BriefChartsQuotesChaptersGlossaryGo nextTranscript
Episode 177 · Deep tech · Press Q&A · 49 min

The next cheque is a purchase order

IIT Madras and Unicorn India Ventures announced a ₹450 crore first close for their deep tech fund on 26 September. At the press Q&A that followed, Prof. V. Kamakoti said money no longer holds Indian deep tech back. What it lacks is Indian companies willing to buy.

Deep tech & hardwareEducation & skillingSales, GTM & growth
KAM
Prof. V. Kamakoti
Director, IIT Madras · with Bhaskar Majumdar and Nat Malupillai
The next cheque is a purchase order: episode thumbnail
48:50
The brief

The argument in sixty seconds

Kamakoti says capital is no longer scarce. Deep tech takes 12–13 years to mature, he says, so the fund is patient money that keeps IIT engineers building instead of drifting into finance. IIT Madras has built the supply side: a student-driven stack that turns ideas into companies, more than a patent a day since FY24, and industry commissioning two-thirds of its new research. The government steps in first, with chip-design grants and, Majumdar says, first orders for his defence and chip startups. Indian corporates don't. Kamakoti asks them for a ₹10 lakh first order. Nat Malupillai adds the other half: startups must earn corporate trust by becoming globally ready.

The argument, drawn

Eight charts, forty-nine minutes

Figures are as said on stage unless marked otherwise; click ▶ to hear them. IIT Madras's own counts come from its patent release, research office and NIRF filing. The fund itself, ₹450 crore at first close, is drawn in EP 176.

Patient money keeps engineers engineering

as told on stage by Kamakoti and Majumdar
Years to maturity
Ather Energy: on campus 2013, listed 202512 yearsMajumdar's earlier fund: 7 years✗ forced exitFrontier Fund I: 10 + 2 years05101213Kamakoti: deep tech takes 12–13 years
One year's civil engineers, his count
5 stayed in civil120 moved out of itIIT spends “around 15 lakh rupees per annum”on a biotech engineer, who then goes intohigh-frequency trading
Without patience, startups from IITs “will not survive”, Kamakoti says. Majumdar's earlier seven-year fund, he says, had to exit a robotics company now heading for an IPO. If every biotech graduate took the trading option, Kamakoti asks, “who would have made the vaccine?”▶ 0:52▶ 3:43▶ 4:44▶ 15:05

Technology is necessary, not sufficient

how startups are picked: Kamakoti's funnel, Malupillai's criteria
1,275 applications100 shortlisted (8%)What gets through, per Malupillai1Serves a national need: space,quantum, batteries2A founding team, not a lone founder33+ years watched in incubation4A paying customer, not just tech“Does your great Nobel Prize-winninganswer have a question thatyou're solving for?”
Kamakoti's figures, for last year; IIT Madras's incubator reports over 1,200 applications and 112 incubated for FY26. Malupillai wants founding teams because “being an entrepreneur is a very lonely job”.▶ 9:50▶ 10:48▶ 11:50

India's quantum opening is in applications

how Kamakoti frames the fund's quantum bet
ClassicalcomputerQuantumacceleratorhard part out, answer back:the way AI work goes to a GPUComputingbasic buildingblocksCommunication“somethingvery big”alreadySensingmore precise,less power,smallerIndia's white space: applicationsFund pick: Quanastra, single-photon detection
He says 10 to 20 sectors came forward wanting quantum as an accelerator. At scale, “every small milliwatt matters”.▶ 7:08▶ 8:19▶ 8:42

Five layers from idea to company

IIT Madras's innovation stack, in the order Kamakoti gave it
01 · Idea
E-Cell
The entrepreneurship cell catches an idea, takes the student to the patent office and tells them: you own it.
▶ 21:55
02 · Design
CFI and its clubs
The Centre for Innovation, run by students: the Agnirath solar car, rocketry, the Raftar race car. Getting in is “tougher than JEE”.
▶ 22:18
03 · Prototype
Nirmaan
The pre-incubator, which turns a design into a working prototype.
▶ 23:24
04 · Company
Incubators
The IITM Incubation Cell, plus fintech, cyber-physical, health tech, bio and rural technology incubators.
▶ 23:37
05 · Mentor
GDC
The Gopalakrishnan-Deshpande Centre, funded by alumni Kris Gopalakrishnan and Desh Deshpande, mentors founders “to become unicorns”.
▶ 23:57
Kamakoti calls it “student-driven”, with a faculty mentor on every project and support from alumni and industry. Ather Energy came out of its incubation cell. Getting through is now “much tougher than the actual B.Tech course”, he says.▶ 24:14▶ 24:31

Patents are a habit; industry leads new research

IIT Madras's own figures, beside what was said on stage
Patents filed by IIT Madras, by year
a patent a day (365)214FY22232FY23378FY24419FY25431FY26Institution fees fell about 80% in 2021:expedited exam ₹8,000, was ₹60,000
Industry's share at IIT Madras, FY25
industryother sourcesNew research sanctions (research office)66%Money received (NIRF, consultancy)29%“Your Saraswati is useful only ifyou generate Lakshmi.”
Said on stage: 384 patents and 70% industry money. The bars use IIT Madras's own counts: 378 filed in FY24, and the 70% holds roughly for new sanctions, not for money received.▶ 19:22▶ 20:45▶ 35:03▶ 35:31

Capital came. Corporate buyers haven't.

who backs Indian deep tech today, by the panel's account
Capital“where to put is the question”✓Government as first buyerDLI grants; first orders, per Majumdar✓Indian corporates“the mindset is not there”✗Kamakoti's ask, against today's launch eventthe event, his estimate: ₹50–60 lakhone first order: ₹10 lakh+ let the startup show your logo
Kamakoti wants corporates to buy on nomination: in open bids, foreign rivals can use “predatory pricing to kill us”.▶ 26:29▶ 28:23▶ 30:05

Most problems don't need the biggest tool

the Shakti processor and frugal AI, his figures
He says 90% of systems need ~100 MHzmotor and AC controllers, micsso, he argues, a 180 nm chip has a marketCost to build the chip, his figuresin the West: ~10×Shakti: ₹4 crore“Free” peripheral IP came with $200,000in maintenance, so the team built its ownFrugal AI: an LLM reaches forall 64 art forms; mostproblems need one.“You never shoot a mosquitowith a revolver.”
Critics wrote off his 180 nm chip, he says, and it worked. “Frugal research will lead to optimal solutions”, he argues, and that holds for AI too: “India can't afford” the energy LLM-scale models need.▶ 39:22▶ 41:49▶ 43:21

Trust runs both ways

Kamakoti asks corporates to buy; Malupillai asks startups to earn it
Corporates should
Startups must
The ask
▶ 45:44
Treat startups as collaborators: “Nurture them… be proud, mention their product.”
Be globally ready: a steady, consistent product, built and designed right, with less jugaad.
The price
▶ 44:08
Pay a little more: “even if I lose some 5 rupees, 10 rupees more, go and buy”.
Get certified, and get data privacy good enough to sell in Europe.
The reason
▶ 45:11
Imported tech is “a black box”: no one can promise it won't do what you don't want.
A big customer has its own global customers at stake.
Kamakoti's bridge between the two is funding that carries startups to TRL 8, because corporates “don't want to take any risk. They are right.”▶ 46:34▶ 47:13▶ 47:40
Worth keeping

Lines that stay

It may have bugs, it will not have malice.

Prof. V. Kamakoti, on chips designed in India ▶ 33:59

In a well, more you drain, more water will come.

Prof. V. Kamakoti, closing his call to corporates ▶ 48:31
Chapters

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.

010:00–2:53 · Deep tech needs patience022:53–6:08 · Keeping engineers in core036:08–9:27 · Where India fits in quantum049:27–12:46 · How startups are picked0512:46–15:34 · Majumdar on patience0615:34–18:58 · Habits and German DNA0718:58–21:48 · The road to a patent a day0821:48–25:15 · The student-driven stack0925:15–29:58 · The purchase-order gap1029:58–34:07 · Government buys; buy Indian1134:07–38:34 · Industry money1238:34–43:42 · Shakti and frugal AI1343:42–46:34 · Why buy from startups1446:34–48:50 · A call to corporates10:0020:0030:0040:0048:50
Venture capitalHiring & talentDeep tech & hardwareEducation & skillingSales, GTM & growthRegulation & policy
01Deep tech needs patience 0:00 02Keeping engineers in core 2:53 03Where India fits in quantum 6:08 04How startups are picked 9:27 05Majumdar on patience 12:46 06Habits and German DNA 15:34 07The road to a patent a day 18:58 08The student-driven stack 21:48 09The purchase-order gap 25:15 10Government buys; buy Indian 29:58 11Industry money 34:07 12Shakti and frugal AI 38:34 13Why buy from startups 43:42 14A call to corporates 46:34
Glossary

The jargon, unpacked

TRL
Technology readiness level: a scale from 1 (basic research) to 9 (proven in operation).
DLI scheme
The government's Design Linked Incentive scheme, which funds Indian chip-design companies. Kamakoti cites Mindgrove and InCore.
Shakti
IIT Madras's family of open-source RISC-V processors.
Expedited examination
A fast track through the Indian patent office's examination queue, for a higher fee. Government-funded institutions such as IIT Madras can request it.
Buying on nomination
Awarding a contract directly to a chosen supplier, without an open tender.
Connections

If this resonated, go here next

EP 176 · Deep tech

Deep tech needs different rules

Bhaskar Majumdar (Unicorn India Ventures), recorded the same day
Deep tech & hardwareVenture capital
Deep dive
EP 92 · Deep tech

The lab was never the bottleneck

Krishnan Balasubramanian (IIT Madras)
Education & skillingDeep tech & hardware
Deep dive
EP 164 · Capital

Early money is solved, Series B is not

Four deep tech investors, from pi Ventures to PremjiInvest
Venture capitalDeep tech & hardware
Deep dive
Full transcript

The whole conversation, searchable

155 segments
KAMProf. V. KamakotiNMNat MalupillaiBMBhaskar MajumdarVKVishal Krishna, UpStreamVoiceQthe floorMODmoderator

Two machine transcripts, reconciled and lightly edited. Floor questions were off-mic and are condensed where hard to hear; [unclear] marks words neither transcript caught. Click a timestamp to open that moment on YouTube.

Checked against the record: India administered about 220 crore COVID-19 vaccine doses (said: 275 crore). Agnirath placed third of eight entries, covering 1,193 km (said: third in the world, of 22 teams). In FY24 private companies outspent government on R&D (said: a quarter as much). Nayara Energy, the oil company that lost email access, went to the Delhi High Court (said: Bombay).