ecosystem, and on an absolute basis says yes: the battery lab at IIT Bombay, the combustion test facility at IIT Madras, the gallium-nitride semiconductor line coming up at IISc Bangalore. 13Three bets, three years out 28:10 Material science, synthetic biology and manufacturing are the forward bets. Rare-earth geopolitics makes materials that avoid scarce metals valuable, and space, robotics, quantum and semiconductors are deliberately excluded because they are already playing out. 14Built from India for the world 30:10 Going global is a macro question of opportunity and founder, not micro tactics: a robotics integrator, an earth-observation company, a decarbonisation business that became India-Europe, a battery recycler following faster adoption abroad. Founders now start global rather than transitioning to it. 15Evangelist first, fundraiser second 32:40 His advice: be tenacious and patient, wear the evangelist's hat to over-simplify what you are solving for customers and investors, and reframe the business as technology, scale-up, market-adoption, customer and go-to-market risks plotted on a timeline against the capital each phase needs. 16Unlearn what you know about India 35:00 Global capital can be reached through platforms like Hello Tomorrow, Entrepreneur First and the HAX and IndieBio crowd, and the message to incoming CVCs is to unlearn twenty-year-old assumptions and take local investors and institutions as guard rails.

Timeline

The arc, briefly

Three Indian venture cycles, and the one that refuses to come with a playbook.

2007-08E-commerce takes off, and runs twelve years. Internet and mobile penetration about fifteen years ago made consumer and fintech markets both large and legible, which is why capital went there first. It could deploy a lot of money against an outcome it was able to underwrite in advance.

2011Freshworks and its peers get started. The companies exist a good while before the belief does, and Rajaram dates the belief two or three years later.

2013-14People sit up. This, he says, is when founders and investors accepted that products could be built out of India, and SaaS got a ten-year run that has lasted until now.

2014-15An EV motor company out of IIT Bombay does well. Eight or nine years before this conversation, and it pulled a cohort of professors into thinking about work beyond academia. He rejects the bias that an academic makes a fine CTO and a doubtful CEO: they have already run centres and raised large grants.

2017-18Deep science and tech starts its clock. His optimistic view is that the ten years following it are about building deep tech out of India for the world.

2018Speciale invests in space. The forward market was pegged at about $8 billion two years out and nobody had a bigger number for it. An analyst report published a fortnight before this taping puts Indian space at $100 billion. That gap is the whole difficulty with deep tech: the markets refuse to be sized in advance.

2023-24The policy lands; the money still cannot. India's deep-tech startup policy was published roughly 24 hours before the recording, and he reads the government using the phrase at all, plus reports of a fund behind it, as the thing worth taking from it. He also says plainly that the country cannot absorb two or three billion dollars of deep-tech deployment this year.

2028Five years from now, he says, absolutely yes. On condition that a host of groundwork happens in between. His three forward bets are material science, synthetic biology and manufacturing, with space, robotics, quantum and semiconductors deliberately left out because they are already playing out.

Takeaways

Ideas to carry out of this hour

There is no deep-tech playbook, only silos

The market for quantum is not the market for space is not the market for green hydrogen, so the reflex that a rising country lifts a whole sector simply does not apply. You have to pick the right vertical. Worse for investors, there is very little replicability between an EV company, a battery company, a robotics company and a quantum-security company; anyone assembling a deep-tech playbook ends up with pages carrying a huge standard deviation from each other. That is also why he says India cannot yet build for scale in this asset class, though in five years it can.

▶ 15:02 ▶ 15:33 ▶ 15:48

The filter is exponential impact, not a better number

Speciale's first question is not whether the work happens in a lab or a workshop but what the technology means: an efficiency improvement in the tens of percent is linear, a few-hundred-percent change in the problem statement is exponential. Only the second kind gets excited attention, because only that kind resets the market it lands in. As a seed fund it is explicitly comfortable entering at lab scale, treating lab-to-factory as the risk it has learned to underwrite rather than a reason to wait.

▶ 1:53 ▶ 2:07 ▶ 2:24

Diligence starts with the founder, not the technology

Thirty to forty hours go into founder background before the science, because deep tech carries the baggage of long gestation, hard fundraising and no rush of buyers. You simply have to live long enough to play. The frame that follows has four parts: what do you know that most people don't know or don't agree with, why hasn't this been done before and why is the timing better now, is there white space around the innovation (checked with IP attorneys, since some founders deliberately keep trade secrets rather than file), and does the exponential claim eventually translate into 50–70% gross margins rather than 10–15%.

▶ 8:51 ▶ 10:21 ▶ 11:09

Professors make better CEOs than the bias allows

The intuitive assumption is that an academic founder is a natural CTO and a doubtful CEO. Rajaram's experience is the opposite: the professors he has backed already coordinated large research centres, raised substantial grants and dealt with people at scale, which is not identical to raising venture capital but is organisation-building all the same. He cites an IIT Madras aerospace department head who now runs his company full-time and is no longer an active professor, and notes that an EV motor company out of IIT Bombay eight or nine years ago inspired a cohort of academics to look past the lab.

▶ 7:20 ▶ 7:51 ▶ 8:21

Capital chased fintech because fintech was legible

Capital moves first towards markets large enough to absorb it and obvious enough to underwrite, and India's internet and mobile penetration made consumer and fintech exactly that fifteen years ago. Deep tech offers no such legibility: Speciale invested in space in 2018, when the forward market was pegged around $8 billion, and an analyst report a fortnight before this taping put it at $100 billion. Cycles are his consolation: e-commerce took off around 2007-08 and ran twelve years, SaaS became credible around 2013-14 after a 2011 start, and deep science and tech began its clock in 2017-18.

▶ 13:12 ▶ 14:00 ▶ 14:45

India's deep-tech edge is inherited, not invented

Roll back to the liberalisation era and India built real depth in petrochemicals, hydrocarbons, chemistry, pharmaceuticals and biotechnology, the direct ancestors of green chemistry, clean tech, battery materials and synthetic biology. Add government departments that are genuinely strong in science, an ISRO that has opened private space and now lends startups testing facilities, talent and counsel, and an OEM manufacturing base that already exports vehicles. Wherever guard rails exist as policy, talent or precedent, he argues, that is the low-hanging fruit: Ather and Ultraviolette in EVs, green hydrogen firms riding the new policy.

▶ 17:39 ▶ 18:08 ▶ 18:54

Sell your risks on a timeline, not your ambition

The world already labels deep tech high-risk, long-gestation and capital-intensive, so his counter is to answer jargon with better jargon: decompose the company into technology risk, scale-up risk, market-adoption risk, customer risk and go-to-market risk, then plot each against the capital and the calendar. Some companies need $10 million to start and some half a million, and knowing which industry DNA you are in is what makes the conversation a fit test rather than a plea for money. The same discipline runs upward. As a GP he has to show LPs which of five or seven validation phases a portfolio company has actually cleared, because revenue is only the last of them.

▶ 33:30 ▶ 34:01 ▶ 26:08

Founders no longer transition from Indian to global

He rejects micro tactics for going global: it is a macro question of the opportunity and the nature of the founder. Where the pain is sharper and the paying propensity higher elsewhere, companies move: a robotics integrator became a global warehousing systems player, an earth-observation startup runs India-US, a decarbonisation company became India-Europe, and battery recycling may find Europe adopting faster than home. Going global beyond that, for vanity, is hard when the market does not need you.

▶ 30:40 ▶ 31:11 ▶ 32:10

The numbers, drawn

What the episode measures

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

How India's space market got re-rated

$ billion

Both figures as stated on air; the report's publisher is not clearly audible in the recording. Speciale first invested in space in 2018, when the larger number did not exist.▶ 14:00

The margin that makes a deep-tech wait worthwhile

% gross margin

Upper bounds of the ranges said on air ('10, 15%' versus '50, 70%'), applied to revenues he concedes may be two years away.▶ 11:55

Four questions before a cheque

the diligence frame, in the order he runs it

Question three is checked with IP attorneys, and not to prove a patent exists: some founders keep trade secrets rather than file, so the point is whether the company is trampling on someone else's claim.▶ 8:51

Sell the risks, not the ambition

the five risks he wants plotted against money and calendar

His answer to being called high-risk and capital-intensive: break the company up and price each piece. Some businesses need $10 million to start, some half a million. Upward, it is the same discipline, since revenue is only the last of five or seven validation phases an LP is shown.▶ 33:30

Worth keeping

Lines that stay

What do you know that most people don't know, or don't agree with you on — that allows you to go build something others won't do, or can't do, or don't know how to do?

▶ 10:21

I'm a big believer of half glass full. If someone asks me whether the deep tech ecosystem in India is perfect, the short answer is no — but I'm satisfied.

▶ 17:04

If you go back and Google the word 'deep tech' in the Indian government ten years ago, I don't know if you'll find a lot of it.

▶ 20:13

I don't think founders transition from being Indian to global. Today founders start their journey saying I'm building from India for the world.

▶ 32:10

The most important thing is for them to unlearn what they already know about India — because the assumptions about India are not very different from what they were 20 years ago.

▶ 36:48

Clips that travel

Short on time? Start here

Deep-tech founders preparing a seed round

Four questions before a deep-tech cheque

The whole diligence frame in three minutes: what you know that others don't, why now, IP white space, and the margin the business must eventually earn.

9:50 → 12:40 · 3 min ▶ Watch clip

Investors sizing an unfamiliar sector

Why capital chased fintech first

Legible markets get funded first, and the space market going from $8 billion to $100 billion in four years is the cleanest evidence that deep tech resists sizing.

12:43 → 16:20 · 4 min ▶ Watch clip

Founders choosing which deep-tech vertical to enter

Where India already has guard rails

The inherited-advantage map: liberalisation-era chemistry and pharma, ISRO's opened doors, an OEM base, and the policy tailwinds that follow.

17:04 → 19:55 · 3 min ▶ Watch clip

Founders and GPs managing long-gestation expectations

Break the company into risks, not a pitch

How to retire risk in phases when no playbook exists, and how an eight-to-ten-year fund reports pre-revenue progress to its LPs.

23:35 → 26:40 · 3 min ▶ Watch clip

Anyone reading the next Indian deep-tech cycle

Three bets, three years out

Material science, synthetic biology and manufacturing, plus the reasoning for leaving space, robotics, quantum and semiconductors off the list.

28:10 → 30:10 · 2 min ▶ Watch clip

Glossary

The jargon, unpacked

Deep tech

In Speciale's usage, hard-to-build technology where multiple streams of science or engineering (AI, vision, material science, new manufacturing) converge to create a moat whose timing had not previously arrived.

Scale-up risk

The risk that a technology working at lab scale fails to survive the move to factory scale. This is the specific risk a seed fund entering at lab stage says it is underwriting.

Founder-market fit

The proprietary insight or domain depth that lets a specific founder build something others won't or can't. It is the thing Rajaram says his diligence ultimately converges on.

IP white space

The room around an innovation where no one else holds patents; checked with IP attorneys not to prove the patent exists but to confirm the company isn't trampling on someone else's claim.

GP and LP

The general partner who runs a fund and the limited partners who supply its capital. Here, the relationship that has to be re-taught when returns arrive on a physics timeline rather than a software one.

Synthetic biology

Engineering biological systems to manufacture products more efficiently or sustainably. One of his three forward bets, aimed at industry, food and bio-pharma.

Rare earth constraint

The geopolitical restriction on how scarce elements and materials move between countries. That is the reason he expects science-led materials that avoid expensive or unavailable metals to matter.

Connections

If this resonated, go here next

EP 164 · Capital

Full transcript

The whole conversation, searchable

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