Minus One — An Investment Thesis

Failure is inevitable. So is success. I back the people who learn faster than they fail.

1. The diagnosis: where Indian ambition leaks out

Technology, properly understood, is how the future gets built — not finance, not arbitrage, not the next layer of abstraction over someone else's infrastructure. And yet capital, the thing that was supposed to build the future, has quietly learned to bankroll the present at a markup. It prefers the legible bet: traction it can underwrite, founders it can pattern-match, sectors with comps. It backs the twentieth food-delivery wrapper and starves the first person trying to build a hard thing that doesn't yet photograph well.

In India this takes a specific shape. A country with a billion people and a genius for engineering has, for a generation, pointed most of its best minds at three exits:

The campus-to-MNC pipeline — brilliant graduates optimizing someone else's roadmap for a foreign P&L.
The arbitrage startup — copy a Western model, localize it, race to the bottom on discounts, call it innovation.
The quick flip — build for the acquisition, not the problem; raise the next round as the product.

None of these is shameful. All of them are small. They take the most capable people we have and aim them at problems whose ceiling is known in advance. The talent is world-class. The ambition is rented.

The gap is not money. India has money. The gap is conviction at the earliest, ugliest stage — the willingness to back a person before the thing is real.

This is a personal thesis. I invest my own money, in small cheques, as one person. There is no "we." There is an "I," and the founders I choose to stand behind.

2. The contrarian question

Every bet I make starts from one question:

What is someone building, right now, that almost everyone who hears about it thinks is a waste of time — and is wrong to think so?

The answer is almost never comfortable. The best -1 ideas sound, on first contact, like one of three things: naïve, impossible, or beneath serious people. Smartglasses sounded like a toy. Reusable rockets sounded like a hobby. A code editor sounded like a feature, not a company.

The job at -1 is to tell the difference between an idea that is wrong and an idea that is merely early and unfashionable. Consensus cannot help you here, because at -1 there is no consensus — there is only a person and their conviction, and your reading of both.

3. What "-1" means, and why I start there

Most capital begins at zero — a prototype, first users, something to point at. The interesting, overlooked moment is the one before:

-1 — Before zero. No product, often no co-founder, no proof except that the founder cannot stop. The idea is half-formed and a little insane. This is where everyone else says come back later.
0 — The idea exists outside the founder's head. A working prototype. The first people who are not friends start to care.
1 — It exists in the world and is useful to it. Not the finish line — the first real one.
X — Whatever it becomes when a founder with conviction is given years and is never abandoned.

Minus One backs the leap from -1 to 0 and stays for the whole walk to X. The cheque is deliberately small — enough to let someone stop doing the thing that pays rent and start doing the thing that matters. The support is deliberately unlimited.

I start at -1 because that is where belief is scarcest and therefore most valuable. A cheque at 0 is a transaction; a hundred people will offer it once the thing works. A cheque at -1 is a bet on a person, and it is the only kind of capital that changes whether the future gets built at all.

The model is the first cheque a believer once wrote to a college student — not because the spreadsheet justified it, but because the person did.

4. The arithmetic of the underdog

This is the core of how I think about the bet.

I believe two things at once: failure is inevitable, and so is success. They are not opposites — they are the same process seen at different moments.

Treat each serious attempt as a shot with some probability of working. If every shot were independent and identical, the probability you have succeeded after N attempts is:

P = 1 − (1 − p)^N

As N grows, P → 1. Given enough tries, success is not luck; it is arithmetic. The more you try, the more certain success becomes. But two things bend this curve.

1. You learn — if you refuse to repeat mistakes

A founder who never makes the same mistake twice raises the odds of every later shot. Let the failure odds decay with each attempt at a learning rate λ:

(1 − p_k) = (1 − p₁) · e^(−λ(k−1))

Then the probability of having succeeded by attempt N has a clean closed form:

P = 1 − (1 − p₁)^N · e^(−λ · N(N−1)/2)

Two engines compound here. N — raw attempts — lifts P linearly in the exponent. λ — learning — lifts it quadratically (the N(N−1)/2 term). Trying matters; trying while learning matters far more. A founder with λ = 0 repeats their failures and converges slowly or never. A founder with λ > 0 turns every failure into fuel for the next shot.

2. At -1, your number of shots is brutally small

You do not get infinite tries. The attempts you can afford are bounded by your runway:

N ≈ B / c

where B is everything you have to spend — capital, time, morale, credibility — and c is the cost of a single attempt. At -1, B is tiny, so N might be 1, 2, maybe 3.

This is the entire reason the first problem must be chosen with conviction. When N is small, learning has had no attempts to act yet, and the formula collapses toward its first term:

P ≈ 1 − (1 − p₁)^N

Now P is dominated by p₁ — the probability your first shot lands. You cannot lean on the law of large numbers when you do not have large numbers. So at -1 you must do two things:

Pick a first problem you can be sure of — high p₁. This is what conviction is, expressed mathematically: a justified belief that p₁ is high enough to clear the bar with the few shots you actually have.
Make each attempt cheap and fast — low c. Low c buys more shots N from the same B, and it tells you quickly whether p₁ was real. At -1 the problem must be solvable quickly.

The complexity ladder

Problems lead to solutions, and solutions lead to problems. Every problem you solve does two things: it reveals the next, harder problem, and it raises B — you bank capital, credibility, a team, and conviction. A larger B buys a larger N, and a larger N lets you survive a lower p₁. So you climb:

Stage 1:  high p₁, low complexity, small B   → win fast
Stage 2:  lower p₁, more complexity, larger B from Stage 1
Stage i:  p₁ falls, complexity rises, Bᵢ rises to pay for it

You begin at -1 with a problem simple enough that success is nearly forced, and you spend each win on the right to attempt something harder. That is how an underdog who starts with almost no shots ends up taking on problems they could never, at the start, have afforded to fail at.

So what I underwrite is not the idea. It is the founder's λ and their honesty about p₁ — a high learning rate, and the discipline to pick a first problem they can actually win. Give that to someone relentless enough to keep spending B, and both failure and success become inevitable. My job is to make sure the failures are the cheap, early, instructive kind — and to keep the founder backed and standing long enough for the arithmetic to finish.

5. Definite optimism

You can be optimistic or pessimistic about the future, and you can be definite or indefinite about it. The indefinite optimist believes things will get better but has no plan, and so diversifies, hedges, and waits. The definite optimist believes the future will be better because specific people will build specific things — and so concentrates, commits, and helps.

Minus One is not a fund; it is a definite optimist with a chequebook. I invest my own money, one person, the way friends and family do — fast, on conviction, off a single memo of what you're building and one conversation. No committee, no quarters of diligence. I do not spray small cheques across a portfolio to harvest a power law I don't understand. I pick a small number of people I believe are building a specific, important, hard thing, and I go all in on them — emotionally, practically, and for as long as it takes.

This has consequences:

Concentration over coverage. Few founders, deep relationships. I would rather be the most useful person on three cap tables than a line item on thirty.
Conviction over consensus. If the round is hot and oversubscribed, I am probably too late and it is probably too legible. I want to be early enough that backing the founder is an act of faith, not a queue.
Years, not exits. Hard things take time. I underwrite the founder for the decade, not the markup for the next round.

6. What I look for

I am looking for a person more than a plan, because at -1 the plan will change five times and the person will not.

The founder:

Obsession over ambition. Ambition wants to be a founder. Obsession needs to build this specific thing and would do it with or without me. I back the second kind.
Builders, not narrators. Show me the thing — the prototype, the repo, the broken first version. Conviction over credentials, always. The fresh graduate, the undergrad, the masters dropout, the self-taught — all welcome, none penalized.
Relentlessness against odds. I am specifically looking for the underdog: the person for whom everything is stacked wrong and who builds anyway. When something is important enough, you do it even when the odds are not in your favour — and those are exactly the people I want to back. "Against all odds" is not a risk I tolerate. It is the filter.
Builds for the thing, not the exit. The founders I want are not trying to get rich. They are trying to make something useful exist, and money is the fuel, not the point. The irony is that this disposition also tends to produce the largest outcomes.

The thing:

Hard tech first, great software welcome. Frontier hardware, devices, energy, robotics, biology, the physical world — and deep software when it is genuinely deep. No fixed sector. I want bets where success requires solving something real, not just out-marketing a clone.
Asymmetric and unfashionable. If it sounds obviously good to everyone, it is probably already backed and probably not large. I want the idea that is easy to dismiss and hard to forget.
A future worth wanting. The test is not "will this make money," it is "if this works, is the world meaningfully better and is India meaningfully stronger?"

7. What I will not back

A thesis is defined as much by its refusals.

D2C and consumer brands. Good businesses, wrong backer. They are won by capital and distribution, not by the kind of conviction I have to offer.
Me-too apps and arbitrage clones. A localized copy of a Western model is a job, not a frontier.
The optimized exit. If the pitch is built around who will acquire it, the founder and I want different things.
Anything where the hard part is raising the next round rather than building the thing.

Saying no to these is not snobbery. It is honesty about where a small cheque and one person's relentless support can actually change the outcome.

8. Nation first

India does not need more wrappers on foreign infrastructure. It needs people who build the hard, sovereign things — the hardware we currently import, the deep technology we currently license, the frontier work we currently watch others do. Every founder who chooses to build a genuinely hard thing here, for the long term, is doing something larger than a startup: they are deciding that the future gets manufactured in India rather than merely consumed in it.

So "nation first" is not a slogan stapled to a financial thesis. It is part of the underwriting. I weight, deliberately, toward founders building things that make the country more capable, more self-reliant, and more technologically sovereign — for humanity, and not for the exit.

9. The bet, plainly

I decide fast, friends-and-family style. One memo of what you're building, one meeting. No fund, no committee, no months of diligence.
I write the first cheque, when it is hardest to write. Small in size, large in meaning.
Then I show up. Intros when needed, time when not asked for, presence when it breaks. First to believe, last to leave.
I concentrate on a few and stay for years. No spraying, no abandoning.
I measure myself by whether founders go from -1 to X — not by whether I got a clean markup at the next round.

If you are building something the world isn't ready for yet — hard, unfashionable, against the odds, for the nation and for humanity — that is the entire thesis. Write to me. No deck required, though send one if you have it.

rakesh@atminusone.com

Minus One — I back founders at -1.

Inspired by my brother — imagineTej.