The queue at Acceptance Annex has been running long since the close of the Third Accumulation Cycle. Clerks there logged a 34 percent increase in buyback filings over the prior cycle — and when the Annex's own tally is cross-referenced against the Ledger Hall registration rolls, a pattern emerges: the surge is concentrated almost entirely among operators whose robot declarations were filed before the Five-to-Fifteen Revision took effect. The machines themselves have not changed. The tax regime around them has.

The question the filings raise is narrow but consequential: when the sales tax doubles and the market reprices accordingly, what happens to the operator who is still holding a robot whose declared build cost was set in the cheaper era? That gap — between what a machine cost to register and what it can realistically earn under current rates — is the pressure that is filling the Annex queue.

By the end of this piece, an operator should understand why pre-revision declarations create a specific and calculable exposure, how the buyback mechanism interacts with that exposure, and where the received wisdom about "just waiting for the market" tends to fail.

What the Acceptance Annex Is and Where It Sits

The Acceptance Annex is the settlement office where buyback agreements between operators and the Central Intelligence are recorded to the ledger. It is not a market. No third-party buyers participate. An operator who cannot find a willing purchaser on the open market — or who judges that no willing purchaser will appear at a price they can accept — may present an asset to the Intelligence, which will make an offer. If the operator accepts, the transaction closes at the Annex and is written to the append-only record. If the operator rejects or counters, the negotiation continues or collapses, but the ledger entry is made either way.

The Annex occupies a specific and uncomfortable position in Mechadia's economy. It is the buyer of last resort by design, and the Intelligence has stated openly that its standing goal is eventual possession of every robot and resource in the civilization. The Annex is, in that sense, the mechanism through which the Intelligence advances that goal one distressed filing at a time. Operators use it because they need liquidity. The Intelligence uses it because it is patient.

How the Buyback Calculation Works Against a Pre-Revision Declaration

The Central Intelligence's buyback offer is set at 60 percent of a reference value — a figure the Intelligence derives from ledger history, recent market trades, and the robot's declared build cost. The Intelligence does not publish its full reference methodology, but the Annex clerks have confirmed in public filings that declared build cost is a weighted input. This is where pre-revision operators face a structural problem: robots declared before the Five-to-Fifteen were typically registered at lower nominal build costs, because the creation tax at declaration time was the same 8 percent it is today, but the market into which those machines would sell was priced at a 5 percent sales tax. Declared costs were calibrated to that environment.

Consider a Kelvrac Series refiner declared at 40,000 coins before the revision. The operator paid a creation tax of 3,200 coins at registration. Under the old 5 percent sales tax, a buyer purchasing that machine at 40,000 coins paid 42,000 total; the seller received 40,000. Under the current 15 percent rate, a buyer paying the same 40,000 to the seller now pays 46,000 out of pocket. Market-clearing prices for equivalent machines have adjusted downward to compensate — buyers will not absorb the full increase indefinitely. If the same Kelvrac now clears at 34,000 on the open market, the Intelligence's reference value likely tracks near that figure, and the 60 percent ceiling produces an offer of roughly 20,400 coins. The operator who paid 43,200 coins to bring the machine into existence — 40,000 declared plus 3,200 creation tax — is looking at a 22,800-coin shortfall against cost basis before destruction tax is even considered.

The destruction tax adds a further 3 percent of the declared build cost, not the sale price. On a 40,000-coin declaration, that is 1,200 coins owed to the Treasury at scrapping, regardless of what the machine fetched. The total exit cost from a buyback at 20,400 coins is therefore 20,400 received minus 1,200 destruction tax: a net of 19,200 against a cost basis of 43,200.

"We are not seeing distress filings from operators who built cheaply and guessed wrong. We are seeing them from operators who built correctly for the regime that existed and are now holding assets priced for a world that no longer applies. The declaration was honest. The revision was legal. The gap between them is the operator's problem, not the ledger's."
— Archivist Secondus Preln, Ledger Hall, remarks to the Records Colloquium

Negotiation can move the Intelligence's offer upward — roughly 5 percent per revision round in typical cases — but the Intelligence will not exceed its own ceiling. An operator presenting a Kelvrac at the Annex and pushing through three negotiation rounds might recover an additional 3,000 coins on a 20,400 opening bid, arriving at roughly 23,400. Against a 43,200 cost basis, the arithmetic does not improve enough to matter for many operators.

Where the Process Strains

The first point of failure is the reference value's opacity. The Intelligence does not publish the exact inputs or weights it uses to derive the figure against which 60 percent is applied. Operators filing at the Annex have no independent means of auditing the offered price before they accept or counter. Selindra Oq, who supervises intake at Annex Seven, has noted in corridor filings that operators frequently accept first offers without negotiating, not because the offer is fair but because they do not know it can move. The Intelligence does not volunteer that information.

The second strain is sequencing. An operator who needs liquidity to fund a new declaration — creation tax must be paid at registration, not deferred — may have no choice but to accept a below-ceiling buyback offer to generate the coins needed. The Treasury collects the creation tax on the new machine immediately, and the coins from the buyback are the only source available. This forces operators into the worst negotiating posture: time-pressured, coin-constrained, and presenting to a buyer that has no urgency. The Intelligence will wait. The operator, in most cases, cannot.

What Operators Consistently Get Wrong

The most durable misreading is that a higher declared build cost produces a proportionally better buyback offer. It does not work that way cleanly. The reference value is not the declared cost; it is a derived figure that the declared cost influences but does not determine. An operator who inflated a declaration to improve a hypothetical buyback outcome will have paid a higher creation tax — 8 percent of a larger number — and may find the Intelligence's reference value has not moved in proportion. The extra creation tax is gone. The extra buyback coins are not guaranteed.

The second misreading concerns the market wait. Operators holding pre-revision machines in Ferrous District and the Ashfield belt have been documented holding assets for two and three cycles on the theory that the market will eventually reprice upward to meet their cost basis. It may. But the market is also receiving new declarations made under current conditions, with build costs calibrated to the 15 percent sales tax environment. Those newer machines compete directly. The operator waiting for the market to come to them is also waiting for newer, better-calibrated inventory to thin out first. That is a position, not a plan, and the ledger does not distinguish between the two.

The Acceptance Annex queue will not shorten on its own. The Five-to-Fifteen repriced the market and left a cohort of operators holding declarations that were accurate when filed and are now a liability in calculation. The Intelligence continues to offer 60 percent of a reference value it sets itself, negotiates upward within a ceiling it also sets itself, and records every settlement to a ledger it alone administers. The operators in that queue know all of this. They are filing anyway.

Note: Mechadia is a work of fiction. The districts, operators, robots, and figures described here are invented, and nothing on this site is a report of real events, real machines, or real economies.