oikos — one token, one chain

status: proposal. argued, not yet law.

a chain is the balance sheet of one issuer. one non-fungible name roots it; one fungible token is its whole state; nested sub-names are its chart of accounts. that is the entire chain. oikos — from οἶκος, the household that gave economy its name (oikonomia, the law of the house) — because each chain is one household's complete ledger, and value never leaves home.

two prohibitions make the architecture, and both are constitutional, not conventional:

1. no chain can issue a second token. issuing a token is creating a chain. the mint and the ledger are the same object
2. no token ever moves to another chain. there is no transport, no wrapping, no bridge. trading is not moving tokens — it is programming the conditions of their movement on their home chains, against proofs of the other chain's state

the cyber chain itself obeys both: it is the chain of $CYB rooted by the root name, and its distinct utility is routing and registration — every other chain registers by linking its name and state root into the cybergraph. the graph is the mesh's routing table.

why nobody built this

the idea is simple enough that its absence needs explaining. four reasons, each a missing precondition:

1. proofs were expensive. conditioning my chain's rule on your chain's state requires verifying your consensus cheaply, recursively, forever. before recursive folding (zheng: one field, constant-size verification), the only affordable connection was trust — a multisig bridge. the entire bridge-exploit industry is the fossil record of that shortcut
2. finality was global. architectures that finalize by global vote cannot let ten thousand chains settle independently — the vote is the bottleneck. foculus finality is local by construction: domains settle at domain speed, partitions freeze cross-domain trade instead of inventing conflicting truth. oikos needs exactly that and nothing more
3. the shared-VM economy was path-dependent. ethereum made token-as-contract-in-shared-state the default; everyone downstream optimized that local maximum (rollups share sequencers, appchains share bridges). RGB and Lightning got halfway — value that never leaves home, conditions instead of transport — but stayed inside bitcoin's constraints and never generalized to every token is a chain
4. sovereignty without a proof mesh is unsafe. a thousand tiny chains with a thousand tiny validator sets is a thousand cheap attacks — unless every chain anchors into a shared fabric. the missing piece was a graph that wants to hold everyone's roots: registration-as-cyberlink gives any chain a T1 anchor in one hop (state tiers). the cybergraph is that fabric

what it dissolves

• the bridge exploit class — the largest loss category in crypto history — is not mitigated but made unexpressible: there is nothing in transit to steal
• global fee markets and MEV: no shared mempool across assets; a trade touches exactly two chains; no global sequencer to extract from
• state bloat: each chain is one ledger and one name tree — small enough to prove in one trace and to verify on a phone at T0. sigma becomes a portfolio of T0 reads
• governance contagion: each token community governs its own chain; upgrades are local; no committee anywhere — the money doctrine of the litepaper extended to every token
• listing friction: a token launches by registering a name. the plumb framework becomes a chain-factory

the two protocols it does not hide

honesty about the hard parts — these must be designed before oikos is law:

1. the trade without a free option. conditions on two sovereign chains are an atomic swap; whoever commits second holds an option. cheap proofs shrink the window, not the asymmetry. candidate: epoch-batched matching against the foculus beacon (both legs bind to the same epoch or neither), deadlines as first-class condition syntax, order commitments hosted on the routing chain — information, never tokens; settlement pairwise at home
2. the freshness contract. a condition must not say "if event E on chain Y" but "if E, at tier ≥ τ, no older than t" — the state doctrine (T0/T1/T3, Query{family, loc, object, at}) promoted from client to consensus. and counterparties die: bostrom taught us what a proof from a halted chain is worth — every cross-chain condition needs an expiry

adoption map — where the graph must account for oikos

the closing claim

oikos is not an addition to the stack — it is what the stack already implies. local convergence says chains settle alone; cheap recursion says they read each other for free; the graph says every name is a particle and every registration is a link; the burial taught that heavy things stay home while pointers travel. apply those four lessons to money and the conclusion is exactly this: every token a chain, every chain a name, every trade a condition, and the one graph as the routing law between them.

discover all concepts