Reflected Power — on capital, towns, and the things that cannot be rushed
part of ongoing series on Maxwell, Marx & Smith
I was twenty, on a chartered train crossing India in the dark, being shown factories the way pilgrims are shown temples. At every stop someone stood up and said a number — we employ this many people, we have been here this many years, our first year we lost money. Nobody ever said our valuation. That word hadn't been invented for us yet.
I've spent the fifteen years since watching capital get allocated well and badly, and I've come to think the whole difference is a physics problem the engineers already solved and the financiers keep forgetting.
the mountain
As a student I was sent to the Tehri dam — 260 metres of rock across a Himalayan river, sitting on a fault line, built to survive an earthquake that hasn't happened yet. My mentors were two Russian engineers, the tail end of a Soviet team that came in the 1980s. They drew before they computed. They treated a safety factor as a moral position.
Here's what Tehri teaches you. A preliminary survey was done in 1961. Power came in 2006. Forty-five years — longer than a career, longer than any fund could ever hold it. The men who ordered the survey were dead before the turbines turned.
You cannot rush a thing like that. It has a time constant, and the time constant is the only number that ever really matters. The Russian money that helped build it arrived because a state visit needed something to sign, and it left when the USSR dissolved — walking out for reasons that had nothing to do with the dam, because it had never really been matched to the dam in the first place. What stayed loyal was the rock, the river, and the obligation.
That's the whole thesis: capital allocated for reasons unrelated to the work will leave for reasons unrelated to the work.
maxwell's rule, borrowed
In transmission-line theory, if the source isn't matched to the load, the power you send doesn't vanish. A fraction reflects — comes straight back at the amplifier that sent it. Push a fast, high-gain loop onto a slow system and you don't get speed. You get ringing, then instability, then a broken amplifier.
The years 2020–2022 were the fastest, highest-gain money ever aimed at the real economy. Raise in a fortnight, deploy in a month. But the actual load — customer habits, supply chains, trust — moves in years. So the loop rang. And when a source can't match a load, the industry doesn't slow the money down. It manufactures a fake load it can deliver into: gross merchandise value instead of revenue, discounts instead of demand, cohorts shown at month three because month twenty-four would tell the truth. The subsidy is the matching network. Switch it off and the reflection arrives.
I was inside some of that, at OYO, at the height of it. Extraordinary people, a real problem, honest work — and, behind us, the largest liquidity experiment in history looking for somewhere to land. It felt like momentum. Momentum is lovely from the inside. There's no moment where a bell rings. That's the trap and I have a lot of tenderness for anyone who walked into it, because I did too.
smith, actually read
Everyone quotes The Wealth of Nations and skips the book Smith wrote first, about sympathy and the impartial spectator you can't out-clever. He never split markets from morals — we did that and put his name on it.
And he called all of this in 1776. On money managed by people it doesn't belong to, he said "negligence and profusion" always creep in — a perfect description of capital three layers deep in other people's pockets. On the price of capital, he understood it selects who gets it: set it wrong and you don't fund more things, you fund different people. And in the moral book there's the man cursed with ambition who sacrifices his peace for a trinket — not a knock on enterprise, a knock on the story enterprise tells about itself.
Jamsetji Tata never read a word of it and got all three right. Rejected by London financiers, he raised his steel money at home in 1907 — the public subscription filled in three weeks, ordinary people bringing small amounts. He had the idea in 1882 and died in 1904; the first ingot rolled in 1912. And before there was any steel, in 1908, he built the hospital. Wide streets, playing fields, temples and mosques and churches — a city, specified in a letter years ahead of the furnace. Not charity. A source deliberately built to wait thirty years, because the load needed thirty years.
The towns are the point. Bhilai, Rourkela, Durgapur — foreign money and foreign engineers who all went home within a few years. The towns stayed. That's what durable capital leaves behind.
who pays, and when
The most important thing Tehri taught me came fifteen years after it started generating.
The reservoir couldn't be filled to full level until 2021 — not because of geology or turbines, but because the resettlement of the families displaced by the water hadn't been finished, and the state wouldn't allow it. Whole communities had been moved; the rehabilitation was substantial but imperfect, and until the debt to those people was settled the asset was simply not allowed to run at capacity.
The social debt was the rate-limiting step on the physics.
I've never found a cleaner statement of the whole idea. You can treat the people as an externality if you like. The asset will decline to perform until you've paid them. There's a strange justice in that — the engineering itself refuses to let you skip the human part.
The same arithmetic, faster, is what a down round does to an employee: the founder keeps control and a salary, the fund writes it to zero against three other winners, and the person who took a pay cut for options holds a piece of paper worth nothing. Except here, unlike at Tehri, nothing in the system forces anyone to settle that debt first.
slow is the return
Then I spent years inside a company building engine controls — the small unglamorous computers that must not be wrong at forty degrees, at altitude, under vibration. And there is no month where that graph goes vertical. You win a customer and then you wait two years for the revenue. When something breaks in the field you send a person on a bus with a box.
It was the best thing I ever did, and here's why: it was matched. The money that came in had a horizon shaped like the work. And I'd tell anyone choosing where to spend their thirties — the patient version compounds and the flashy version doesn't. Not because patient people are better. Because patient work is impedance-matched, and flashy work is buying its matching network on credit.
Sridhar Vembu is the living proof it's a choice, not a sector. Zoho — bootstrapped, no venture capital, never listed, a billion dollars of revenue, run from a village near Tenkasi where he built a campus and a school that trains rural kids with no degree. Read his logic and Jamsetji's 1902 letter and they're the same document: put the good work where the multiplier lands on people with nowhere else to earn, and accept that it takes twenty-five years. Zerodha did it in finance, bootstrapped and profitable. DMart did it in retail — eight quiet years, then hundreds of stores.
Slow isn't the price of the returns. Slow is the return.
sanand
If I had one image for all of it: a field in Gujarat.
Tata's Nano plant moved there in 2008 after the land fight at Singur collapsed — built in fourteen months. The Nano itself failed, honestly and completely, a real product the market just didn't want. But the town didn't fail. Ford came and went; Tata took the plant back. And in 2026 a $2.75 billion semiconductor facility opened on the very same soil — India's first under the chip mission, thousands of jobs.
Farmland → a failed little car → memory chips, in under twenty years, and the place never emptied out. Because the durable asset was never the product. It was the ecosystem around it — the graded land, the power, the water, the trained hands, the supplier who moved his shop, the school that got better because teachers could afford to stay. Capital put there doesn't evaporate when the thesis changes. The next thesis inherits it.
The whole current build-out — the fabs at Dholera, the chip town planned in Assam for forty thousand workers, thirty thousand people (mostly women, mostly first jobs) hired near Bangalore — is being sited on exactly that century-old logic. And the space startups flying real rockets are behaving more like Tata than like the flameouts, because a launch is the most honest event in commerce. Nothing to fake.
three questions
I'm not a philosopher. I'm an engineer who learned finance, and I've ended up with three tests I run on everything:
Does the money reach the work? Not the story — the actual asset, the plant, the people, the lab.
Does it survive one full cycle with no new cheque? The honest ones go quiet and do arithmetic. The others start talking about momentum.
Who benefits, and where? In a district, for a generation — or in a metro, once, at exit? Tenkasi or a term sheet. Sanand or a shopping mall on the bones of a mill.
I got a front-row seat mostly by luck. A train at twenty. A mountain and two Russian engineers at twenty-one. A monetary experiment I mistook for momentum. A company that took years to become obvious. And now, watching capital come back to atoms — to fabs and rockets and things with mass — knowing it'll be misallocated again, because it always is, and that some of it will build towns anyway.
The mountain doesn't care what you called the round. It just tells you, eventually and without malice, whether the thing you were shouting into could absorb what you sent.
For a while we were shouting into a mismatch and calling the echo growth. The good news is that the echo fades, the rock remains, and the people who build for the load are still, quietly, building.
Longform posted here - https://www.linkedin.com/pulse/reflected-power-kunjan-gandhi-dshde/
