What qualifies me to be a venture capitalist is my experience. I have 26 years across consulting, research, investment banking, corporate banking, and on the buy side sitting on the boards of companies and helping them scale.

That career began on a debt desk, not a cap table. Lending to multinationals at HSBC and DBS, then mid-market corporates, taught Rajeev to read cash flow and downside before he ever underwrote upside. He moved to sell-side investment banking at Edelweiss, now Nuvama, and SMC, then to the buy side at FidelisWorld, a roughly $100 million consumer fund backed by Middle Eastern LPs. There, the instinct that would define Cactus was born: back the infrastructure, not the gamble.

Private equity or venture capital requires one way up. You don’t want fluctuations. So the strategy focused on the ancillary and the infrastructure side.

Check out the video of the conversation here or read on for insights.

The scar that built the thesis

At FidelisWorld, Rajeev backed Raw Pressery, the premium cold-pressed juice brand. The product was excellent and growth was fast, until COVID froze the capital markets and the company hit a cash crunch, forcing a sale to Wingreens Farms. It now runs at roughly 4x its acquisition-era revenue. But the durable lesson was about physics, not pandemics.

You’re effectively transporting volumes of water across the country. The costs are too much and it makes it very difficult to get your margins.

That insight rewired a later call. When portfolio company Auric started in Ayurvedic beverages, Rajeev pushed it toward effervescent tablets and lighter form factors to escape the same logistics trap.

His read on the Indian consumer is just as blunt. There is no single market of 1.4 billion buyers, he says, there are “three or four Indias.”

The top 10% of the population has the ability and the intent to pay. Below the top 100 million, it’s very difficult. You don’t have too much disposable income.

The data backs the instinct. Goldman Sachs projects India’s affluent cohort, those earning over $10,000 a year, will grow from about 60 million in 2023 to over 100 million by 2027. Cactus underwrites only to that top slice, where margins survive.

Why asset-heavy became the new asset-light

The genuinely contrarian move came in 2021, when the herd was paying up to 25x revenue for software. Cactus walked into factories.

We were early investors in manufacturing in 2021, when investors were giving crazy valuations. We look at the tailwinds. There’s a China plus one benefit coming our way.

The bet is paying off. The set-piece is Showroom B2B, where two former furniture entrepreneurs mapped India’s fragmented apparel supply chain.

They identify where the best yarn is made at the lowest cost. They send it to the next place where it gets dyed, the third where it gets woven. They go and aggregate unused capacity. There are lots of small factories using only 50% of capacity.

In February 2026, Cactus led a $17 million Series A (about ₹150 crore in equity and debt) into Showroom B2B, the firm’s 12th investment from Fund I. Elsewhere in the portfolio, electronics manufacturer Brandworks Technologies posted ₹258 crore in FY25 revenue at a 43% revenue CAGR; Cactus led its $7 million Series A in 2025, since extended to $11 million total. Intangles, a physics-based predictive-AI firm for fleets, manages over 400,000 vehicles across North America, Europe, Southeast Asia and the Middle East, predicting engine failures up to a month ahead with about 95% accuracy; Cactus followed on in its $30 million Series B led by Avataar Venture Partners. The clearest proof the model produces liquidity, not just paper marks, was an early exit from Rubix Data Sciences at a 48% IRR.

The discipline of saying no

Cactus runs every deal through a framework Rajeev calls the five Ts: Team, TAM, Tech, Traction, and Transaction. The last one is where the banker shows.

Every business is good at a certain price. If the business is looking really good but the valuation expectation is out of whack, we’d be happy to not participate.

This connects to the intellectual core of the fund. Rajeev enters after product-market fit, which changes what he is betting on.

We are basically underwriting growth risk. We are not taking mortality risk. Your biggest concern is, will you not be able to scale.

It also shapes how he reads aggregators. He strips out inflated GMV and prices the business on its real take rate, typically 5 to 6%, and its unit economics, not the throughput routed through its balance sheet. The same restraint governs his AI exposure, drawn from a gold-rush analogy that runs through his whole career.

We prefer the picks and shovels strategy to the gold prospecting one. The gold prospecting one is binary. The large names just become stronger and stronger.

So he funds application, not foundational models. Kapture CX, an agentic-AI customer-support platform in the portfolio, now draws 35 to 40% of revenue from its AI products. The logic returns, every time, to consistency over lottery tickets. He models a downside tranche returning 1.5 to 2x, a hard middle at 4 to 6x, and a handful of outliers at 8 to 12x, and accepts missing far more than he catches.

We will miss 50, 100 of them, but we have to do 15. We need to make sure these 15 give us the returns we need to give our investors.

Cactus Partners closed its first fund at over ₹630 crore, about $77 million, with 60% from domestic LPs including SIDBI, and reports zero write-offs to date.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

He agreed. Five weeks later, the investor pushing for the change walked away anyway.

It was like they dropped a brick into the pond, watched the ripples, and then ran away. They didn’t support us when we really needed it.

The man who lost his company this way went on to build the next one so it could never happen again.

Check out the video of the conversation here or read on for insights.

The engineer who never wanted to be a founder

Amrit’s drive runs three generations deep. His grandparents migrated from India to Kenya, then to the UK, working three factory jobs at once. His father became an optometrist with his own practice. Amrit, the third generation, returned to risk.

My grandparents worked really hard to make it better for my parents. My parents worked really hard to make it better for us. So what am I doing for the next generation?

He earned a First-Class Honours in Chemistry at Birmingham, top of his class, then hit the 2008 crash and the depersonalised graduate-recruitment machine.

I found the process so inhumane. After doing one application I said, this is not the path for me. I’m going to work out my own path.

A PhD in chemical engineering, a £500 university grant, and a UN scientist who told his research group the world was beyond saving pushed Amrit toward mission instead of fatalism. The product that came out of it was simple to explain. Conventional lithium battery modules are spot-welded and glued, so one dead cell kills the whole unit.

Imagine you’re driving and the tyre bursts. Imagine if that tyre was welded to the car and you couldn’t repair it. You had to throw the whole car away.

Aceleron’s patented compression technology used nuts and bolts, letting modules be opened, serviced, and reused. Amrit calculated the approach could extract up to ten times more value per cell than a welded battery, and the company reached a manual assembly capacity of one megawatt-hour per day with 35 employees.

The right technology, the wrong capital structure

The trap was never the engineering. Aceleron raised over £15 million across seven years from Mercia, the Business Growth Fund, and Toyota Mobility 54, with rounds nearly every year. By the 2022 Series B, Amrit and co-founder Carlton Cummins held roughly 30% combined. He names the structural cause directly.

In the UK the feeling is very much optimizing to reduce downside, not optimizing for maximum return. So they take larger stakes earlier.

That early dilution pre-determined the ending. Asked whether the cap table was the real villain, Amrit pointed deeper.

There are only two types of problems in business. People problems, and problems you don’t yet realize are people problems.

Aceleron entered administration in September 2023, owing around £880,000 to creditors. In April 2024, India’s Advik Hi-Tech, a tier-one auto supplier with roughly $131 million in FY25 revenue, bought the assets and IP. Amrit refuses to call any of it a tragedy.

It was one of the most challenging experiences, but genuinely one of the best gifts I could ever have received. I’d never want to go through it again. But it was a present.

From atoms to bits, and a company that needs no one’s permission

In mid-2024 Amrit founded Lorefully with two former Aceleron colleagues, Barry Diffin and Paul Jennings. The name comes from lore, knowledge passed person to person by word of mouth. The first product, an AI knowledge tool for field engineers, died in enterprise procurement because no multinational has a budget line for “knowledge management.”

The pivot that worked came from a free booth. Lorefully approached the organisers of InstallerSHOW, a UK trade event with over 31,000 attendees, who handed over exhibition space in exchange for a debrief. On the floor, human operators hold consented, structured conversations that AI transcribes and categorises, with human editors reviewing every output. It is research disguised as conversation.

Unlike a survey, where you ask one question and get a response, this is a conversation starter. We come back with a rich data set.

The insight underneath it is what Amrit calls thought equality.

So many people in the room have an informed view who never get the chance to express it in a way where it’s recorded.

The timing fits the market. Event-management software is forecast to grow from $17 billion in 2025 toward $96.5 billion by 2036, while post-event survey response rates sit near 5%. Lorefully occupies the gap as a live, on-floor extraction layer, and the numbers are moving. The company is on track for £500,000 in revenue this calendar year and expects to be profitable this quarter, on roughly £500,000 raised entirely from angels and zero institutional VC. The team is six full-time, supplemented by paid students through its Future Voices Programme, which gives newcomers a structured first rung. The platform has captured over 5.5 million words by May 2026, up from one million in January. Events range from 10 people to 45,000, with a pipeline of more than 100 events in 2026 and a first-of-its-kind National Shipbuilding Office study that gathered 500 structured inputs.

Bigger opportunities in Southeast Asia, with organisers running 500 to 700 events each, would mean turning the internal tool into a self-serve product. When the host suggested funding that through upfront customer payments or debt rather than equity, Amrit did not flinch.

Why would you want to repeat the same mistake twice?

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

That man is Dr. Siddharth Panwar. The pile of “scrap” he collected, one printout at a time, is now the largest proprietary neural dataset of its kind on Earth: over 60,000 hours of continuous EEG from more than 25,000 patients. It is the training fuel for MANAS-1, a 400-million-parameter foundation model that NeuroDx unveiled at the India AI Impact Summit in February 2026, India’s first sovereign foundation model for the human brain.

Check out the video of the conversation here or read on for insights.

The Instrument the West Gave Up On

Electroencephalography is the oldest brain-reading technology that exists, invented by Hans Berger decades before MRI or CT. In the 1960s and 70s, skilled technicians could read tumours and lesions from the wavy lines alone. Then MRI arrived. Structural pictures held up in court. EEG, which requires interpretation, did not. The skill died with its practitioners.

The West, in our opinion, has given up on EEG far too soon.

Siddharth’s frustration is not nostalgic, it is economic. An EEG machine in India costs about ₹3 lakh. An MRI requires tearing down walls to install. EEG is the only brain diagnostic that can realistically reach a primary health centre in rural India. His conceptual leap was to stop treating EEG as a medical chart and start hearing it as language.

I don’t see EEG as anything other than a language. It’s a sequence, very similar to speech. And in fact, there are 20 electrodes, so it’s like a chorus. Twenty people singing.

The implication is large. The same transformer architectures powering ChatGPT, the math of sequential probability and self-supervised reconstruction, work on brainwaves because both are sequences with grammar. MANAS-1 was trained the same way large language models are trained: hide portions of the signal, ask the model to reconstruct them, and let it learn the inherent syntax of neural activity. On the standard benchmark of 25,000 publicly available EEGs that any researcher can scrape, MANAS-1 is the best-performing model in the world. For epilepsy detection, it hits 80 to 90 percent accuracy on the EEG alone, and crosses 95 percent when paired with basic clinical questions.

The Raddi Moat and the Milestones

The competitive advantage is not the architecture. It is the data, and the data is the product of patience.

They delete it every month. It’s just getting washed out of their hard disks. It’s like raddi. Like papers we used to sell. You can buy it raddi ke bhaav.

There were no procurement contracts. Just a decade of waiting rooms and relationships. Eventually, two men noticed. Rajesh Kamra, an IIT Kanpur alumnus and founder of Koovs, and Sandeep Singh, an applied AI veteran from Walmart Labs, brought in their company Newron on a sweat-equity arrangement. Compute, engineers, an office. No venture dilution. Singh is now NeuroDx’s CTO.

🏆 In February 2026, MANAS-1 was unveiled at the India AI Impact Summit. NeuroDx (operating legally as Intellihealth) was selected as 1 of 12 IndiaAI Mission Champions, alongside Sarvam and the IIT Bombay-led BharatGen consortium, and the only physiological foundation model on the list. The model was released open-source on Hugging Face. The full cohort received a two-hour audience with the Prime Minister at his residence.

Taking on Neuralink, with Less Money

Elon Musk’s Neuralink drills holes in skulls. Siddharth listens from the outside. Invasive systems win on signal quality but lose on everything else: surgery, scarring, cost, geography. The global brain-computer interface market is currently valued in the $2.6 to $4 billion range and is forecast to grow at 14 to 18 percent annually through the next decade, with non-invasive systems holding the dominant share. NeuroDx is hardware-agnostic by design, the intelligence layer that can sit on any commercial EEG headset.

He’s doing it invasively. We are trying to do it non-invasively. What he’s doing invasively, we can get 70 to 80 percent of that done without surgery. And we can make it available to a person sitting in a village.

NeuroDx is raising $10 million. In Silicon Valley, this is a rounding error. Meera Murati raised at a billion-plus pre-revenue. Sarvam raised roughly $40 million pre-revenue. NeuroDx is being told $10 million is not possible because they have no earnings yet. Indian deeptech funding grew 58 percent in 2025 to $1.22 billion, and deeptech now represents 15 percent of Indian PE/VC activity, up from 4 percent a decade ago. Yet the appetite for pre-revenue, frontier scientific risk remains thin.

I am willing to die, but I will not scale down my ambition. If my failure has to inspire the next round of great companies, I am fine with that.

The Brain Scan Ahead

The pipeline runs well beyond epilepsy: ADHD and autism for children, depression and anxiety for working-age adults, dementia screening for over-55s. The vision is a five-minute headset session in any clinic that returns a same-day cognitive report. No prescription. No specialist queue. The cost structure of a blood test, not an MRI. MANAS-2, scaling to 2 billion parameters and over 100,000 EEG datasets, is what the round is meant to fund.

The most important AI lab of the next decade may not be in San Francisco or Hangzhou. It may be in a shared office in Bengaluru, run by a man who once sat outside a doctor’s room waiting to be trusted with the trash.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Inside a risk meeting at Indifi Technologies, somebody made a call that, depending on how you looked at it, was either insane or the most disciplined decision an Indian fintech had made in a decade. They turned the tap off for 30 percent of every customer they would have approved the month before.

We stopped lending to about 30% of the people we would have lent to in August 23, in one stroke. Our monitoring systems were doing their job, and our organisation culture was allowing us to listen to those signals.

Twelve months later, the rest of the industry would learn what Indifi already suspected. The 2024-25 cycle hit the unsecured MSME sector harder than anything since COVID. Industry credit costs escalated 70 to 80 percent. Indifi’s? Around 30.

The man behind that September call is Alok Mittal, Co-founder and CEO of Indifi.

Check out the video of the conversation here or read on for insights.

The 85% Problem

Alok is not a career banker. He is a coder by training, an early-career engineer at Hughes Software, and the Co-founder of JobsAhead.com, which Monster acquired in 2004. He then co-founded the Indian Angel Network and ran Canaan Partners’ India business from 2006 to 2015. Inside those deal rooms, the thesis for Indifi got built.

When we talk about entrepreneurs, we are only talking about venture-backed entrepreneurs. Whereas there are 60 million other entrepreneurs in this country who are running their business. And those entrepreneurs are underserved for capital.

SIDBI’s 2025 report puts formal institutional credit penetration among Indian MSMEs at 18 percent. The credit gap is roughly 30 lakh crore rupees. The retail and trading sector, which has almost no fixed assets to pledge, faces a 33 percent operational financing gap.

Eighty-five percent of Indians, your countrymen and mine, are not credit-unworthy. We just have not been able to figure out which ones are and which ones are not.

How Indifi Underwrites

A new NBFC borrows at 14 to 16 percent. A bank borrows at 5 to 6, because it gets to use savings deposits paying near-zero interest. Banks lend at 11 to 12 to the safest customers, leaving the rest for younger NBFCs that have to lend at 22 to 28 and be very good at telling who will pay back. Indifi’s average yield sits at 22 to 24 percent, risk runs at three and a half to four percent annually, and cost of funds is around 10 percent, achieved by diversifying across 37 institutional lenders.

The deeper reason traditional lending misses the middle is psychological.

Human underwriters are good at looking at a policy and saying yes or no. But they are not good at looking at a file and saying, is the probability of default here 2% or 5%? So you won’t serve the 5% risk customer, even if that customer is willing to pay you higher.

So banks default to narrow bands. Sub-2 percent risk, 17 to 20 percent pricing. Everyone outside is invisible. Indifi runs 50 to 60 parameters through machine learning to price risk across a wide spectrum and stays profitable on a portfolio basis.

Half of Indifi’s loan book originates through digital ecosystems: Swiggy, Zomato, Amazon, Flipkart, payment players. A Swiggy restaurant’s monthly receipts tell you everything about its ability to pay back. During demonetisation and COVID, lenders relying on branch verification went blind. Indifi did not.

The other discipline Alok hammers on is vintage pool analysis. NPA, the headline metric, is mathematically gamed by growth: new loans take 12 to 18 months to go bad, so doubling your book dilutes your NPA.

In a good cycle, every lending company looks good on those metrics. But if you look purely at one cohort, January to March 2023 originations, that cohort is fixed. How are losses rising in it? Are they stable across cohorts? That’s vintage pool analysis.

It is, conveniently, how Indifi spotted something the market had not, sometime in mid-2023.

The Red Button and What Survived

In September 2023, Alok and his team pressed it. AUM, which had peaked at 2,178 crore rupees in March 2024, drifted to about 1,998 crore by June 2025. FY25 closed with a 45 crore rupee net loss, weighed down by provisioning and an accounting transition. GNPA hit 4.56 percent in Q1 FY26.

What survived was the rest of the book. Operating revenue grew 22 percent year-on-year to 360 crore rupees. EBITDA climbed from 90 to 107 crore. Tier 1 capital adequacy stood at 23.92 percent, gearing at 2.2 times, both far stronger than the 15 percent regulatory floor. Indifi returned to operational profitability in Q2 FY26. In March 2026, BlackSoil extended a 40 crore rupee venture debt facility.

Across its lifecycle, Indifi has raised over 129 million dollars, including a 35 million dollar Series E led by ICICI Venture in June 2023, and disbursed over 497 million dollars across 73,000 loans to more than 55,000 MSMEs.

Our credit performance is actually better now than what it was in 2021–22. But the team is still very cautious, because their sense is that at the market level, this hasn’t sorted itself out.

The macro made the discipline look prescient. US tariffs on Indian exports climbed to 50 percent on textiles, gems, and leather through 2025, contracting exports to the US by 37.5 percent in the worst-hit segments. The Middle East conflict inflated freight, LPG, and fuel costs. Small manufacturers and kirana stores absorbed the squeeze, distress that does not show up in a Cibil score but does show up in a restaurant’s daily Swiggy receipts.

Indifi is not a unicorn. It has not filed for an IPO. What it has done is survive the first proper Indian credit cycle of the digital era with its capital structure intact. For a fintech founder in 2026, that may be the more interesting story.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

While Ninjacart’s operating revenue declined 19% year-on-year and WayCool laid off over 200 employees, FarMart quietly turned EBITDA-profitable in Q4 FY26 at a ₹3,600 crore annualised run rate, on $78 million of total capital raised. That is roughly a fifth of what its better-funded competitors burned through.

Check out the video of the conversation here or read on for insights.

The Grandfather and the Pivots

The trigger was personal. Alekh Sanghera’s grandfather, a six-decade farmer near Jalandhar, told his city-educated grandson that despite India’s tech revolution, farming had become an unrespected, unprofitable life. The data backed him up. Over 85% of Indian farmers are smallholders, fewer than 20% have institutional credit access, and machinery rentals alone consumed up to a third of seasonal earnings.

Alekh, who had spent his MicroSave years walking Western UP hinterlands auditing Direct Benefit Transfers for the Gates Foundation and the World Bank, understood why rural cash flows broke in ways most agritech founders never bothered to learn. In 2015, with boarding-school friend Mehtab Singh Hans, he founded FarMart.

What followed was four companies inside one. FarMart began as an Uber-for-tractors rental platform. It died on credit defaults. It pivoted to a BNPL credit card for farmers, profitable but impossible to scale before UPI rails matured. It pivoted again into a Khatabook-style ledger for village retailers. The product was sticky, but rural India would not pay for SaaS. Then in 2020, a small marketing-SMS feature inside the bookkeeping app went viral.

We had no idea about healthcare or education, so we picked agri. Then we built four different companies inside one before we figured out that the money in India is in commerce, not subscription.

The 10,000-plus village retailers using FarMart’s software were not customers to be monetised. They were a decentralised procurement army to be armed.

The Single Truck Theory

Today FarMart operates across 40-plus dry commodity categories, sourcing for Reliance Retail, Britannia, Parle, and a roster of biofuel distilleries riding India’s ethanol mandate. The mental model is simple.

Fundamentally, our unit is one truck. Imagine the multiplier effect of that single unit to get to 500 million dollars in scale. At any given moment, about one lakh trucks are on shipment. Lose one truck and your margin structure for the next seven days is gone.

Everything is engineered around bulletproof economics on one truck, then multiplied. Zero owned warehouses. Zero owned trucks, just 3,800 fleet partners. Zero retail storefronts, just SaaS-armed village aggregators. Ninety quality inspectors cover 110 districts as relationship managers. Computer-vision AI trained on five million images grades batches with 85-97% accuracy without anyone leaving the village.

The compensation philosophy is equally contrarian.

If you try to make money on the trade, you’ll never make money. Your spread will either hurt the farmer or hurt the buyer. We make money on offtake fees, platform fees, finance origination, logistics fees. The services stack is where the company lives.

If the model is so clean, why hasn’t capital replicated it? Alekh’s answer reframes the question. B2B agri is structurally a positive working capital business. Buyers pay in 30 to 60 days. Farmers demand instant liquidity. Most agritechs died trying to bridge this gap with their own balance sheets.

FarMart’s solution: securitise institutional invoices through SEBI-registered NBFCs, route capital from asset managers into farmer payments, charge origination fees of 0.5 to 1.5% to both sides. Britannia and Adani paper carry top ratings. Alekh turned the creditworthiness of his buyers into the working capital for his sellers. In domestic output linkage, only ITC is larger, moving 16-20 lakh tonnes annually against FarMart’s 11-12. Functionally a duopoly.

The Tailwinds and the Runway

Three forces are compounding in FarMart’s favour. India hit E20 ethanol blending in 2025, five years ahead of schedule, with E27 targeted by 2030. Grain-based ethanol capacity for ESY 2025-26 exceeded 13,000 million litres, eclipsing sugarcane molasses. FarMart became one of the largest grain aggregators feeding distilleries, quietly monetising India’s energy independence policy.

The Digital Agriculture Mission has generated 84.8 million farmer IDs by early 2026, collapsing verification cost the way Aadhaar did for fintech. And the 2.7 million sq ft Bharat Mart at JAFZA Dubai opens late 2026, where FarMart already runs an office and where its consumer brand FarMart Pantry, 27 SKUs across atta, basmati and millets, launched first before hitting Zepto and Blinkit shelves in India.

That last move changes the margin story. Bulk B2B trade caps gross margin at 11-12%. Processed exports hit 25%. Quick-commerce private label hits 35-40%.

Vertical integration is a moat. If you really want to be a very large, profitable business, the revenue has to be the sum of the parts. The B2B services give scale. The processing and brand give the absolute profit pool.

The numbers do the talking. ₹2,340 crore paid directly to rural nodes in 2025. 18,000 tonnes of food loss prevented. 23,000 tonnes of carbon emissions averted. Revenue per employee crossing $1 million. A spot on the Avendus-Hurun India U35 List 2025 and Fortune India 40 Under 40. The IPO timeline, by Alekh’s own articulation, is six quarters away.

We want to be a public company because we want to show this can be done. Globally, $18.5 billion has gone into agritech in the last decade and there haven’t been outcomes. We want to be the outcome.

In a funding winter, the founder who survives is rarely the one with the most capital. It is the one with the cleanest unit.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Check out the video of the conversation here, or read on for insights.

The Excelan pricing playbook

Kanwal’s first decade in America was three job losses in 24 months and a move west to what was not yet called Silicon Valley. He rose to senior staff engineer, but when he asked for a management track, the answer was blunt.

I asked my bosses about management. They said, ‘No, you guys are not managers. Who would work for you Indians?’

A subordinate he had trained, David Jackson, left to start Altos Computers. In 1982, Kanwal decided to do the same.

That year, IBM was shipping PCs at roughly $2,000, down from the $50,000 mini-computers and half-million-dollar mainframes that preceded them. Kanwal saw the obvious thing the market had not processed: desktops were useless alone. They had to talk to the mini-computers and mainframes where corporate data actually lived.

He co-founded Excelan with two partners, each putting in $10,000. VCs said no a hundred times before John Bosch wrote a $2 million cheque for 50 percent of the company. Bosch later made more than 100x his money.

The product bet was contrarian. Intel, Xerox and Digital had announced Ethernet as the future standard, with chips promised in three to five years. Kanwal built a board-level Ethernet implementation immediately, using off-the-shelf components. Then he bundled the Department of Defense’s TCP/IP protocol onto it, a move the industry considered a category error.

The conventional wisdom was that TCP/IP was ill-suited for Ethernet. It was built for a slow, error-prone military network. My logic was: if there are no errors, there are no retransmissions. TCP/IP has no inherent limit on speed. I was the only person who bet on it.

The product worked. The company did not, at first. The CEO went hands-off, the board fired him, and Kanwal stepped in with a pricing turnaround. He replaced à la carte pricing (a $2,000 board, $60 software, transceivers bought elsewhere) with a single bundled solution at $14,995, exactly half of Digital’s $30,000 equivalent. Hardware cost of goods was roughly $300 per unit. Gross margins moved to 80 to 90 percent.

In 1987, Excelan became the first Indian-American-founded, venture-backed company to IPO on NASDAQ. Revenue was $22 million, profit was $3 million, valuation was around $125 million, and the stock popped 50 percent on opening. One detail worth preserving: weeks before the IPO, the board temporarily replaced Kanwal with a white CEO because “there was a fear an Indian CEO may not play well on Wall Street.” The replacement was fired six months later.

In 1989, Excelan merged with Novell in an all-stock deal worth roughly $210 million. Novell’s stock 10xed within a year, eventually reaching a $12 billion market cap, making it the world’s second-largest software company after Microsoft. Kanwal effectively ran Novell by his late 40s, bought Unix from AT&T, and was passed over for CEO when Ray Noorda retired in 1995. On his way out, he made Unix open source, a decision that seeded Linux and the modern cloud.

TiE, and the arithmetic of angel investing

In December 1992, eight Indian-American executives were waiting at the Santa Clara Marriott for a bureaucrat who had missed his flight. While they waited, they started talking.

We realised we all had similar journeys. Nobody believed in us. Someone suggested we form a dinner club, not to invest, but to encourage our youngsters to become entrepreneurs instead of job seekers.

That dinner club became TiE, The Indus Entrepreneurs. The first workshop in March 1993 expected 100 attendees. 500 showed up. Today TiE operates 61 chapters across 14 countries with roughly 15,000 members, including 3,000-plus charter-member mentors. TiE also lobbied Delhi in 1999-2000 to create the Foreign Venture Capital Investor asset class, the legal plumbing for modern Indian VC.

Kanwal became an investor because the first generation of Indian founders needed capital the Valley would not write. An early angel cheque into Exodus Communications peaked at 1,000x during the dotcom bubble. He sold on the way down and settled for 86x.

I felt like an idiot, because at one point it had been a thousand times. Later I realised 86x isn’t bad. But it’s not a thousand.

In 2007 he co-founded Inventus Capital Partners, which has raised about $208 million across three funds, seeded nearly 90 teams, and produced seven unicorns including Headspin and Cohesity, alongside exits like Poshmark (NASDAQ IPO, 2021), redBus (acquired by Naspers for $138 million), Nutanix, PolicyBazaar, and Sierra Atlantic. His current vehicle, Silicon Valley Quad, is a four-partner syndicate writing $2 to $3 million seed cheques for 25 to 30 percent of the cap table, with a deliberately harsh assumption: 75 percent will fail, so the winners must return 40 to 50x.

His founder-picking framework is built on elimination, not prediction.

I bet on the person, not on the TAM. Technology changes every day, competition emerges every day. The only constant is the person. I can never tell you who will win. But I can tell you, with 90 percent confidence, who is not going to.

The screen: intellectual honesty (owns failures), humility (arrogance is an auto-no), fairness in equity allocation, willingness to hire people smarter than yourself, and revenue-per-employee literacy for software founders.

10 million founders by 2047

Kanwal turned 80 in August 2025. His memoir The Groundbreaker was published by Diversion Books in February 2026, with a foreword by Congressman Ro Khanna. He has donated $5 million to Michigan Tech and $3 million to IIT Bombay, which named the Kanwal Rekhi School of Information Technology after him.

His stated goal is to live to 102, to see India’s 100th anniversary of independence. His current institutional bet is KREST, a Rural Entrepreneurship Center in Nizamabad. India, he argues, will not be built by 200 unicorns in a handful of metros. It will be built by 10 million entrepreneurs, many in tier 3 and tier 4 cities. One of his examples is an IIT graduate who returned to his village, redesigned primitive farming tools, doubled productivity, and now sells to 70,000 farmers with revenue around 5 crore and a line of sight to 50 crore.

Per-capita GDP in the US is roughly 30 times India’s. Kanwal’s thesis is that the gap only closes if the founders come from everywhere.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Somewhere on Alok Goyal’s second monitor, there is a spreadsheet. It does not track portfolio valuations. It lists every company that went on to become great - that he evaluated and passed on.

One name on that list is Postman, the developer tooling company now valued in the billions. Alok understood APIs. He just could not imagine the market getting large enough. That failure of imagination, he says freely, was the lesson.

This is the unusual thing about Alok Goyal, co-founder and Partner at Stellaris Venture Partners: he treats his misses as primary data.

Check out the video of the conversation here or read on for insights.

From Broke Consultant to COO to VC

Alok describes his career as “Brownian motion,” and the description holds. He started a PhD in video streaming at the University of Texas at Austin in 1992, quit midway, came back to India to start a company that failed, then pivoted into strategy consulting at McKinsey India and later the McKenna Group in Silicon Valley - a firm born alongside Intel and Apple. When the 2001 dot-com crash hit, the NASDAQ fell from 5,000 to 1,000, and his firm, which served only high-tech clients, ceased to exist. He spent 18 months unemployed and broke, carrying the weight of two MBA loans.

He finally took a sales job he didn’t want.

It’s like an experience. Until you try, you don’t know whether you like something. I really loved the kick of doing a deal.

That accidental discovery led to nine years at SAP India, eventually as COO, closing roughly 2,000 enterprise deals a year. When personal circumstances brought him back to India and SAP began pushing him toward Germany, he worked with a coach to ask what came next. One hypothesis: early-stage tech investing. In January 2013, he joined Helion Venture Partners, barely knowing what Flipkart was.

Helion was, by Alok’s telling, among the best firms in India at the time - home to MakeMyTrip, India’s first venture-backed IPO. But venture firms are not corporations. Partners develop different views on economics, decision-making, and succession. In 2016, Alok left with Ritesh Banglani and Rahul Chowdhri to start Stellaris. They were the first to break away from a first-generation Indian fund in nearly a decade, and four independent funds ultimately emerged from Helion’s diaspora, including Fireside Ventures and Fundamentum.

Stellaris launched with $100M in Fund I (2017), grew to $225M in Fund II (2021), and closed a $300M Fund III in 2024-25, with up to $150M earmarked for AI-native startups. Total AUM now exceeds $600M.

Three Mistakes, Repeated

The investment philosophy at Stellaris is built on focus: 8 to 10 deals a year, over 60% at inception stage - before a product is built, sometimes before a line of code exists. But Alok is more interesting on what he has consistently got wrong than on what he has done right.

After 13 years and roughly 4,000 companies evaluated (the firm invests in approximately one out of every 400), three anti-patterns keep showing up.

The first: over-indexing on market size. Great founders expand their market’s definition over time. Trying to size it at inception is often the wrong question.

The second: weighting market over founder when the category is undefined. In technology, many markets don’t exist yet at the time of investment. In those spaces, the quality of the person is the only real variable.

The third is the one that has cost him the most: fearing large incumbents. The question changes costume across eras - “Why can’t SAP do this?” becomes “Why can’t Google?” becomes today’s version: “Why can’t OpenAI?”

Large companies have so much baggage, friction, low decision-making speeds. Even when the answer is obvious, it never really happens. Almost always, when I have over-indexed on large competition, I have been wrong.

The 22 Believers

The clearest illustration of Alok’s founder-first thinking is Axtria, a pharma analytics company founded by Jassi Chadda. When Alok backed him, Jassi had zero revenue after 18 months, a misread market, and a high-pressure investor. His response was to return the investor’s capital. The recapitalisation vehicle: all 22 employees put in their own savings to fund the buyback. Most had followed Jassi from his previous company. They quit their jobs to do it again.

How insane a belief those 22 people would have had in that founder. He is a guy you follow to war knowing you might not make it. That was the prime part of my investment thesis.

In September 2025, Kedaara Capital invested $240M in Axtria via a secondary transaction, delivering significant returns to those early believers.

The broader Stellaris portfolio reflects the same conviction. Whatfix, backed when there were only two founders - one building, one selling - raised a $125M Series E led by Warburg Pincus in 2024, reaching approximately $900M in valuation with a 4.5x increase in ARR since 2021. Mamaearth became India’s first D2C brand to go public in 2023. Stellaris was in at Series A.

Work, Not Humans, at the Centre

When ChatGPT launched in November 2022, Alok recognised the wave but misread its pace. He had seen cloud, mobile, and the early internet. Those pattern libraries, he says, became liabilities. His response was deliberate: he wrote no new checks in 2024, using the year to understand the shift rather than react to it.

The insight he arrived at is structural. Old software made humans better and charged per seat. New software does the work and charges per outcome - what he calls the move from co-pilot to agent.

A human was still at the centre of my imagination for software. But I think work is at the centre today. It is not necessarily the human.

Portfolio bets like Pibit.ai reflect this: an AI-native insurance underwriting platform that raised $7M in Series A in November 2025 and claims to cut underwriting cycle times by 85%, charging for work delivered rather than seats sold. With the US-India Interim Trade Agreement announced in February 2026 creating cleaner market access for Indian SaaS firms, the macro conditions for this thesis are better than they have been in years.

Alok is still willing to hold his own framework up to scrutiny. Asked whether his US-market preference contradicts his stated belief in backing founders over markets, he doesn’t deflect.

Both our assets and our liabilities are our experience. All learnings are biases by definition.

That willingness to name his own blind spots, in public, after 13 years, is what makes the spreadsheet of misses worth paying attention to.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

That floppy disk became Quick Heal.

I never thought of it as an antivirus product altogether. Customers would come in with systems that wouldn’t boot, files behaving strangely. I started writing small utilities to remove infections. That’s all it was.

Sanjay traced the full arc of a company built without venture capital for most of its life, without a distributor willing to back it, and without a product category India had yet learned to pay for.

Check out the video of the conversation here or read on for insights.

The Rejection That Built a Network

Quick Heal version 1.0 launched in 1995, priced at Rs. 450-500. For nearly three years, it sold almost nothing. Piracy was rampant. People paid for hardware, not software. But Sanjay kept improving the product, adding features global competitors had ignored: resumable updates for slow Indian connections, bootable tools for users running unlicensed operating systems.

By 1998-99, small offices and DTP shops in Pune were paying. The bigger test arrived when Quick Heal tried to scale nationally. Every major distributor - firms handling Oracle, Microsoft, and McAfee - turned them away. No one would sell an Indian software product.

Nobody was willing to sell a product built in India. They never trusted that something of that quality could come out of here. But that was a blessing. It forced us to build our own network.

What followed was a city-by-city expansion that most funded startups would not attempt. They opened a branch in Nasik first - not Mumbai - to test the model cheaply. It broke even within three to four months. Then a new city every six months, each branch capturing the market before moving to the next. By 2009, Quick Heal held 30% market share in paid antivirus across India, with roughly 50 internal salespeople managing a channel ecosystem that would eventually grow to 25,000 partners. Revenue was approaching Rs. 100 crore. The company had never raised external capital.

The IPO, the Margins, and the Enterprise Bet

In 2010, Sequoia Capital approached them. Sanjay did not go looking. The Rs. 60 crore infusion mattered less than what came with it: board structure, financial discipline, and the governance machinery that makes a public company possible.

In 2016, Quick Heal listed on the NSE and BSE - the first Indian cybersecurity software product company to go public. For Sanjay, the significance was practical rather than symbolic.

Being listed is a different feeling altogether. Every decision is under public lens. You’re accountable not just to customers but to thousands of shareholders. That changed how carefully we take decisions.

Listing gave Quick Heal something no marketing budget can manufacture: share value as acquisition currency, credibility in enterprise procurement, and trust signals that matter when bidding for national infrastructure contracts. In November 2025, the company secured a Rs. 64.25 crore contract with the NFSU Research and Innovation Council - a direct product of that institutional standing.

The financial story since listing has been one of deliberate reinvestment. At peak consumer antivirus, Quick Heal ran EBITDA margins close to 45%. Today they sit in the low single digits, as capital moves into next-generation enterprise security under the Seqrite brand. FY2024 revenue was Rs. 313 crore, with the current year projected around Rs. 350 crore and market capitalisation near Rs. 2,500 crore. The consumer business, once 95-99% of all revenue, has been compressed by OEMs bundling security into operating systems. Sanjay does not dispute this.

The enterprise bet is the counterweight - and the numbers justify the pivot. Enterprise renewal rates exceed 80%. Consumer renewal rates have fallen to 20-25%. The math is decisive.

Thirty Years of Escalating Threats

Sanjay’s account of how threats have evolved is less a technology briefing and more a study in how criminal industries mature. Early viruses were nuisance code, written to demonstrate skill. Email and broadband made malware global in minutes. Then monetisation arrived: credential markets, dark web data sales, ransomware-as-a-service.

The 2022 attack on AIIMS Delhi is the example he returns to. Months of silent reconnaissance preceded a single-night strike that locked every OPD, operating theatre, and administrative system simultaneously. Recovery took months, not days.

Fraud is becoming more personal, more convincing. Earlier, scams were generic. Today, messages are tailored, voices can be cloned, videos can be faked. AI is being used to scale deception.

Quick Heal’s response includes GoDeep.AI for predictive threat detection, SIA - an AI assistant that lets non-technical administrators configure security policy in plain English - and a recently launched Digital Risk Protection Service that monitors dark web activity, fake apps, and fraudulent social media handles outside an organisation’s own network. The DPDP Act, with its rules notified in November 2025 and a compliance deadline of May 2027, adds regulatory pressure that directly accelerates enterprise security adoption. Non-compliance penalties reach Rs. 250 crore.

Sanjay’s counsel to founders, shaped entirely by what almost ended Quick Heal in its first decade, is consistent with how he has always operated.

Whatever you’re good at, focus on that. Get the maximum market share. Scarcity makes you focus only on the pain points customers are actually paying for - not the fancy ones.

Cybercrime costs are projected to reach $10.8 trillion globally by 2026. The man who first debugged a virus in 1990 on a borrowed floppy is now defending hospitals, banks, and power grids. The product category has changed beyond recognition. The instinct that built it has not.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Ahmedabad now has over 500 pickleball courts. India has roughly 2,500. The United States has north of 50,000. For Siddhant, that gap is not a sporting statistic. It is the entire business.

Check out the video of the conversation here or read on for insights.

The sport that makes the economics work

Pickleball started in 1965 on Bainbridge Island, Washington, invented by two friends with bored families and a backyard. For decades it was dismissed as a sport for older people. Then, sometime in the 2010s, it stopped being that.

It is played on a court the size of a badminton court. The paddle is flat and solid, slightly bigger than a table-tennis bat. The ball is plastic with holes, which limits bounce considerably compared to tennis. The serve must be underhand. Both sides must let the ball bounce before volleying in the opening exchange. A seven-foot “kitchen” zone near the net forces short, tactical play rather than power.

The things that make pickleball very age-agnostic and very easy to pick up are the two-bounce rule and the kitchen. Because of these, the sport becomes easy to play for anyone.

A 10-year-old and an 85-year-old can rally on the same court. Games take 45 minutes. You need four people. That accessibility is the commercial foundation of everything Siddhant is building.

The yield argument

The central insight Siddhant returns to is yield per square foot. A standard tennis court footprint can fit four pickleball courts. Four times the players, four times the bookings, on the same piece of land.

₹5–7L

Construction cost per court

₹9–10L

Gross monthly revenue, 4-court facility

8–10 months

Typical break-even period

All figures stated by Siddhant Jatia in the Founder Thesis podcast.

After rent, electricity, and manpower, a well-run four-court facility nets roughly ₹3 to ₹3.5 lakh a month. The total build is around ₹20 lakh. Divide one by the other and you have an 8-to-10-month payback. That is not a sports investment. It is a real estate monetisation strategy.

A lot of people look at this like a lottery ticket. But unless certain fundamentals are in place, the right location, the right income group nearby, the right construction standards, the end product isn’t great, and it takes a toll on the player experience.

The locations being converted are varied: hotel podiums, school grounds open for pay-and-play after hours, rooftops, and in one detail Siddhant mentions without much ceremony, legally disputed land parcels where owners can generate returns while proceedings continue. The infrastructure is modular. The payback does not wait.

When I suggested this sounded a lot like the co-working industry, Siddhant did not pause. “100%,” he said. “You take underutilised real estate, unitise it, earn more per square foot, and layer services on top.”

What Picklebay is actually building

Siddhant launched Picklebay in May 2025, bootstrapped from his own capital after seven months of development. The founding frustration was personal. He had become a regular player and found the entire ecosystem running on sprawling WhatsApp groups, unverified court listings, and events with no digital infrastructure.

The entire pickleball community was only on WhatsApp. People were finding courts, finding players. I saw a real business here. I wanted to go vertically down, not multi-sport, and create a tech stack that solves the pain points I personally faced as a player first.

Picklebay now lists 650 to 700 courts across seven cities. The platform physically verifies each facility, flagging courts that fall short of regulation dimensions. That verification is not a feature. It is the product. There is an infrastructure pipeline of 750 courts from operators seeking construction consultancy, and Siddhant says he is engaging at least 10,000 players across the company’s own WhatsApp communities.

On the software side, Picklebay has built a Tournament Management System (TMS) that digitises the entire event lifecycle, already deployed in the Picklebay India Tournament Series 2025. A Venue Management System (VMS) is in pilot ahead of a mid-2026 rollout, giving venue owners dynamic pricing tools and occupancy analytics. Siddhant describes the logic simply: “It’s like an airline managing inventory. You’re optimising yield end of the day.”

The missing layer he is building toward is larger. In hospitality, channel management software gives a hotel owner one interface to update rates across all booking platforms simultaneously. In sports, no such layer exists. Every aggregator, Huddle, District by Zomato, Playo, gives operators a separate backend. Siddhant wants to build the system above all of them.

The longer logic follows from there. Once the tech stack is proven for pickleball, you launch PaddleBay for padel, ShuttleBay for badminton. Vertical depth first, horizontal replication second.

Why this moment matters

In April 2025, the Ministry of Youth Affairs and Sports officially recognised the Indian Pickleball Association as the sport’s National Sports Federation. The Delhi High Court upheld that status in February 2026. With NSF recognition, the IPA can now sanction leagues, standardise coaching certifications, and build toward Olympic inclusion, which the global pickleball community is targeting for 2032 or 2036. India has submitted a letter of intent to host the 2036 Summer Olympics in Ahmedabad. The alignment is not accidental.

Siddhant’s cultural argument for the sport is worth taking seriously.

Pickleball has the potential to become the new gully cricket of urban India. When we were growing up, you’d call your friends, play a quick game in the evening, spend an hour and a half. You didn’t need anything. That’s what gully cricket was. Pickleball is that, but organised and built for the city.

Gully cricket has largely disappeared from urban India. Space is gone, scheduling is hard, and you cannot get eleven people to commit. Pickleball needs four. It costs roughly ₹250 a head split across a booking. It requires no prior athletic ability and no serious fitness. Siddhant thinks it can become the second most-played sport in India by participation, not spectatorship, not leagues. Just by the number of people actually playing.

He has been building for a year. The sport is still in its early urban phase, still played primarily by upper-middle-class India. The infrastructure, the software, the governing body, the Olympic conversation, and the wellness shift among younger generations are all moving in the same direction. Siddhant’s bet is simply that he built the rails before the train arrived.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Check out the video of the conversation here or read on for insights.

The Industry’s Original Sin, and the 85% Number Nobody Talks About

The story of Indian wealth management, as Manu tells it, is a story of structural conflict dressed up as service. In the decade from 2000 to 2010, foreign banks built the relationship manager model in India. High-trust, low-transparency, and enormously profitable. Then 2008 happened, and clients began discovering the gap between what they’d been promised and what they actually owned.

The damage that followed was concentrated in the 2012 to 2018 cycle, when real estate AIFs promising 17% to 20% returns proliferated. Almost two-thirds of them failed to perform anywhere close to those promises. Some didn’t return principal. Family offices were born not from aspiration but from broken trust.

But the flaw Manu spent 20 years watching accumulate goes deeper. Wealth distributors who want to scale inevitably pivot to manufacturing their own financial products. Once you manufacture the product, you can never objectively advise a client to exit an underperforming fund you created. The conflict is architectural, not accidental.

Centricity’s founding decision was to never manufacture. Manu built an open platform working with 80 to 85 alternate asset managers, eight to ten product options in every investment category, no proprietary product to protect.

That decision led directly to the number that defines his entire thesis.

Nearly 80 to 85% of PMSs and AIFs across genres actually underperform the index, over 15 years of data. Only the top 10 to 15% beat it. That’s the hard reality. One has to be very cautious.

The industry sells alpha, the promise of beating the market through superior fund manager skill. Manu’s platform has processed lakh-crore worth of transaction data, and the data contradicts that premise, consistently. His practical conclusion: a wealth manager’s job is not to find the next star fund. It is to build a portfolio that stays above the median and keep the client from panic-selling during a 30% correction.

Lazy investors actually make money. Every time you do a transaction, buy and sell and trade, how much money do you actually make? You don’t.

Two Platforms, One Bet

Centricity runs two platforms on 80% shared infrastructure.

One Digital is the B2B2C engine, built for India’s 17,000+ independent financial advisors and mutual fund distributors. These are micro-entrepreneurs with strong local client relationships in Tier-2 and Tier-3 cities, but historically limited to mutual fund distribution. Manu gives them institutional-grade access to PMS, AIFs, unlisted equities, bonds, and pre-IPO investments, along with CRM tools, portfolio reporting, and ongoing training. Training, he says, accounts for 40% of the platform’s proposition.

Full plug-and-play. Like a co-working space. You take your laptop, go and sit, everything is available.

Invictus is the B2C engine, targeting “First Families,” portfolios of Rs 100 Crore or more. Each private banker manages a maximum of 20 families, a hard cap, not a soft guideline. The reasoning is practical: if markets are down 2% and you have 40 clients to call before the session closes, you cannot do it well. Manu’s private banking team, 30 senior bankers hired in the second half of 2025, operates at that boutique ratio across Delhi, Mumbai, and Bengaluru.

The metrics behind the business are worth pausing on. Centricity reached operational break-even in February 2024, just 11 months after commencing formal operations in April 2023. AUM has moved from approximately Rs 4,500 Crore in FY24 to approximately Rs 10,000 Crore ($1.2 billion) in FY25, with a target of Rs 18,000 to 20,000 Crore in FY26. Revenue went from Rs 20 Crore in FY24 to a projected Rs 60 to 100 Crore in FY25, targeting Rs 175 to 200 Crore in FY26. The monthly SIP and STP recurring book has crossed Rs 50 Crore per month, and 60% of total AUM is classified as Annual Recurring Revenue, the slow compounding kind that survives market downturns.

In September 2024, Lightspeed India Partners led a $20 million Seed round at a $125 million valuation, a sixfold jump from Centricity’s $20 million pre-seed valuation of 2022. Co-investors included the MS Dhoni Family Office, Burman Family Office, Paramark VC, NB Ventures, and MMG Group. Manu’s three-year targets: 50,000 to 60,000 distributors on One Digital, up from 17,000+ today, and 1,200 to 1,500 First Families on Invictus, up from 135 to 150 currently.

The Philosophy That Holds It Together

What distinguishes Manu from most founders building in financial services is where he places the emphasis. Not on growth, not on scale, but on what he can and cannot control.

Markets are not controllable. Geopolitical is not controllable. Asset allocation and risk, that I can control. If I’m not doing my basic core job of what is in my control, believe me, I’m just not doing the job.

This applies to hiring. Manu screens out high-revenue bankers with a transactional mindset and hires for longevity and client engagement. His CTO, Kamal Kishore, comes from Tata Digital’s AI engineering practice, brought in specifically to embed machine learning into portfolio risk visualization. His compliance head is expected to be “drop dead serious.” His finance head is expected to push back on every expense.

The bet Manu is placing is long-dated: that an industry built on opacity, product pushing, and one-time commissions will lose share over time to one built on ARR, transparency, and compounding client relationships. It does not pay off in a quarter. It pays off the way good portfolios do.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Manu J. Nair, co-founder and CEO of the Ethereal Exploration Guild (EtherealX), has a precise word for that situation. He does not call it a monopoly.

What we have today is not market dominance. It is market dependence. That is not how we envision the future of space.

Check out the video of the conversation here or read on for insights.

The distinction matters more than it sounds. SpaceX’s Falcon 9 is not blocking anyone. It is simply too busy. With the Starlink constellation consuming much of its own launch cadence, every other satellite operator is essentially a guest on someone else’s vehicle, booking whatever slots remain, often designing their product around a bottleneck rather than solving it.

EtherealX, incorporated in May 2022, was built to remove that bottleneck. Their vehicle, the Razor Crest Mk-1, is a 71-meter fully reusable medium-lift rocket, roughly the height of a 20-story building, capable of carrying up to 24.8 tonnes to low Earth orbit in expendable configuration and 8 tonnes when both stages are recovered. No other company in the medium-lift category is attempting full, two-stage reusability.

The market appears to agree that this matters. Manu and the team closed a $20.5 million Series A in December 2025, co-led by TDK Ventures and BIG Capital with participation from Accel and Prosus, oversubscribed from an initial $15 million target. Post-money valuation: $80.5 million, a 5.5x jump. Total capital raised across four rounds stands at $26.3 million. More telling is the $130 million in signed launch contracts secured from six international customers, including Japan’s SpaceBD, Taiwan’s Space Agency, and SABR, a European government launch aggregator, all before the rocket has flown once.

Those customers are not buying a rocket. They are buying insurance against a world where one country controls access to orbit.

The Physics Trick at the Heart of the Business

The reason EtherealX’s economics can work where others have struggled comes down to one engineering decision: what to do with the upper stage.

On a standard Falcon 9 flight, the upper stage costs roughly $12.5 to $15 million to manufacture and is expended after delivering its payload, burning up in the atmosphere. That figure represents 30 to 35% of total vehicle cost per flight. Falcon 9’s partial reusability, where only the booster returns, set a new industry standard. But it did not solve the upper stage problem.

Every company building a reusable upper stage treats re-entry heat as the enemy, relying on heavy ablative shielding or fragile ceramic tiles to survive it. Manu’s team asked a different question entirely.

Why fight the re-entry heat when we can utilize it? The damage is also advantageous, until a point where it is not. Then you replace. It basically follows a bell curve, just like an afterburner on a jet.

The result is the Full Flow Segregated Cooling Cycle (FFSCC), a proprietary engine feed cycle that routes re-entry plasma heat through the system to drive the engine turbines, executing a propulsive braking burn on the way down. Every kilogram of heavy shielding eliminated becomes a kilogram of revenue-generating payload. Manu claims this is the first new rocket engine feed cycle developed in six decades of rocketry, a claim that is pending IP protection.

The only globally comparable effort is Stoke Space in the US, which recently crossed $1 billion in total funding and is working toward upper-stage recovery in the small-lift category. Manu is the only founder pursuing this in the medium-lift segment.

From a Dinner Table to 150 Acres in Andhra Pradesh

Manu’s path here was not a straight line. Trained as a mechanical engineer, he was one of 12 candidates selected globally for Project PoSSUM, a NASA-aided scientist-astronaut program. He went on to work at ISRO’s Human Spaceflight Centre, where he met co-founder Shubhayu Sardar, a 10-year ISRO veteran. At Manastu Space in Mumbai, he met co-founder Prashant Sharma, who led propulsion. The company started, as Manu tells it, over a dinner conversation and money borrowed from his father.

Three years on, the infrastructure is real. BASE-001 in Cuddalore, Tamil Nadu, is India’s highest pressure-rated privately built rocket engine test facility, constructed from scratch because no adequate public infrastructure existed. BASE-002, on a 150-acre parcel in Space City, Andhra Pradesh, acquired with state government support, will handle integrated manufacturing, assembly, and stage-level testing for the full vehicle, located within an hour of the launch pad.

The Technology Demonstrator Vehicle orbital launch is targeted for 2026 to 2027. Manu has been deliberate about skipping suborbital test flights: the FFSCC re-entry profile only produces meaningful data at actual orbital velocities, around 7.8 kilometers per second. Anything less does not replicate the thermal conditions the system is designed for.

The recognitions have followed the milestones. In 2025, Manu and Prashant were named to Forbes 30 Under 30 Asia in the Industry, Manufacturing and Energy category. EtherealX received the Aegis Graham Bell Award from India’s Ministry of Electronics and Information Technology, was listed among Forbes India’s Select 200, and became the first startup outside the US and Europe invited to Beyond Gravity’s Launchpad Program.

On pricing, Manu’s target starts at $350 per kilogram, against an industry average of $5,000 to $6,000 today. A fleet of two to three vehicles, each targeting a 72 to 96-hour turnaround between flights, is how he plans to get there. His advice to the deep tech founders watching all of this is characteristically direct.

Not every great idea is a money-making idea. And not every money-making idea is a great idea. Building a launcher company is different from building a venture-backed, commercially scalable launcher company. People do not realize those are two different things.

The rocket does not exist yet. The land does. The contracts do. The test campaigns begin this year.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

India’s healthcare infrastructure has a design problem hiding inside a geography problem. While 70% of the population lives in rural areas, close to 65% of healthcare infrastructure is concentrated in cities. The government’s five-tier system was built to bridge this, but it runs without enough doctors, because, as Priyadarshi Mohapatra explains plainly, market forces are more powerful than policy mandates.

Only 1% of people get through government college PG seats. The rest take on huge debt through private institutions. When a doctor is reeling under a debt trap, how do you expect them to extend empathy to others?

This structural diagnosis is what Priyadarshi built CureBay around. But the spark was more personal. During the COVID-19 lockdown, Priyadarshi, formerly Country General Manager at Microsoft and a Google Cloud executive, needed a teleconsultation for his wife. The private hospital asked for his Skype ID.

I used to head Microsoft’s consumer business. Skype was a product of mine. Who is using Skype in today’s day and age?

He started working on his thesis the next day. CureBay was founded in 2021.

Check out the video of the conversation here or read on for insights.

The model, and the product nobody believed

CureBay operates 200+ eClinics across Odisha, Chhattisgarh, and Jharkhand, each running on a monthly opex of ₹35,000 to ₹55,000, covering rent and the salaries of two certified healthcare workers, a nurse-phlebotomist and a pharmacist. The patient experience is deliberately designed to feel like an urban polyclinic: vitals first, clinical note, then a doctor, who is remote 80% of the time. A digital stethoscope connects the nurse’s hands with the doctor’s eyes and ears on an integrated platform. Prescriptions are digital. Diagnostics follow a Pareto logic: a point-of-care machine handles the 20% of tests that cover 70% of rural requirements. For the rest, blood is drawn, centrifuged on-site, cold-chained to partner labs, and the report arrives on WhatsApp.

Above this sits a network of 1,500+ Swasthya Mitras, retired teachers and NGO workers who carry a gig app into village households, booking consultations and running AI-powered screenings before anyone walks into a clinic. For tertiary referrals, a CureBay concierge accompanies the patient to a partner hospital and handles the admission paperwork, solving the problem that causes most rural patients to give up after a first wasted trip.

Then there is Kavach. For ₹499, members receive unlimited GP consultations, limited specialist consults, ₹1,000 in Cure Coins redeemable against medicines and diagnostics, ₹500 per day hospitalization cash cover, and access to an interest-free emergency loan. Priyadarshi’s answer to the “this can’t work” question is the most revealing thing he says in the entire interview.

This is designed on the same principles as insurance. You take an assumption of utilization, build in checks so there is no misuse, and price it so you’re still in the money. A consultation is free, but you still need a doctor’s prescription to get the medicine. That’s the right thing in healthcare anyway.

The hospitalization cover is backed by a partner insurer. The interest-free emergency loan is subsidized using hospital referral fees. But the real insight Priyadarshi makes is simpler: the product only works because the clinic already exists. Insurance companies have more capital but cannot fulfill a rural healthcare promise at a patient’s doorstep. CureBay can.

110,000+

Active Kavach members

60%+

Kavach renewal rate (first cohort)

₹128.4 Cr

ARR, late 2024

18-20 mo

Clinic breakeven

First operational circles in Balasore and Puri (Odisha) are already profitable at the cluster level.

The business case, and what the money signals

CureBay has raised approximately $37 million across five rounds. The Series B in May 2025, $21 million led by Bertelsmann India Investments with British International Investment (a UK sovereign fund) and Elevar Equity, puts the post-raise valuation at roughly $75 million. There is no unicorn status to announce. What matters more is what the investor profile signals: a sovereign fund and a global institutional investor do not back rural India plays out of charity. They back them when the unit economics hold.

Priyadarshi’s answer when asked about the pitch is worth quoting directly.

There is no two different kinds of pitch. Rural India is a billion-people problem. We are solving it commercially. You need non-trivial money to solve non-trivial problems, and to get that money, you have to show people they will make money.

CureBay’s 1.4 million patient profiles, collected through 200 clinics staffed by trained workers rather than scraped synthetically, have also produced a second asset that Priyadarshi did not set out to build. Six to seven AI models trained on real clinical data, a dermatology screening tool, a dental screening app deployed through Swasthya Mitras who do not even know they are running an AI product, and a fully integrated care platform that simply did not exist before CureBay had to build one for itself. Inbound conversations are now happening around licensing that platform to other physical networks.

You should see some of my Swasthya Mitras. They don’t know it’s an AI product. They think they’re just taking images and uploading. It’s AI giving their neighbor a report instantly.

Asked what stops a well-funded competitor from copying the model, Priyadarshi says he would welcome them and share every lesson he has learned. There are 100 CureBays required to solve this for a billion people. The model he has built in Odisha, Chhattisgarh, and Jharkhand is coming next to Bihar, Madhya Pradesh, Telangana, Andhra Pradesh, and eastern Uttar Pradesh. He will have scratched only the surface. The story of CureBay is, in that sense, not about one company. It is about whether India finally figures out how to move the doctor to the patient rather than the other way around. Priyadarshi’s bet is that a nurse, a screen, and a ₹499 membership card might be enough to start.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Ten years later, 3one4 Capital has backed 5 of India’s 125 unicorns, seen its first fund return nearly 6x, and is deploying a fourth fund of $225 million. Pranav Pai, the firm’s founding partner, unpacks on what Indian venture capital gets wrong, and what comes next.

Check out the video of the conversation here or read on for insights.

The Bet Nobody Else Was Making

Pranav’s founding thesis in 2016 rested on three minority views. First, that UPI and Jio would enable technology to solve population-scale problems across India, not just for affluent urban users. Second, that Indian VC was drowning in “outside-in” investing, funding clones of American ideas: 160 Flipkart copies, 190 Swiggy/Zomato copies, 48 Ola copies. Third, that a fund run by operators, not finance professionals, would make better judgments about what India actually needed.

Our best investments are all two guys on a laptop when they come to us. That’s how early we go, and we hope we can stay invested all the way to IPO.

The firm has stayed deliberately sub-$250 million across all funds. The math is deliberate: of 200,000 startups registered in India over 15 years, only 56 have IPO’d. The number of VC firms with more than one winner from that pool is vanishingly small. A $500M fund needs $2B back. That requires 5-6 portfolio companies to IPO between $4B and $7B. The base rates, Pranav argues, do not support that at scale yet. So 3one4 targets $2B-$5B IPOs, the mid-to-large compounders, not the moonshots.

Two Unicorns From One Fund - Neither Looked Safe

Fund I produced Licious and Darwinbox. In 2016, buying meat in India meant a black plastic bag, no traceability, no cold chain. Vivek, one of Licious’s co-founders, had worked at another VC fund before starting the company. He knew every investor in the ecosystem. He still got rejected everywhere. 3one4 had done grassroots research on India’s grocery baskets. They understood the market and said yes.

What followed was not just a good investment thesis - it was a case study in founder execution. Licious ran a weekly lottery letting customers print their face on chicken boxes. Stigma became identity. Early-neighbourhood retention hit 93-94%, almost unheard of for a weekly perishable basket. India’s meat market has since grown from $20 billion in 2016 to over $55 billion today. Licious, now a unicorn, posted INR 795 crore in revenue in FY25, up 16%, with losses down 45%.

Imagine going from black-packet stigma to customers fighting a lottery to put their face on your box. That is a 180-degree inversion. That is how founders defy what the market thinks is possible.

Darwinbox was the enterprise version of the same contrarian logic. The pitch: displace Oracle, SAP and Workday for Asian enterprises. Pranav and co-founder Anurag had sold HR software to HP, EMC and VMware as operators. They had seen from the inside how broken the incumbents were: six months to go live, six months to decommission, no mobile experience. Darwinbox is now a unicorn, and in 2025 secured a fresh $40 million from Canada’s OTPP to fuel AI development and North American expansion.

Where Pranav Comes From - and Why It Matters

Pranav’s family traces its roots to Goa, through generations shaped by the Portuguese era. His grandfather was an orphan. His father was the first to earn steady money. The lesson absorbed was specific: nothing is guaranteed, and the cost of losing is not abstract. He grew up in Bangalore competing with what he calls the “shit generation” of 2000-2010, a cohort who faced reservation policy upheaval, the global financial crisis and 26/11. Most of the people he competed hardest against left India and never came back.

To compete with 25 million kids every year, through every exam and contest - everything is a hunger game. That is not a safe environment. And the strongest thing I’ve seen in a founder is the ability to harness that for the right outcome.

He stayed. He got into Stanford for electrical engineering (2011-2013), arrived just as the Facebook IPO happened, Peter Thiel was running his Zero to One course, and a16z was becoming a household name. He joined a Vertec SaaS startup in L&D as employee number three, saw it acquired at roughly 10x revenue, exercised his ESOPs and came back to Bangalore in early 2016. He decided not to become a founder (”a founder does anything but build”) and instead looked at the capital itself as the problem worth solving.

What the Next 10 Years Actually Look Like

The macro context matters here. In early 2026, what the industry is calling the “SaaSpocalypse” is playing out: agentic AI is compressing per-seat SaaS pricing, Indian IT stocks shed INR 2 lakh crore in weeks. Pranav’s thesis for the next decade is built around three pillars. First, what he frames as “Service as Software” - companies positioned as the destination, not just the tool. Not software that helps lawyers work faster, but a company that becomes a better law firm. 3one4 has backed Nyayanidhi, a Bengaluru startup building India’s litigation operating system, seeded at $2 million.

Second, manufacturing and deep tech. 3one4 has invested in Agnit Semiconductors, focused on gallium nitride technology, and is actively pursuing what Pranav calls the “fabless” model applied beyond chips - specialty chemicals, biomanufacturing, pharma - where Indian startups hold the IP and contract manufacturing to globally compliant plants. Third, technological sovereignty. Every country is now scrambling to reduce dependency on others; India is no exception, and the government is slowly getting out of the way.

If you work backwards from the question of what India’s technological sovereignty depends on, there are a lot of VC-fundable opportunities there. Some are just getting started. It is going to be a big theme for the next 10 years.

The Only Score That Matters

Near the end of the conversation, Pranav cuts through two hours of nuanced thinking with a single standard:

To judge me, you look at: has he made money? Has he scored the goals? Has he won the games? And we have.

The 3one4 Summit 2026 marks the firm’s 10-year milestone. For a fund built by an operator with no finance pedigree, betting on India before it was fashionable, that scoreboard is hard to argue with.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

Simran Mulchandani stood at the front of that classroom in 2019, a man who had once traded European currencies before the Euro existed, and felt something shift. “I’m thinking to myself, damn, these kids are screwed by a birth lottery,” he told me on the Founder Thesis podcast. “This was literally the result of a birth lottery.”

That evening, he sat down with his colleague Karishma and made a list of everything they hated about the world. They had no idea what they would build. But they knew they had to build something.

Watch the full conversation with Simran Mulchandani here.

From Trading Floors to Nightclub Floors

To understand how Simran arrived at that classroom, you need to understand the two lives he lived before it.

After graduating from Brown University with a computer science degree, he joined J.P. Morgan in New York during the pre-Y2K, pre-Euro era when the European trading desk handled ten separate currencies. His boss sent him to Singapore right before the Asian currency crisis hit in 1997. “I’m sitting in New York watching volatility,” Simran recalled. “And that’s what you want as a trader, right? Volatility going through the roof in Singapore.” He was hooked. Three days after arriving, the global heads offered him a permanent role.

But something nagged at him. “Every year is a race,” he said. “It’s a race to bonuses.”

Then a childhood friend named Ashutosh Fatak showed up with an idea about changing Indian music. Independent artists and session musicians deserved a platform beyond Bollywood. Simran moved back to India with his wife and two children. They started Blue Frog.

Blue Frog became an icon. The entrance was designed like something from Charlie and the Chocolate Factory, a tiny door in a mill compound that opened into a cavernous, magical space. Forbes did a deep-dive on their fifth anniversary. Time Out Mumbai wrote that the city’s music scene now had two eras: before Blue Frog and after Blue Frog. The Independent UK named it one of the top ten music venues in the world, alongside the Sydney Opera House and Madison Square Garden.

We built an incredible brand, a brand that even 10 years after closing still has its cachet. From the outside, it must have looked amazing, but I can tell you it was really hard work.

Then came Delhi. And Pune. And Bangalore. They expanded too fast without understanding their new markets. “We went into Delhi to build for ourselves, unfortunately,” Simran admitted. The financial strain weighed down the entire company. In 2015, after ten years, he exited with his laptop, a pair of forgotten sneakers, and one person: Karishma, Blue Frog’s first employee, who would become his co-founder at Rangeet.

The Mountain and the Classroom

The next chapter began at 15,000 feet.

Simran and his family chose Mount Kinabalu in Borneo for his recovery. For most climbers, it’s manageable. For Simran, carrying ninety kilograms and the residue of running a nightclub for a decade, it was brutal. His daughter developed altitude sickness. His wife told him to keep going with their son Ishan. At 14,000 feet, under moonlight, Simran was stopping every twenty steps. He was ready to quit. Ishan, fifteen years old, refused to let him.

He said to me, ‘Come on, Dad, we’ve come this far. I’m not letting you give up.’ He was 15. He’s a cricket player. I support him. But he supported me that day, and I wouldn’t have made it to the top.

At the summit, Simran realized the peak wasn’t his destination. It was his launchpad.

After that classroom in Malad, he and Karishma began building workshops on confidence, creativity, communication. Three-hour sessions on gender, climate, mental health, societal inequity. Then they went to Bangladesh to meet BRAC, one of the largest NGOs in the world. BRAC took them to schools in rice fields and elite international schools in Dhaka. Then they delivered a verdict that changed everything.

“They told us one very simple thing that was blindingly obvious,” Simran recalled. “They said, ‘How important is this Rangeet thing to you? Because you will never scale as a three-hour workshop. Your schools don’t have time.’”

It was the Blue Frog lesson again. They had built for themselves, not for their customer. So they rebuilt. The SEEK curriculum, Social, Emotional, and Ecological Knowledge, emerged from that failure. And they made one decision that set them apart from every EdTech company of the previous decade: the app would be for teachers only. Children would never touch a screen.

No machine can ever teach you how to be a human being. So right from day one, Rangeet was built for teachers.

The Model That Shouldn’t Scale

Here is how a Rangeet classroom works.

The teacher has the app. It contains instructions, training videos, tutorials for every lesson. When the teacher needs to show something to students, they press a button. The content appears on the classroom’s smart TV. No WhatsApp notifications. No phone calls. Just the lesson. If there’s no smart TV, Rangeet prints resource books with only the child-facing content.

The unit economics are startling. Rangeet costs approximately ₹500 per child per year, just 1% of what India spends per child in the public school system. Training a teacher means reaching forty children. Training a master trainer means reaching hundreds of teachers. The entire model is built on the understanding that teachers are the multiplier, not the bottleneck.

The results speak in partnerships and numbers. Oxford University Press approached them to write “My Happiness and Me,” India’s first well-being workbook series for grades 1-8. The Brihanmumbai Municipal Corporation adopted Rangeet across 250 schools. Pratham, one of India’s most respected education NGOs, now executes their programs. The UN recognized them as a best practice for SDGs 4 and 5. The Brookings Institution featured them in a 2022 policy report. In January 2025, Simran spoke at the World Economic Forum in Davos.

Rangeet has now reached over 500,000 children across India and Bangladesh on approximately $500,000 in annual revenue. In 2026, they’re expanding into four African countries.

The market tailwinds are real. India’s Social and Emotional Learning market is projected to reach $1.8 billion by 2033, growing at 17.5% annually. NEP 2020 mandated holistic student development. The post-COVID learning gap, where children returned to school having lost years of social-emotional development, created urgent demand for exactly what Rangeet offers.

If someone tells us, ‘We need some time to get back to you,’ I’m like, ‘Good job. Let’s do this slow.’ Because we’re seeing all of these events taking place because we are investing time.

The Long Game

Most EdTech companies of the 2015-2021 era believed scale came from removing humans from the equation. Gamified apps. Personalized AI tutors. One device per child. Rangeet bet the opposite. Scale comes from empowering the teacher.

Simran’s advisor Sean Bellamy, founder of the Sands School in the UK, put it starkly: “Is maths useful if there are no animals left to count? Is language useful if there are no forests left to describe?”

Anir Choudhary, the friend who gave them their first break in Bangladesh, has now joined Rangeet to lead growth. He built a community of 600,000 teachers in Bangladesh starting with 23. His challenge for the team: reach 100 million children.

They’re building AI features now, but with a clear constraint. AI will support teachers with better feedback and coaching. It will never replace them.

Life is so long. If we can just be committed to the things we love, be open, and not worry about time, everything works out.

For a man who traded currencies through the Asian financial crisis, who built and lost an iconic music venue, who nearly gave up on a mountain in Borneo, patience has become the strategy. The children losing birth lotteries every day now have 500,000 peers learning what no app can teach on its own: how to be human.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

This is not a future projection. It is already happening, at scale, inside the credit pipelines of India’s largest private banks.

Check out the video of the conversation here or read on for insights.

The man who built this system is Prateep Basu, Co-founder and CEO of SatSure, a Bengaluru-based Earth intelligence company at the intersection of satellite imagery, artificial intelligence, and financial infrastructure. SatSure has raised approximately $27.7 million across 12 rounds, counts ICICI Bank, HDFC Ltd, and Kotak Mahindra Bank among its investors, and is currently valued at approximately Rs. 512 crore. Its flagship platform, SatSure Sage, has analyzed over 2.1 million farmer plots across 1.95 lakh villages in India.

The numbers matter. But the origin story matters more.

The Problem Nobody Had Framed Correctly

Before SatSure, getting a loan as a farmer required a physical visit to the land, third-party ownership verification, a manual crop assessment, and a credit officer with enough agricultural knowledge to decide. The process took 30 days. For urban borrowers, banks were competing to offer pre-approved loans by WhatsApp.

People like you and me get WhatsApp messages ten times a day from banks about loan eligibility. Why can’t that exist for farmers? That was the gap we were filling.

Prateep came to this problem from an unusual direction. He spent his early career as a propulsion systems engineer at ISRO, working on the GSLV MK-III. After a Master’s at the International Space University in Strasbourg, he joined Northern Sky Research, a Boston-based consultancy advising global aerospace investors. There, he identified what he calls the industry’s core problem.

The industry was supply-centric, not demand-centric. People were launching satellites without knowing precisely what problems they were solving, because whatever happened, defense would buy something out of it.

He returned to India in 2017 and co-founded SatSure with Abhishek Raju and Rashmit Singh Sukhmani. Their first pilot was in Srikakulam, Andhra Pradesh, made possible by a letter of introduction from local MP Ram Mohan Naidu to the district collector. That letter came without money. It came with access, to farmers, agricultural scientists, and the texture of how rural credit actually worked. Naidu is today India’s Union Minister of Civil Aviation.

By late 2017, SatSure had won the AP AgTech Summit challenge, backed by the Bill and Melinda Gates Foundation, and secured a Rs. 5 crore purchase order from the Andhra Pradesh government. Prateep and his co-founders then bootstrapped for four and a half years before raising their first institutional round.

What 40 Pixels Know That a Bank Manager Doesn’t

The technical question at the center of SatSure’s model is simple on its surface: can a satellite identify a specific crop on a one-acre plot?

The answer involves physics that goes well beyond what any human eye can detect. Each pixel in a multispectral satellite image carries 13 bands of spectral information, including near-infrared and short-wave infrared wavelengths that measure chlorophyll content, soil moisture, and crop stress. The same 40 pixels are re-imaged every five days, building a time series that reconstructs multiple growing seasons from a single screen.

Risk is all about history. You cannot send a drone back in time to collect data. That is the unique value of satellite imagery.

This time-machine effect is what makes the data genuinely useful for credit assessment. Prateep’s team combines crop identification with market pricing data, minimum support prices, and cost-of-cultivation benchmarks per crop per hectare to construct a season-by-season income estimate for each farmer. That feeds into an alternate credit score.

The result: a loan approval process cut from 30 days to 10 to 30 minutes, with auto-approval thresholds set by the banks themselves.

Getting to 90 to 95 percent crop identification accuracy in India required far more work than in markets like the US, where farms are large and uniform. Indian agriculture is fragmented, with adjacent plots often growing different crops. Prateep built SatSure’s training dataset through statistical field sampling and data-barter arrangements with early banking partners, accumulating approximately 300,000 ground-truth data points across major lending states.

With 40 pixels and 13 spectral bands per timestamp, re-imaged every five days, you end up with incredibly dense information. That is the secret sauce.

One Stack, Three Markets, and a Bet on Sovereignty

SatSure’s product architecture reflects a deliberate decision to maximize return on a single data infrastructure. The same satellite stack powering farm credit scoring also drives SatSure Skies, which helps the Airport Authority of India automate airspace cartography that previously took months, and SatSure Sparta, which helps FMCG companies and global banks prove their commodity supply chains are not sourced from recently deforested land, a compliance requirement under the EU Deforestation Regulation.

The product is everything that happens behind the client interface. The data models, the reusable infrastructure, the algorithms. The thin client layer for the end user? That is the consulting part.

The company’s most consequential move, however, is upstream. In 2022, Prateep launched KaleidEO, a subsidiary developing proprietary satellite payloads planned for Very Low Earth Orbit at approximately 380 km altitude. In 2025, SatSure joined the Allied Orbits consortium alongside Pixxel, Dhruva Space, and PierSight, winning an IN-SPACe contract to build India’s first privately led national Earth observation constellation, a 12-satellite system representing a Rs. 1,200 crore investment. The consortium declined up to Rs. 350 crore in government viability gap funding, choosing to self-fund in exchange for global commercialization rights.

The logic is direct. India currently spends approximately Rs. 800 crore annually buying satellite imagery from foreign companies. Prateep’s argument is that controlling the camera in orbit is the only way to control the economics and the continuity of the business underneath it.

In February 2026, Prateep was named one of the Top 10 Global Innovators at the India AI Impact Summit for SatSure’s work in satellite-based credit scoring, a recognition that places a quiet, technically grounded company at the center of a market the global satellite EO industry values at $13.2 billion today, growing to $19.6 billion by 2030.

Forty pixels. Five hundred kilometers up. Every five days. It turns out that is enough.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

That gap is what Nalin Agrawal has spent eight years building toward. His company, Snapmint, now serves 7 million monthly active users across 23,000 pincodes, financing 1.5 million purchases every month. In October 2025, General Atlantic led a $125 million Series B. FY25 revenue grew 80% year-on-year to Rs 158.5 crore. The company posted its first net profit: Rs 15 crore. None of it was obvious when the idea began, and the whole thing nearly died before it found its footing.

Check out the video of the conversation here or read on for insights

The Anomaly That Started Everything

In 2016, Nalin and his two IIT Bombay co-founders, Anil Gelra and Abhineet Sawa, were running a programmatic advertising business. The three had previously built a patent analytics firm (acquired by Netscribes) and payment systems for Yes Bank. Data and transactions were their native language.

An e-commerce client asked them to run a campaign for a new credit card EMI feature at checkout. Other campaigns delivered 8-9% lift in consumers reaching the payment page. The EMI campaign, which used the word prominently in the creative, delivered a lift of 80 to 120%. Consumers were clicking through in large numbers. Almost none transacted.

The geo analytics explained why: respondents were overwhelmingly from tier-2 cities, and they did not have credit cards. Demand existed, at real scale. The product to serve it did not. That insight led directly to what Snapmint is today: an EMI payments platform requiring nothing more than a PAN number and a UPI mandate.

How Interest-Free Credit Works, and Why It’s Hard to Copy

The first question Snapmint always has to answer is the obvious one. If consumers pay 0% interest, 0% processing fees, and 0% late fees, who absorbs the cost?

The merchant does. When a consumer buys a Rs 20,000 television on 6-month EMI through Snapmint, the brand pays a commission of 3.5% to 10%, depending on category and tenure. For high-ticket items like furniture or premium electronics, conversion lifts of 20-40% make that commission a straightforward trade. Nalin often invokes Bajaj Finance here: it has used the same offline subsidy model for decades. Snapmint rebuilt it for the internet, with machine learning replacing the human agent in a store.

In one model, you rely on the consumer making mistakes, and that is what you earn from. We are working in exactly the opposite way. We don’t make money from their mistakes.

The margins are thin. A typical 6-month transaction at 7.5% commission leaves roughly 2.5% for credit losses, 3% for cost of capital, and operational costs on top. Net margin per transaction sits around 1.5%. What makes the business defensible is not the model itself, which others could replicate, but the precision of the underwriting behind it.

Snapmint’s credit loss rate is under 2.5%. The industry benchmark is 6-8%. The model evaluates each transaction against approximately 3,000 variables, with 180-200 data points available in real time, and delivers a decision in under one second. It has been trained on over 9 million loans, 8 million of which are now fully closed. The fraud engine runs parallel to underwriting, watching for pattern anomalies, three identical smartphones shipping to the same pincode inside 15 minutes, and responding in 5-10 minutes before losses compound. This is not the kind of thing a competitor can purchase or hire quickly. Nalin estimates it takes two to three years to build a comparable system, and that Snapmint keeps improving while anyone else is catching up.

For anyone to compete, they’d first have to beat our threshold of 70 to 80% approval rates while keeping credit losses below 2.5%. Marrying those two is the key to this business.

The Term Sheet That Disappeared, and the Investor Who Stayed

After getting its NBFC license in 2019, Snapmint had a Series A term sheet for $5 million signed on March 19, 2020. India locked down on March 24. The money evaporated. For a company that had just committed to building a lending book, the timing was near-fatal.

One investor held. Prashasta Seth, later Managing Partner of Prudent Investment Managers, had been watching the model closely. He came in with $1 million not because the market was good, but because he believed the model would survive.

I love your business model. If you can come out of this period and show me it works, I’ll be happy to double down.

Prashasta Seth to Nalin, March 2020

Snapmint came out of it. Credit losses on the COVID cohort settled at 4.8%, well below any comparable benchmark. Seth returned in 2021 with $10-12 million. Revenue at the time of that investment was approximately Rs 4 crore. By FY25 it was Rs 158.5 crore. Seth’s firm led the $18 million pre-Series B in December 2024, then participated again in the $125 million General Atlantic-led round in October 2025.

The Next Bet: EMI Without a Card, on Any QR Code

The company is now building credit on UPI: a consumer scans any merchant QR code, including a PhonePe code, and sees EMI as a payment option through a pre-approved credit line held with a partner bank. RBI restricts credit lines to banks rather than NBFCs, so Nalin’s structure is a fee-based partnership: the bank holds the exposure, Snapmint manages the underwriting and collection.

We believe India will leapfrog credit cards and go straight to EMI on UPI. We’ve enabled brands to increase sales by 10 to 20%. We want to bring this to more than 100 million consumers in the next few years.

The argument is structural. Credit card penetration in India serves the top 2-3% of consumers. UPI reaches almost everyone. With 10-15 million existing Snapmint users as a starting base and the General Atlantic capital behind it, Nalin’s three-year target is Rs 1,000 crore in revenue. There is no unicorn valuation to report yet; General Atlantic holds 18.8% after the Series B, and the company has not disclosed a number. The milestone Nalin points to instead is the Rs 15 crore net profit in FY25. The model works. Now it needs to scale.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

For most founders, this would signal product-market fit. For Sreeraman Mohan Girija and his co-founders at ShopSense, it was a lesson about customer delight versus sustainable business.

We thought, from both a customer and brand perspective, is there something interesting with technology that we can do inside the store that will increase engagement time.

Twelve years later, his company Fynd powers 2,300+ brands across 20,000+ stores, does ₹300-350 crore in annual revenue, and employs 1,100 people. Yet most consumers have never heard of it.

Check out the video of the conversation here or read on for insights.

The Hardware Trap

The ShopSense Match product looked perfect. Fashion retailers paid ₹10,000 monthly per device. Customers loved it. But the economics were broken.

Each touchscreen cost ₹50,000 upfront. Stores needed rewiring. Pre-Jio internet was unreliable. The fatal flaw was attribution.

While we had data showing a 10-12% uplift, it was difficult to convince a brand because the transaction did not happen on the device. Transactions are happening on the point of sale system.

By 2015, ShopSense was doing roughly ₹1 crore in revenue after raising $1 million from Kae Capital and IIFL.

They pivoted to real-time inventory data they already had. What if Tier 2 cities could shop from Mumbai stores? They built a hyperlocal marketplace with 2-3 hour delivery and “Find a Fit” where customers ordered two sizes, tried both, and returned one.

In fashion, the biggest problem is fit.

By 2016-17, after raising $10 million from Google, the marketplace hit ₹5 crore in revenue. Then came the wall: customer acquisition costs, logistics, discounting wars.

Everyone loves a discount. That became the default expectation. That requires a lot more capital.

Meanwhile, the technology Fynd built (storefront builders, inventory systems, marketplace connectors) was valuable to brands wanting omnichannel. Reliance Brands Limited had started using Fynd’s tech.

When Reliance Industries invested in 2019, they weren’t buying a marketplace doing ₹5 crore. They were buying infrastructure powering their operations.

Reliance acquired 87% for ₹295 crore. Investors made 8x returns. The consumer marketplace shut down. Fynd became pure B2B infrastructure.

Scale as Strategy

The first year was rough.

You’re talking about one of the largest employers in the country. And here we are like a startup, move fast and break things. It took us a good year to culturally adapt both ways.

But once aligned, Fynd got scale impossible for an independent startup.

When we started working with Reliance, we understood there is this edge case, this use case, this format. How do you build a platform that serves all those purposes?

The product explosion was dramatic. Fynd built D2C websites, point-of-sale systems, endless aisle tablets, AR try-on, order and warehouse management, marketplace connectors (ranked #1 by Amazon), AI support (Kaily), AI photography (Pixelbin), and AI fashion design (Fynd Create).

Revenue grew from ₹5 crore to ₹300-350 crore. Small brands pay ₹10,000-20,000 monthly with transaction fees. Enterprise clients pay lakhs.

As they scale, we charge something on top of transactions, which they’re happy to spend.

Today, 80% revenue is enterprise. Out of 1,100 employees, 600-700 are engineers and designers.

The Designer’s Edge

Sreeraman has a Master’s in Interaction Design from IIT Bombay. This shapes everything.

When you are building a product, you only optimize for the end customer. Not your engineering team’s capabilities.

He leads a 100-person design organization. His hiring test is empathy. If candidates defer to managers or can’t articulate customer-first reasoning, it’s a red flag.

Sreeraman has strong views. CRED is “fantastic design but value has gone down.” Amazon is ugly but fast, making it great. Figma beat Sketch through distribution, not design.

Design is not beauty. Design is what works.

This shows in Fynd’s conversational commerce approach. Sreeraman believes shopping will shift to chat but is skeptical of simplistic visions.

Shopping behavior is going to change. As a brand, are you ready to provide that experience on your website?

Need-based shopping (groceries) works in chat. Want-based shopping (fashion) requires visual browsing. Fynd’s AI stack plugs into whatever interface wins.

Going Global & The 2030 Vision

By 2022, Fynd had saturated India. They chose Middle East first (MAF, Rivoli, FNAC), then Southeast Asia (Superindo). In 2024, Sreeraman walked over 100 UK stores.

When we started conversations with retailers, lot of people in tech had some Indian connection. They’d worked with Reliance.

International revenue is under 10%. Canada is in pipeline with shoppable TV.

Sreeraman is thinking about “AI-native unified commerce” by 2030. India’s e-commerce should hit $345 billion. AI in retail will grow 14x. ONDC threatens to drop marketplace commissions from 25-35% to 5-8%.

Fynd’s bet: brands need zero-touch operations, conversational discovery, open protocols, composable infrastructure. The platform processes transactions for 20 million+ customers at Reliance scale.

Why Infrastructure Wins

The irony of Fynd’s story is deliberate obscurity. While Flipkart and Amazon battle for mindshare, Fynd quietly powers thousands of brand experiences behind the scenes.

These ideas are really fragile when they start off. In a meeting, it’s very easy to rip apart an idea.

The same applies to infrastructure businesses. In a world obsessed with consumer metrics, choosing pipes over destinations requires conviction. Twelve years in, that conviction produced a ₹300 crore business, a strategic acquisition, and a platform touching millions of transactions daily.

Not bad for a designer who started with touchscreens in fashion stores.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

It was 1999, and Sreevathsa Prabhakar, a fresh engineering graduate working as a service technician for BPL, stood trapped in a 100-square-foot house in Mumbai’s Dharavi. The man who’d bought an expensive television set was furious it wouldn’t turn on. He wanted his money back, and he wanted the owner of BPL to come personally. Sreevathsa was just a junior engineer who barely spoke Hindi.

He was almost about to hit me.

After two hours of pleading, the customer finally let him examine the TV. The problem? A loose cable connection. When the screen flickered to life, everything changed.

He went from wanting to hit me to offering me a drink. He even offered me money. That day, I realized something powerful: if you solve a customer’s problem, they will go out of their way to reward you.

Check out the video of the conversation here or read on for insights.

Sreevathsa didn’t set out to be an entrepreneur. While his engineering classmates landed coveted software jobs at Infosys and TCS, he spent the late 1990s fixing televisions and refrigerators in customers’ homes across Mumbai. His territory included Dharavi, Sion, and Matunga. Not exactly the glamorous career his family had envisioned.

But those years taught him what MBA programs don’t: the economics of after-sales service, the psychology of frustrated customers, and the operational nuances that separate good service from fraud. He learned that service was always treated as a cost center, an afterthought in product companies focused on distribution and sales.

In 2009, after his father’s death and a callous comment from his Nokia supervisor, Sreevathsa quit corporate life and founded The Service Solutions (TSS). The concept was simple but novel for India: bring McDonald’s-style standardization to device repair. Apple became his first major client, trusting this unknown Indian entrepreneur to launch the country’s first Apple Care Center.

Apple is Apple because they really believed in the dream that I sold, saying, I will bring the Apple Store experience to India.

TSS scaled to 385 locations across India and the Middle East, managing service for Apple, Samsung, OnePlus, and HP. Revenue hit ₹160 crores with strong margins, all bootstrapped. But the journey nearly broke him.

When German conglomerate B2X offered to acquire TSS in 2014, the acquisition should have been the happy ending. Instead, it became the origin story for something bigger.

Device Protection Reimagined

Inside B2X’s German bureaucracy, Sreevathsa pitched an idea: what if device protection could be reimagined as a white-label technology platform instead of traditional insurance? The corporate answer was no. So he bought back his shares and started Servify in 2015.

The problem was obvious to anyone who’d bought electronics in India. Extended warranties existed, sold by insurance companies through retailers, but customer experience was terrible. Claim rejection rates hovered around 70%. Repairs took weeks.

Servify’s model flipped the equation. Instead of selling “Servify Care,” the company would power “AppleCare,” “Samsung Care Plus,” and “HP Complete Care” as the invisible orchestration layer. The customer sees only the brand they trust. Behind the scenes, Servify handles everything: distribution (onboarding 150,000 retail outlets), underwriting (partnering with insurers like ICICI Lombard), decisioning (AI-powered claim approval in real-time), and fulfillment (routing repairs to OEM service centers).

The economics work because everyone wins. On a ₹10,000 device protection plan, retailers keep 40-45%, OEMs get 10-15% royalty, insurance partners receive 25-28%, and Servify retains 12-15%. At scale, it’s a 15-18% EBITDA business.

How? By running almost entirely on software.

Our entire ops team, which manages all these claims and dealer onboarding, is about 15-18 people on a ₹2,000 crore business. Because we automated almost everything.

The platform now manages 740 million devices across 120+ product categories, processes 3 million transactions monthly, and serves 30 million users in multiple countries. The shift to global markets has been dramatic. International business now generates 70% of revenue, up from just 25% two years ago. The U.S. has become Servify’s largest market, followed by Europe and the Middle East.

The technology advantage is real. Servify holds over 20 patents, including video-based diagnostics that can assess device damage remotely. The platform integrates with OEM serial number databases to prevent fraud (a customer can’t claim warranty on a broken device bought secondhand), optimizes parts inventory across service networks, and uses confidence scores to monitor service center performance.

The Path to Public Markets

The company expects to raise $250-300 million at a valuation between $1.5-2.3 billion. Ahead of the IPO, Servify is finalizing a $100 million pre-IPO round that will officially crown it a unicorn.

The company has raised $165 million to date from investors including Iron Pillar, Blume Ventures, and DMI Sparkle Fund. Much of that capital went toward geographic expansion, regulatory licensing (device protection is a regulated business in most markets), and building integration infrastructure. A single telco partnership like AT&T requires integrating 7-8 different systems. Every new country requires fresh licenses and local compliance frameworks.

The financials show the inflection point arriving. FY24 revenue reached ₹759 crores, up 24% year-over-year. Net losses dropped 59% to ₹94 crores as operating leverage kicked in. FY25 is expected to be Servify’s first profitable year, with net revenue hitting ₹900 crores and 2-3% margins.

But Sreevathsa is explicit about playing a longer game. AT&T’s device protection program just launched and will take three years to ramp fully. Samsung’s consumer electronics division just went live. HP is expanding across countries. Apple, a client since 2008, continues adding markets where Servify powers AppleCare in regions without direct Apple retail presence.

This ₹900 crore net revenue will be at least ₹4,000 crores in the next few years. What will that valuation be? That’s what matters. In our business, I can’t value based on what is today.

The discipline extends beyond financial projections. Despite raising over $165 million, Servify operates with bootstrap-era frugality. Sreevathsa still recalls an incident where an employee exploited approval limits to place a $2 million parts order. His response? Make her personally negotiate with the supplier for a free return or pay out of pocket.

The DNA is like no wastage. I am okay to spend on celebrations or people, salaries. But where there is no need, there is no need.

The example: if you solve 9 customer tickets in under 2 minutes but one takes 5 hours, your average looks fine. But 10% of customers received failed service. Averages lie. Every customer interaction matters.

When Sreevathsa visits Apple Park in Cupertino and sees a “Welcome Servify” banner, he doesn’t see validation of past success. He sees proof that solving customer problems, even unglamorous ones in after-sales service, can build businesses that power the world’s most valuable brands.

The IPO will bring scrutiny, public market discipline, and governance requirements. Sreevathsa welcomes it.

An IPO isn’t a funding event. It’s a maturity event.

For a company born from a locked room in Dharavi, that kind of maturity has been 25 years in the making.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

The physics PhD who had spent years studying “squishy molecules” at the University of Pennsylvania had just discovered India’s innovation problem wasn’t technical. It was cultural. The country had world-class scientists, state-of-the-art labs, and brilliant minds. What it lacked was anyone who cared about turning research into businesses.

Fast forward to January 2026. Ritu is managing ₹1,200 crore ($150 million) across three funds at Ankur Capital, the deep science and agritech-focused venture firm she co-founded in 2013 with Rema Subramanian, a cost accountant with 30 years of CXO experience. Her portfolio company Captain Fresh just withdrew its ₹1,700 crore IPO application, not because of weak fundamentals, but to complete an acquisition of Frime, a Spanish seafood company. When it refiles in 2026, it will mark Ankur’s first public exit at a valuation exceeding $1 billion.

The numbers tell the Captain Fresh story: ₹3,421 crore in revenue for FY25, up 145% year-over-year, and profitable with ₹42.4 crore in PAT. It’s a rare feat for new-age tech companies.

That research institute that kicked Ritu out? It’s now among the domestic entities filing patents at record rates. India logged 100,000 patent applications in 2024, with 50% filed domestically for the first time in history.

Check out the video of the conversation here or read on for insights.

From Lab Bench to ₹1,200 Crore

Ritu’s path to venture capital reads like an accident report. After her PhD, she spent a decade at Unilever and Philips bringing products to market globally, from soaps to televisions. But working inside mega-corporations taught her a frustrating lesson.

The risk appetite for innovation in such a mega company is limited. And I really should be in this more risky startup world.

When she returned to India in 2013, the venture ecosystem looked nothing like today. The word “family office” didn’t exist as a formal concept. Angel networks were nascent. And the idea of giving someone else your money to manage for seven to ten years?

I have been in people’s offices where you’d call it a family office today. And in those days, you just talked to the family or the proprietor. There was no family office person. It was just called, I’m a rich man, basically.

Fund I closed at ₹50 crore in 2016, scraped together from angels, friends, and family. The first check was $100,000. Ritu met a stranger in Colorado who wired $150,000 on the spot. That’s what it took.

By 2020, Fund II closed at ₹350 crore with institutional backing from CDC Group (now British International Investment), SIDBI, and the MacArthur Foundation. Fund III, currently raising with a ₹1,200 crore target, adds the U.S. Development Finance Corporation to the LP roster.

The Multiplication Insight

Most venture capitalists obsess over valuation. Ritu obsesses over multiplication.

Can I tell you my biggest learning? It’s multiplication. How much money we make in a company is multiplied not just by the valuation, but how many shares we hold, too. It sounds very silly, but it’s a reality.

This simple math drives Ankur’s entire strategy: Write the first institutional check. Take double-digit ownership. Follow aggressively through Series B to defend that stake. Target 4-6x returns with a net IRR of 30%+ over eight to ten years.

The insight shapes everything. Ankur doesn’t spray capital across 50 companies at low ownership. It concentrates on 15-20 bets where it can meaningfully influence outcomes and retain significant equity through multiple rounds.

Two Buckets, Two Playbooks

Ankur operates with two distinct investment theses, each requiring different underwriting skills.

Digital India: B2B platforms digitizing fragmented mass markets using India Stack infrastructure. The risk isn’t technical. It’s operational execution.

Captain Fresh and Vegrow, a B2B fruit marketplace that raised $40 million in January 2025, fall here. Vegrow posted $150 million in revenue for FY24, targets $500 million by FY26, and handles 300 tonnes of fruit daily across 30,000 connected farmers.

Saswat Finance, a rural fintech startup, uses milk yield data to provide loans to dairy farmers. It has disbursed ₹55 crore across 6,000+ customers and partners with Amul and Hatsun.

The thesis hinges on what Ritu calls “codification.” In unstructured markets, there’s no standard SKU for fish, fruit, or fabricated parts. Technology creates that standard language through computer vision and sensors. Once quality is codified, trust can be digitized.

Deep Science: Global IP plays in biology, chemistry, and physics. The science is validated pre-investment. The question is techno-commercial scale.

Offgrid Energy Labs raised $15 million in September 2025 for its ZincGel battery technology, a safer and cheaper alternative to lithium-ion. It’s establishing a 10 MWh demo plant in the UK, planning a GW-scale facility in India, and holds 25+ IP families with Archean Chemicals as a strategic investor.

String Bio converts methane into protein via its SIMP platform. It posted ₹16.5 crore in revenue for FY24, operates a commercial plant in Tumkur, and has partnerships with Woodside Energy and Skretting.

Vimano, Ankur Fund III’s first investment, manufactures membranes for flow batteries and green hydrogen electrolyzers. It raised ₹25 crore in April 2025.

These companies solve global problems from an Indian context, then scale worldwide.

The Operational Reality

Ritu describes the work as building a “village” around each startup. Ankur co-founded ThinkAg, an industry platform connecting agritech startups with corporates and policymakers, and hosts the annual Deep Science Forum (third edition: January 13, 2026 in Bangalore).

Once we’re invested, we have to do everything. You call me up and say you want to connect with somebody. I don’t know the person, but I need to hustle to go find someone who does.

The work splits into four categories: strategic guidance, tactical operations, fundraising support, and founder emotional support. Midnight calls are common.

You have not invested in a nine to five business.

Her advice to founders is blunt: Think like an investor, not just an operator.

You are putting capital into projects. Your assessment of why you’re putting these into things, there has to be some discipline to that piece. If you don’t think of yourself as the investor, you lose the plot.

And avoid mediocrity at all costs.

Mediocrity attracts mediocrity.

Hire either the absolute best or hungry hustlers who learn fast. Never the safe corporate middle ground.

The Tailwinds

The macro environment is shifting in deep science’s favor. India’s deep tech sector is projected to attract over $10 billion in funding by 2029. The National Deep Tech Startup Policy announced a ₹1 lakh crore corpus for R&D financing. States like Karnataka and Tamil Nadu released dedicated deep tech policies for 2025-2030.

Geopolitically, the “China Plus One” diversification is accelerating. The US Biosecure Act is redirecting global biotech supply chains. Indian talent, democratic stability, and cost advantages are pulling in global partnerships.

The domestic capital base is maturing too. Family offices are allocating meaningful capital to venture funds, reducing reliance on volatile foreign sources. The IPO market, while selective, is open to profitable growth stories like Captain Fresh.

Ritu’s journey from studying soft matter in a Pennsylvania basement to orchestrating ₹1,200 crore in capital deployment spans just 13 years. That same institute that showed her the door? Its researchers are now pitching her their patents.

The physicist has learned to love the mess. And she’s built the formula to profit from it.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt

This is the kind of experience that mixed reality has been promising for a decade. The difference is that this one actually works, and a company called Flam AI figured out why the others didn’t.

Watch the full conversation with Amit Gaiki on the Founder Thesis podcast for the complete story of Flam AI’s pivot, their Samsung campaigns, and the distribution strategy fueling their growth.

Check out the video of the conversation here or read on for insights.

From Consumer Dream to Enterprise Reality

Amit Gaiki, co-founder and CTO of Flam AI, didn’t set out to build an advertising company. He and his co-founders, Shourya Agarwal and Malhar Patil, started in 2021 with a thesis about how humans consume content.

Content on internet has evolved since the past two decades. First it was all text content. Then we transitioned into images, Facebook, Instagram took over. Then it was all videos with TikTok and reels. So obviously it has to evolve into something better.

The “something better” they identified was interactive, three-dimensional content that exists in your physical space. Not virtual reality, which cuts you off from the world, but mixed reality, which layers digital objects onto the real environment around you.

They raised a $3.5 million seed round on this vision: build the next TikTok, but for mixed reality. They created a consumer app with avatar systems, motion capture technology, and the ability to make printed photos come alive when scanned. Users loved it. The product was sticky.

Then 2022 happened. The funding environment turned hostile to consumer moonshots without clear revenue paths. Amit and his team faced a calculation many founders avoid until it’s too late: roughly ten months of runway remaining.

Most founders in that position keep pitching investors. Amit’s team asked a more useful question: where is there a business hiding inside what we’ve built?

We realized we need to pivot this into a business-facing entity. Consumers will follow, but you have to build an ecosystem first. And the consumer is not the best place to do that.

The answer was advertising. Mixed reality advertising had existed for years but never scaled. Brands would run a flashy AR campaign, generate some press, then never do it again. It was a gimmick, not a business.

The diagnosis was straightforward. There were two ways to deliver an AR experience. Native apps offered great performance but zero accessibility, because no one downloads an app to watch an ad. Browser-based AR offered accessibility but terrible performance, working on perhaps 60% of Android devices in India with laggy, pixelated results.

Apps have very good performance but accessibility is very low. Browsers have high accessibility but performance is very low. We get the best of both worlds.

Flam AI built a native mixed reality bundle that installs in a single click without requiring an app store visit. Scan a QR code, and within one second, high-fidelity 3D content is playing through your camera. The company achieved 100% device compatibility on Android. The rendering is, in Amit’s words, “pixel perfect.”

The Business and the Numbers

With the friction problem solved, brands actually wanted what Flam AI was selling. The value proposition was clear: every touchpoint where a brand communicates with a consumer could become an interactive mixed reality experience.

The pricing model reflects how brand marketers think. Flam AI charges a fixed fee, typically around 20% of media buying spend, rather than per-impression rates that create unpredictable costs. The platform is self-serve and no-code. Brands create 3D assets using tools they already know, upload to Flam AI, and the system handles device optimization automatically.

What changed everything was repeatability.

There is a lot of repeatability, and there’s also a lot of confidence in launching flagship products with Flam AI. You want to cut out all the noise and be very exclusive, very premium.

Samsung used Flam AI for their Circle to Search launch, creating what Amit describes as the first interactive print ad in India. When Samsung launched voice search, they returned to build what the company claims is the first voice-interactive mixed reality ad anywhere. Google is rolling out Flam AI codes across all their Indian retail outlets. ICICI Bank used the platform to deliver mutual fund presentations through WhatsApp, with a fund manager appearing to stand in the user’s living room.

Flam AI has crossed $10 million in annual recurring revenue. The company raised a $14 million Series A led by RTP Global in May 2025, bringing total funding to approximately $22 million. Revenue is growing “in multiples, not in percentages,” according to Amit.

The majority comes from India, where digital advertising has become the largest ad medium at around ₹49,000 crore in FY24-25. Flam AI has also expanded to the Middle East, Japan, and the United States. The timing has helped: when Meta shut down Spark AR for third-party creators in January 2025, brands suddenly needed alternatives.

Flam AI’s distribution strategy centers on creative agencies rather than direct enterprise sales alone. Agencies can use the platform free with a watermark, build proof-of-concept campaigns, and pitch them to brand clients. When the brand wants to run without the watermark, they pay Flam AI. It’s product-led growth applied to enterprise sales.

The Glasses Question

Every conversation about mixed reality eventually arrives at the same place: what about glasses? Mark Zuckerberg has bet Meta’s future on mixed reality headsets. Apple’s Vision Pro represents a similar wager.

Amit’s perspective is more measured.

Headsets won’t replace phones for sure. Glasses might be the next interface, but only if convenience matches phones. We haven’t seen anything in market yet. You can’t predict it, it can only be experienced.

He points to a pattern anyone who owns a Meta Quest will recognize. Exciting for the first few weeks. Unused two months later. Phones are always with you, always charged, always ready. Until glasses match that, the transition won’t happen.

Flam AI’s bet is that phone cameras are the transitional interface. The company is building the ecosystem, the publishing tools, the brand relationships, on the platform that exists today. If glasses become dominant, Flam AI will already have the content pipeline in place.

The Underlying Bet

Amit and his team are wagering on something specific: that the shift from passive to interactive content is inevitable, and that the company controlling the low-friction distribution layer will capture disproportionate value. They’re not trying to own the creative tools or the media buying. They’re trying to own the moment when someone scans a code and content appears.

It’s a narrow wedge, but narrow wedges have a way of expanding. The advertising use case funds the infrastructure. The infrastructure enables new use cases. Amit started with a vision of billions of consumers experiencing interactive content. He might still get there, just through a door marked “enterprise” rather than “consumer.”

The full conversation with Amit covers the specific technical decisions behind Flam AI’s architecture, details of their Samsung voice campaign, and a deeper discussion of when glasses might actually replace phones. Watch the Founder Thesis episode for the complete story.

Amit Gaiki is co-founder and CTO of Flam AI. The company has raised $22 million and serves brands including Samsung, Google, ICICI Bank, and Dabur across India, the Middle East, Japan, and the United States.

Listen now!

Other ways to listen:

Apple Podcast | Amazon Music

Your Feedback matters

As always, I’d love to hear your thoughts! Whether it’s about this episode or ideas you’ve been playing around with, shoot me an email at ad@thepodium.in. Your feedback keeps these conversations going, and I’m always up for chatting about your startup ideas too.

Until next time,

Your Host,

Akshay Datt