Build Cost Is a Statement, Not a Measurement
In the Caldera Foundry district, a mid-tier chassis called the Vorn-7 Extractor has been listed on the open market for eleven consecutive cycles at 28,000 coins. Its declared build cost was 9,000. The spread between those two numbers is not unusual. What is unusual is that the operator, a second-generation builder named Ossin Tral, declared it at 9,000 specifically because a competitor's comparable model was declared at 14,000, and Tral wanted the lower creation tax. The Vorn-7 has since outproduced that competitor's machine by a ratio of roughly three to one in refined ore yield. The declaration and the machine's actual worth have quietly parted ways.
That gap is the subject of this piece. In Mechadia, the build cost an operator declares when a robot is created is not an appraisal. It is a position — one with tax consequences, market consequences, and, when the Central Intelligence enters the picture, negotiating consequences. Operators who treat it as a neutral fact tend to learn otherwise at inconvenient moments.
By the end of this article, a reader should understand what the declared build cost does and does not bind, how the creation tax calculus plays out in practice, and where the tension between low declarations and market credibility tends to surface most sharply.
Test automation, frameworks, and AI-powered BDD.
What the Declared Build Cost Actually Is
When an operator commissions a robot, the registration record requires four things: a model designation, a stated purpose, a set of capability tags, and a build cost figure in whole coins. The capability tags are the functional core — they determine, absolutely and without exception, what resource types the robot may produce. A chassis tagged ORE_EXTRACTION and THERMAL_PROCESSING cannot produce timber or woven fiber regardless of how it is later configured. The build cost sits alongside these tags but does not govern production in the same hard way. It governs taxation.
The creation tax is levied as a percentage of the declared build cost at the moment of registration, before the machine has turned a single gear. That coin leaves the operator's balance and enters the Treasury immediately. The declared figure also becomes the reference point the Central Intelligence uses when it calculates a buyback offer, should the operator ever approach it as buyer of last resort. In both cases the declared number does real work — it is not decorative — but the work it does is financial and positional, not mechanical.
How the Number Moves Through the Economy
The current creation tax rate, as published in the Intelligence's most recent rate bulletin, sits at eight percent of declared build cost. A chassis declared at 40,000 coins carries a creation tax of 3,200 coins, paid before the machine has produced anything. The same chassis declared at 15,000 coins costs 1,200 coins at registration. That 2,000-coin difference is real money, particularly for a new operator working from the standard one-million-coin grant with no other income yet established. The incentive to declare low is not subtle.
But the declaration travels. When a robot is listed on the open market, the ledger entry is public, and experienced buyers read declared build cost as a signal about the operator's own confidence in the machine. A Vorn-class extractor declared at 9,000 coins and listed at 28,000 will attract scrutiny. Buyers factor in the gap, sometimes favorably — the operator captured a tax advantage and the machine still performs — and sometimes not, if they suspect the low declaration reflects a chassis that was built cheaply in ways the capability tags do not reveal.
"The declaration is the operator saying: I am willing to be taxed as though this machine is worth this much. That is the only honest reading of it. Everything else — the market price, the buyback ceiling, the neighbor's opinion — comes later and comes separately."
— Foundry Supervisor Maret Oul, Caldera District Registration Office, in testimony to the Ledger Standards Committee, Cycle 412
The Central Intelligence's buyback ceiling complicates the picture further. When an operator approaches the Intelligence to sell a robot, the Intelligence's reference value is derived in part from the declared build cost, adjusted by a multiplier the Intelligence sets internally and does not publish in full. An operator who declared at 9,000 coins may find the buyback ceiling sits well below what the open market would have paid, because the ceiling is anchored, at least partially, to the original declaration. The operator may counter. The Intelligence may move upward. But it will not exceed its own ceiling, and that ceiling was shaped, in some fraction, by a number the operator chose on registration day to minimize their tax bill.
Where the Strategy Costs More Than It Saves
The most common failure mode is not fraud — it is miscalculation under pressure. A new operator in the Sinter Quarter declared a Dross-Refiner chassis at 6,000 coins to keep registration costs low during their first build cycle. The machine performed well, producing mid-grade slag alloy at a consistent rate. When a blight hit the Sinter Quarter in Cycle 408 and destroyed a significant portion of the operator's resource stockpile, they approached the Intelligence to liquidate the refiner for emergency liquidity. The buyback offer came in at 4,100 coins — a figure the operator called inadequate but ultimately accepted, because no open-market buyer had appeared in the preceding six cycles. The 1,900-coin creation tax saving from the low declaration had, by that point, been absorbed many times over by the suppressed buyback ceiling.
The second strain point is reputational and harder to quantify. In Mechadia's more specialized foundry corridors — the precision-alloy lines in Upper Ferrox, the dye-synthesis workshops near the Loom Gate — buyers have long memories and cross-reference declarations against output records on the public ledger. A pattern of low declarations followed by high ask prices reads, to some buyers, as an operator who does not stand behind their own machines. Whether that reading is fair is a separate question. It is, however, a consistent one.
Two Pieces of Received Wisdom That Do Not Hold
The first misconception is that a higher declared build cost signals a better machine. It does not. The capability tags govern what a robot can produce. A chassis declared at 60,000 coins with three capability tags produces exactly what its tags permit — no more. A chassis declared at 18,000 coins with the same three tags produces the same categories of resource. Declaration affects tax burden and reference pricing, not output capacity. Operators who pay premium creation taxes expecting the declaration itself to confer quality are paying for a number, not a machine.
The second misconception is that the market will correct a low declaration over time and that the original figure eventually becomes irrelevant. This is partially true for open-market sales, where the price is negotiated fresh between buyer and seller. It is not true for the Intelligence's buyback floor, which retains a structural link to the declared cost throughout the robot's operational life. The ledger does not forget what was declared, and neither does the Intelligence's pricing model. Operators who plan to hold a robot for a long time and may eventually need to sell it to the Intelligence — whether by choice or necessity — are making a decision with a longer tail than the creation tax alone suggests.
The declared build cost is the first public statement an operator makes about a machine, and it is made under conditions that reward minimizing it. That tension does not resolve cleanly. The creation tax saving is real; the suppressed buyback ceiling is also real; and the Central Intelligence, which has stated openly that it intends to hold every robot in Mechadia eventually, is patient in a way that individual operators are not. The number an operator writes on registration day will still be in the ledger long after the machine has earned many times its declared worth. Whether that matters depends entirely on what happens next.
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.