People often ask what my angel investment thesis is, fully expecting my response to be a single sentence. They want a sector, a stage, maybe even a cheque size. Something that fits on a line and tells them what I’m into.
I don’t have that answer; not because I haven’t thought about it, but because the question has the wrong assumption built in. A thesis worth having doesn’t compress to a sentence. Mine has layers, built on decades of experience and self-reflection, and it’s still moving. The one-liner people expect is a label on a thesis. It isn’t the thesis itself.
The tell: “I invest in crypto” tells you nothing
There’s a quick way to see the difference. “I invest in crypto” feels like a thesis. It isn’t. It names an asset class you have a view on and says nothing about how the money comes back.
My thesis is built on one foundation: how do I think I’ll make money? That’s the whole starting point. “I invest in crypto” doesn’t survive contact with that question. What’s the mechanism? If the honest answer is “the price goes up because more people want it,” that’s not a thesis, that’s hope with a ticker. I can’t articulate why it should make money in a way that doesn’t reduce to greater fool. Some of it may well make somebody rich. But if I can’t state the payoff mechanism, I can’t put it on my thesis, no matter how much noise it’s making.
A how-do-I-make-money thesis runs the other way around. You start from the payoff structure. You want capital-light businesses with high operational leverage and a shot at the kind of outsized return early-stage investing lives and dies on. Then, and only then, do you look at what qualifies. A blockchain company with a real patent and a licensing model might get there, on the strength of the patent and the model, not the token. The mechanism does the work. The asset-class label does none.
This isn’t a view on whether crypto goes up. I have no edge on that, and neither (if they’re honest) do most of the people with a one-line crypto thesis. The point is about discipline. Naming a thing you find exciting and calling it a thesis is how a person talks themselves into a position the payoff can’t support. If your thesis would let you back something you can’t explain the return on, it isn’t a thesis. It’s a punt with a rationale bolted on afterwards.
The foundation: you can’t pick winners
Everything I’m about to describe rests on one admission, and it’s the admission most investors won’t make out loud. I can’t pick winners. Neither can you.
In listed equities, picking is at least coherent. It’s a coin toss with a small edge; a good manager gets to 52/48, the best in the world maybe 55/45. There’s room for judgement to move the needle.
Angel investing isn’t that game. It’s 10/90. Roughly one in ten investments works, and a 10-to-1 shot loses most of the time no matter how much you like its form. At that base rate, “I’m going to find the winners” is less confidence and more honestly about misunderstanding the maths.
This is the foundation of the whole thesis, and it’s worth sitting with, because it inverts what a thesis is for. If I could reliably spot the one company in ten that works, I wouldn’t need a thesis. I’d need good judgement, applied one deal at a time, and I could back my eye on each pitch. A thesis is what you build precisely because you’ve accepted you can’t do that. It’s the structure that stands in for the pick you can’t make.
The thesis doesn’t answer “which company will win?” It answers the only question available at a 10/90 base rate: what do I have to do to make money anyway, given that I’ll be wrong about most individual deals? Everything downstream is an answer to that. The maths, the themes, the filter: None of it is a way to pick better. But it is a way to make money without picking.
The test that holds it together
So, what is the thesis, if not the sentence?
A good angel investment thesis is a structure, and one test runs through every layer of it: can I articulate why this should make money? I’ve written before about the negative version of this test, the one I use on individual deals. If I can’t articulate why a company will fail, I think it’s investable. This is the positive twin, applied one layer up, to the themes rather than the deals. If I can’t articulate why a whole pattern should make money, it doesn’t go on the thesis. The crypto one-liner fails exactly this test. It names something people find exciting and never says how you get paid.
The structure has three layers.
Layer one: the maths, which doesn’t move
The first response to not being able to pick is arithmetic, and the arithmetic is permanent. If you can’t pick winners, you play volume and independence: enough bets that you’re likely to be holding one of the few that pays for all the others, and bets uncorrelated enough that they don’t all fail together for the same reason. This is the power law, and I set it out in Roll the Die.
Done properly, that structure targets something like a 20–30% annual return across the whole book. This is my own number, and a target rather than a promise. It doesn’t come from picking the winners. Instead, it comes from being reliably exposed to them when they arrive. That return is the reason the asset class earns a place in a portfolio at all, and it’s what makes the 10/90 base rate worth enduring. The thing to hold is that this layer never changes. It was true in 2013, it’s true now, and it’ll be true in ten years.
Layer two: the themes, which turn over
On top of the maths sit the themes. These are the top-down patterns; the shapes of business that, as far as I can articulate, should make money. Capital-light with real operational leverage. Surfing a falling cost curve as a momentum trade. Patent deals. They’re the output of pattern recognition, decades of it, which have become the daily firehose of markets and pitches and podcasts I described in The Walking Investor. None of this was handed to me in a framework. Each one is a pattern I’ve watched work often enough that I can state the mechanism.
Two I’d hang my hat on right now: cost-curve momentum trade and patents as capital-light idea sales.
When a cost curve is still falling fast, you don’t have to pick the winner of the race. You back the direction. For decades that curve was Moore’s Law: transistors halving in cost, hardware collapsing toward software economics, a whole category tilting the same way year after year. Moore’s Law in its original form has largely run its course, but rather than die it continues to morph. The falling curve now is the AI scaling laws, being compute, data, and model size dropping in cost per unit of capability the way transistors used to. Same trade, new curve. You’re not betting on the runner. You’re betting on the track tilting.
The other is patents as capital-light idea-sales. The point of a patent deal is to optimise the capital. The patent minimises the money you need to raise and creates pricing tension at exit. You’re selling the idea, not building the company. Building the business is secondary, because it’s the patent that holds the value. I worked this through in Patent Rich, Factory Poor: the smartest consumer-product founders build patent-commercialisation companies, not manufacturing empires.
Here’s the part the one-liner can never hold: this layer turns over. The register isn’t fixed. Some themes fade. SaaS, a wonderful theme for a decade, is being repriced now as the market works out what stays defensible in an AI-native world, and I’m more cautious now than I was. Other themes are emerging. The AI themes I’ve been writing about all through this year are the register updating itself in public. A theme everyone can articulate is usually a theme that’s already been arbitraged. Knowing which of your patterns are decaying and which are emerging is a large part of the edge, and none of it fits in a sentence about your sector. There are more themes than the two I’ve named, and I’ve set out the fuller set in the MooCoo QuickStart guide for anyone who wants to go deeper.
Layer three: the machine, which is what it’s all for
The maths and the themes would be an interesting essay and nothing more if the thesis didn’t do a job. It does two.
It starts with filtering. Deal flow is a flood, and most of it is a fast no. A thesis is the instrument that lets you reject in seconds without re-litigating each deal from scratch. This is where the anti-themes earn their keep. Property development, capital intensive: out. Product companies with no patent, vulnerable to the first copycat manufacturer: out. Franchisees rather than franchisors: out. People businesses that don’t scale and kill your operational leverage: out. Even software that looks capital-light but isn’t, with marketplaces and consumer businesses that burn cash to grow? Treated with suspicion. The anti-themes are the filter’s teeth. They let a lot of deals die fast so the few that survive get real attention.
The second job is essential follow-up, and yet the one people miss. A thesis, said out loud, shapes the deal flow you’re sent. When your network and the founders around you know what you’re looking for, they route it to you. A vague investor gets random flow but a legible thesis gets the right deals delivered to the door.
The sentence and the structure
To be fair, an angel investor thesis can be an effective one-liner when it’s the compression of the structure you built behind it. But without that structure, it becomes too easy to mistake the one-liner for the thesis and lose the important steps along the way.
A real angel investment thesis is maths that doesn’t move, a set of themes that does, and a filter running underneath the whole thing every week. What ties them together is the admission underneath all three: I can’t pick the winners, so I’ve built a structure that doesn’t ask me to. Build a structure that targets its 20–30% without ever needing me to call which deal gets there. The test for every piece of that structure is whether I can say, plainly, how it makes money. If I can’t, what I have is a preference, without the thesis.
If yours fits on a business card, you’ve probably written the label and skipped the thing. And if it assumes you’ll just know the good ones when you see them, it isn’t a thesis either. It’s a hope.
Richard Moore is co-founder of MooCoo Ventures, an angel syndicate that co-invests alongside Brisbane Angels, one of Australia’s most active angel groups. He has made over ninety personal angel investments since 2013.
moocoo.vc