The Economics of Things That Have Never Existed

There's a spreadsheet that gets built for every frontier venture, and it is always a work of fiction. It's called the TAM - Total Addressable Market - and for anything genuinely new, it is roughly as reliable as asking a fortune teller for a P&L. This doesn't stop anyone from building it. It stops almost no one from believing it.

The problem is structural, not lazy. A TAM is arithmetic performed on a market that already exists: count the buyers, multiply by what they spend, apply a plausible slice. It's forecasting by rear-view mirror, and for mature categories that's completely fine - the rear-view mirror is showing you a road that hasn't changed much. But point that same method at a category the world has never had, and you get a number with the confidence of data and the substance of a horoscope. What was the addressable market for electric light in 1878, priced against the candle industry? The honest answer - the one nobody could write on a slide - was that the candle market told you almost nothing, because light was about to stop being a thing you bought and start being a thing you assumed.

This is the trap. New categories don't compete for an existing market; they dissolve the boundaries of the old one and pool somewhere unexpected. The first serious estimates of the mobile phone market, done by very serious people, landed at a fraction of a percent of what actually happened, because they modelled it as a car-phone for executives. They forecast a product. What arrived was a change in what it means to be a human being who is reachable. You cannot TAM that. The category ate its own forecast.

So what do you actually do, if you're trying to price, fund, or forecast something that has no precedent? You stop pretending you're measuring and admit you're reasoning. Three moves have consistently earned their keep. The first is to forecast the problem, not the product - the pain is real and estimable even when the solution is imaginary. Nobody could size the market for orbital manufacturing by looking at orbital manufacturing; you size it by asking how much value is trapped in compounds we currently make badly, expensively, or not at all because gravity keeps interfering. The problem has a number. The product is still a rumour.

The second move is to price the option, not the outcome. Frontier ventures are not bets on a known payoff; they're bets on the right to a payoff if the world tilts a certain way. Finance has a whole mathematics for this - real options theory - that most strategy decks quietly ignore because it doesn't produce a single satisfying number. But it's the honest frame. You're not buying a market. You're buying a position at the table of a market that may or may not get invented, and the value of that seat rises with uncertainty rather than falling with it. This inverts the usual intuition, which is exactly why it's useful: the more genuinely unprecedented the category, the more a credible early position is worth, not less.

The third move is the one people find hardest, because it feels like an admission of failure: forecast in scenarios, and refuse to collapse them prematurely into an average. The expected value of a coin that pays nothing or pays everything is not "half of everything" - that number describes an outcome that cannot occur. Frontier economics is bimodal, sometimes worse, and averaging the modes produces a figure that is precise, defensible, and completely wrong. Better to hold the whole distribution in view and price your commitment to the shape of it.

None of this makes the fiction go away. You will still build the spreadsheet, because capital speaks in spreadsheets and you have to say something. But there's a difference between a forecast you believe and a forecast you deploy - a story sharp enough to move money while everyone in the room quietly understands it's a hypothesis wearing a suit. The failure mode isn't fiction. Fiction is unavoidable here. The failure mode is forgetting you wrote it.

Which leaves the question I keep coming back to, and haven't fully answered: what is a TAM worth when the market itself has to be invented? My working answer is that its only honest value is as a shared hallucination precise enough to coordinate a small number of brave people and enough capital to find out. That's not nothing. Most of the important things that exist were, at one point, exactly that.