The Parent Contributes Knowledge, Not Capital

Counter-Positioned by Construction ended with a list of channels. A digital attacker launched inside an incumbent is closed by the incumbent because its results sit on the parent's accounts, its funding sits in the parent's budget, its brand carries the parent's obligations, and its board is the parent's board. Each of those is a path along which the parent's collateral-damage calculation travels to the decision to continue, and Finn and Bó were closed along them. A Counter-Positioned Clone is defined by having none of those paths. This memo says what replaces them.

The corpus does not usually write about deal terms, and the reason it does here is that the terms are not commercial preferences. Two things are invariant: the parent has no path, through capital, governance, brand, information or dependency, to the clone's decision to continue; and no party holds a cheap option that caps the investors' upside. Everything else in this memo is an instrument proposed to meet those two, derived one at a time from the failure each prevents, and an instrument can be replaced by another that meets the same invariant. An operator who understands why a term exists can defend it in a negotiation or substitute for it. An operator who only knows the term will trade it away for something that looks like a concession and is the parent's veto under another name.

The parent is paid for what it knows

The parent's contribution is the Operating Inheritance: the decision rules, exception history and commercial terms The Incumbent Wrote the Mandate and Never Encoded It scoped. That contribution has to be valued and paid for in equity, because a contribution treated as goodwill is a contribution the parent can withdraw, and a contribution paid for in cash makes the parent a vendor of knowledge rather than a participant in what the knowledge becomes. A minority stake — enough to matter to the parent's own accounts, not enough to carry a vote on the clone's commercial decisions — is the instrument. The parent is on the capitalisation table for what it supplied. It is not on it for what it will later wish to protect.

The knowledge itself travels under a licence with a scope. The reach chapter said a mandate states what an agent may touch and on whose authority, and the inheritance is a mandate's content; the licence is where its authority is written. Where the inheritance has to touch records as well as rules — customer histories, contract terms — the licence names the records, the purpose and the limits, because those records carry consents and obligations that do not transfer by being useful.

Capital comes from outside

The clone is funded by investors whose interest is its growth. This is the term that removes the budget-line channel, and it is the one the parent will most want to reverse, because funding the clone itself looks like control at low cost. It is control at the cost of the clone. A parent-funded clone is a line in the innovation budget, and Counter-Positioned by Construction showed what happens to that line when the parent's quarter turns.

Outside capital carries its own requirement, and the parent's purchase right has to be written to meet it. The structure has to be one investors will fund, and the requirement it has to meet is that the likely acquirer holds no cheap option, because an option at a pre-agreed price caps their upside and makes every subsequent round a discount. So the purchase right is priced at the greater of two numbers: fair market value set by independent valuation, and a floor expressed as a multiple of the last round so that the parent cannot wait for a down cycle. That is a floored price, not a market price, and the memo says so; the floor is what makes the right acceptable to the round, and the valuation is what makes it acceptable to the parent. When the floor binds, the parent pays more than the independent valuation, and whether a parent accepts that is a question of what the clone is worth to it, which the architecture does not settle. The floor, the valuation method and the exercise window are negotiating variables; the invariant is that the parent holds no cheap option that caps the investors' upside, and any pricing that meets the invariant is admissible. Whether a given round accepts a named acquirer on those terms is a question the round answers, not one the memo can; what the memo can say is that a round will not accept the alternative. The stake and the right are one package, priced together as the parent's return on the inheritance: the inheritance is the contribution, and the equity and the right are what it buys, valued jointly so that the same contribution is not given two independent prices. A right at a floored price has value even when its exercise price is not discounted, and that value is part of what the package pays for. The invariant sits apart from the pricing — no cheap option that caps the investors' upside — and how the package is priced within that invariant is for the three parties to settle.

Two further conditions make the right real rather than a veto in disguise. It expires: a right that lasts forever is a discount on every future round, so it runs for a fixed window and lapses unexercised. And when it lapses or is declined, the clone is freely sellable to anyone, including the parent's competitors, with the parent bound by a drag-along. A parent that can neither buy nor block has exactly the position the counter-position requires: a route into the new model, and no route into the decision to continue.

The trigger is revenue, not share

The parent will want its right to vest on market share, because share is the metric by which its own board measures threat. Share is the wrong trigger for three reasons the corpus can state without a lawyer. It is hard to measure in a fragmented market and easy to dispute. It is the metric most likely to be disputed in a concentrated market at exactly the moment the parent wants to buy. And it measures the clone against the parent's book rather than on its own terms, which is the Finn calculation written into the contract. A trigger on the clone's own revenue — a threshold, or a multiple of the parent's comparable line — is measurable from the clone's own ledger and auditable without the parent's data. It removes none of the notification or competition-law obligations an exercise carries; those attach to the transaction whatever the trigger says, and a parent that exercises faces them on the day.

What the parent asks for and must not get

Three terms will appear in the parent's first draft, and each is a channel from Counter-Positioned by Construction under a new name.

A non-compete on the parent's client base. This is the first thing the parent's counsel writes, and it dissolves the counter-position entirely, because the parent's clients are the ones already paying the Coordination Tax the clone exists to remove and are the readiest to move. A fenced clone is a clone that can only sell to people the parent has not reached, which is the market the parent is not worried about. What the parent can have is an endorsed referral path and early sight of which of its own clients are moving, which is worth more to it than a fence it would in any case be unable to enforce against its clients' own choices. A different and narrower thing — a restriction on the parent's own future origination in the clone's lines, as distinct from a fence on its clients — is a separate question this memo does not take up; if it is ever written, it is a proposed mechanism whose permissibility is for transaction-specific competition-law and contractual review, not one the architecture settles.

Operational information rights. A parent with full visibility into the clone's ledger holds the playbook, and the whole argument is that the parent cannot execute the playbook inside its perimeter. Some parents will try anyway, and the Integration Paradox says the attempt will cost them, but the clone should not fund the lesson. Investor-grade reporting — the figures any minority holder receives — and nothing operational. The Audit Surface is for the clone's Steward, not the parent's strategy team.

Board control. An observer seat until the right is exercised, and no vote on commercial decisions. A parent with a veto over pricing, hiring or market entry is the funder problem restored through governance: every decision that threatens the parent's book becomes one the parent can block, and the parent's calculation has reached the decision to continue by a different door.

The builder's position

The memo has so far described the parent and the investors. A clone also has a builder — a studio that extracts the inheritance, encodes it into the architecture and supplies the first Steward — and the builder's compensation has to align it with both sides at once or it becomes one side's agent.

A builder paid only a success fee on the parent's exercise is the parent's agent: its incentive is to hit the trigger, not to maximise the company, and the investors will price that in. The structure that holds is layered. A build fee from the round, for encoding the inheritance. An operating fee, monthly and declining, for the period the builder supplies the Steward. An equity stake, which aligns the builder with the investors on valuation. And a success fee paid by the parent on exercise, which is harmless once the builder also holds equity, because the builder then wants the exercise price high and not merely the exercise. Two of the four are paid by the parent, two by the round, and the builder is neither side's agent.

The term that makes this a business rather than a transaction is the platform. The builder retains the architecture and licenses it to the clone; on exercise, the parent acquires the company and a licence, not the platform. The Arco Flywheel is the corpus's name for why that matters: each build generates reusable infrastructure that makes the next build faster, and a clone built from one parent's inheritance is most of the architecture for a clone built from the next parent's. What that does to the studio's economics is a question for another memo.

The Operator's Verdict

An operator inside an incumbent who takes these terms to a board will be asked why the firm should accept a minority position, outside capital, no fence and no veto in a business built from its own knowledge. The answer is the two closures in Counter-Positioned by Construction. Every term the board would prefer — funding it, branding it, fencing it, governing it — is a term under which the board itself will close the business the first time its collateral damage becomes legible. The structure is not a concession to investors. It is the only configuration in which the firm's knowledge reaches the new model and the firm's barrier does not.

None of this describes an Arco transaction. No parent has contributed an inheritance, no round has been raised, and no purchase right has been written. The terms are derived from two public closures and the corpus's principles about where a boundary has to live, and whether Arco originates this way is a decision the Log does not announce. What the Log records is the configuration under which it could be done without the parent's calculation reaching the clone.

Technology changes what is possible. Who holds the capital determines whether the clone survives the parent.