Methodology

How a thesis is built.

A repeatable discipline: begin with a plausible future, expose the assumptions it rests on, follow the value down the stack, and end with the specific assets that own it.

The spine

Every thesis moves through four stages.

01 · The narrative

Begin with a contradiction, not a category.

A thesis opens where two things are true at once and cannot stay true together: the world says it wants one outcome, but the system is still built for another.

From that tension the argument builds in a fixed order. It shows why the current system is under structural pressure — cost, scarcity, dependency, energy, regulation, security — rather than resting on imagination. It establishes why the moment is different now: the specific unlock that changed the timing, feasibility or urgency. It then renders the future state concretely enough that it can later be tested: what becomes more valuable, what becomes less, and which bottlenecks emerge.

The opening ends not with a summary but with a question: if this future arrives, what becomes scarce, strategic and hard to replace — and who owns it?

02 · Assumptions

State the assumptions, and what would break them.

Every thesis exposes its load-bearing assumptions rather than hiding them. Each is written so it can be checked, and each is stacked so the dependencies are visible: the future state stands only if the layers beneath it hold.

Assumptions are not asserted and left alone. They are paired with the strongest counterargument, taken seriously, and with the observable condition that would prove them wrong.

Device · Assumption Stack
What must be true for the thesis to hold
Illustrative. Each layer has to hold for the one above it to stand.
FS
Future state — the thesis is realised
04
Capital mobilises behind the shift
03
A structural constraint turns binding
02
The incumbent layer cannot absorb the change
01
The commodity layer stops capturing the rent

03 · Value accrual

Follow the value down the stack.

A theme being real is not the same as a theme being ownable. Value rarely accrues evenly along a chain: it thins where a layer can be swapped with an API call, and concentrates where replacement is slow, regulated, data-rich or physically scarce.

The map isolates where the profit pools are likely to sit — the bottleneck layers that become unavoidable if the future emerges — as distinct from the visible layer that attracts the attention and the capital.

Device · Value Accrual Map
Where the profit pools may emerge along the chain
Illustrative. Value thins where replacement is easy and concentrates where it is hard.
Upstream Control layer Data Applications VALUE POOL
Value concentratesThinner margin

04 · Ownership

Centrality is not ownability.

The Ownership Blueprint maps each future-state layer to the specific assets that own it, then keeps two questions apart that are easy to run together: how central a layer is to the thesis, and whether the asset can actually be bought, at an acceptable price, on acceptable terms.

Because those two axes rarely line up, the blueprint tends toward a barbell rather than a single basket: own the layers that are both central and ownable now; price the assets where dominance is already reflected in the entry; and watch the frontier options, sized as small, high-upside sleeves. Selectivity is the discipline — the blueprint is not a directory, and an asset earns its place only where the reason for exposure is specific.

Device · Ownership Blueprint
Future-state pillars, mapped to specific asset exposure
Illustrative. The bottleneck layer on the left, the way to own it on the right.
Compute Energy access Control layer Physical execution Listed control point Private round High-conviction owner Real asset
High-conviction pathSupporting exposure

05 · Disconfirming evidence

A thesis worth holding can be broken.

Conviction without falsifiability is only enthusiasm. Every thesis names the conditions that would make it wrong, each one observable, and each attached to a signal to watch over a stated window.

The bear case is mandatory, not decorative. A thesis is never presented as inevitable: it shows what must be true, what could break it, and what would weaken it before any capital is at stake.

What would make this wrong
  • The commodity layer keeps capturing the rent, and value never migrates down the stack.
  • An incumbent absorbs the control layer into an existing suite before a challenger can own it.
  • The structural constraint relaxes faster than expected, removing the scarcity premium.

06 · Research standards

The standards every thesis is held to.

Attribution integrity. Borrowed data is attributed in place. The credibility of a thesis depends on distinguishing original analysis from corroboration, and single-source or stale figures are flagged as such.

Evidence over assertion. Claims are traceable. The argument is confident because the thinking is visible — the evidence, the assumptions, the framework — not because the language is loud.

Verify time-sensitive facts. Roles, valuations, rounds and filings change. Live facts are re-checked before distribution, and every named asset is treated as a diligence candidate requiring independent judgment. Research is not investment advice.

The lexicon
Language the imprint avoids
Generic words are banned unless immediately backed by evidence.
  • well-positioned
  • innovative
  • promising
  • game-changing
  • massive opportunity
  • the next big thing

See the method applied.

Thesis 001 works the full discipline end to end — future state, assumptions, value accrual, falsifiers and a selective Ownership Blueprint across eight assets.