This is the one section that changes. Everything above describes XChain's durable design and vision — it stays valid as the network grows. Here live the parts that move over time: what is running today, what is in progress, the roadmap, and the network read straight from the chain — so those numbers never go stale. If any sentence here ever looks dated, the live panel is the ground truth.

The live network

Reading network status from www.xscan.org…

Read directly from www.xscan.org/rpc|lcd/aN (CORS open, so it loads in your browser). If the network is briefly unreachable, a link to the explorer is shown instead of stale numbers. The fault-tolerance verdict is derived from the live power distribution — with two validators where the largest holds more than a third, losing it halts the chain; add nodes until each holds no more than a third and the verdict flips automatically. No node/validator count is hardcoded anywhere in these docs — this panel is the single source.

✅ Here now — running, and verifiable by anyone

The Testnet MVC is live on real infrastructure (current node/validator set and fault tolerance in the panel above). 6 of the 7 primitives are real on-chain modules (UIR, UAC, XMP, DA, VLC, SC); the MVC acceptance passes end-to-end on real nodes + a real relayer. The explorer www.xscan.org reads chain-id and denom at runtime.

🟡 In progress / stand-ins — stated plainly

Item	Where it really is	What that means for a reader
Proof system	The real Groth16/BN254 verify is merged into `master` and proven end to end on devnet nodes — but the public network still runs the pre-upgrade binary, so the prover on the live chain is MOCK. `ProofSystem.MOCK` is blocked off-testnet by a fail-closed chain-id guard	Two distinct steps remain open, not one: the live network has not been upgraded to the verifying build, and the circuit is still a stand-in `MiMC(prev,tx,new)` (~164-byte proof), not the real SP1 wrap and verifying key. Closing both is exactly what Phase 4 — Zero-knowledge, end to end means below
EVM	Merged to `master`, not live	Going live on A1 needs a re-genesis (coin-type 118 → 60), so a `master` build differs from the running testnet; to match the live network, check out the tag `v0.2.1-testnet`
Public repository	Not yet published — the `xchain` org name is held by others	The clone URL in Run a Node is a placeholder until then

Do not read the running network as having real cryptographic verification yet. The verify path is real and devnet-proven, but the public chain has not been upgraded to it — and even after that upgrade the circuit remains a stand-in until the real SP1 verifying key lands. Two honest steps, not one, and both are written down here rather than glossed over.

🔭 The roadmap — 12 phases, sequenced by outcome, not by date

The roadmap is the founding team's opening proposal, nothing more. Every phase — its scope, its order, even its existence — can be amended, extended or replaced by community vote through on-chain governance (x/gov), by the process described in Governance & Community. And a phase is complete when its outcome is verifiably real on a live network, never when a date passes.

Arc	Phase	The outcome that closes it
I · An open network — from founding-team project to infrastructure no one controls	1 · Resilience	The network survives the loss of any single machine, operator or region — proven by drills on the live network, not in theory
	2 · Openness	Anyone can audit the source, run a node and join the validator set using nothing but public artifacts — its operation no longer depends on trusting the founders
	3 · A universal execution surface	The already-built EVM layer goes live; any Ethereum developer deploys with the tools they already use
II · Verifiability at full strength — make "verify, don't re-execute" true end to end, under attack	4 · Zero-knowledge, end to end	Proof production moves from stand-in to real, on open prover infrastructure — anyone can produce a valid proof
	5 · The Data Layer at scale	The data Domain carries production-scale volume under its full Tier policy, with published costs and third-party-verifiable inclusion
	6 · Adversarial resilience	Supply, anti-replay and containment invariants hold under continuous attack — fuzzing, chaos drills, a public attack catalogue
	7 · Independent assurance	External auditors and paid attackers examine the system; findings and fixes are public
III · Mainnet — from rehearsal to real value	8 · Community-ratified economics	Every economic and governance parameter is fixed by community vote — zero unreviewed defaults, zero founder-set parameters
	9 · Mainnet	Launch with Arc I's bars met from block one: a distributed validator set, no party above ⅓ power, the full acceptance suite passing on the real network
IV · A self-sustaining ecosystem — make the founding team unnecessary	10 · Builders	Outside teams ship production applications on XChain without ever talking to the founders
	11 · Horizontal scale	Capacity grows by adding chains and provers, not bigger machines; proving is decentralized across independent parties
	12 · Community stewardship	Upgrades, treasury and infrastructure run through governance as routine; the successor to this roadmap is authored by the community itself

Diagram

Figure 14 — The dependency map. Phases within an arc may overlap; Mainnet is gated by independent assurance and ratified economics — never by a date.

The proposal also names three explicit non-goals, themselves open to amendment by the same vote: no permissioned mode (permanently out of scope, not deferred) · no inflation switch (emissions beyond the pre-funded reserve are out of scope) · no date-driven launches (a phase that misses expectations is re-scoped by vote, not shipped unfinished).

Read together, the four arcs are one movement: progressive decentralization. Governance widens from the protocol (x/gov, already live) to the project's direction, strategy and this whitepaper — and Phase 12's outcome is precisely that the community, not a founding team, authors what comes next. Capacity grows the way Phase 11 describes, by adding chains and provers rather than bigger machines; any figure such as "1 billion TPS" remains aggregate nominal throughput (Execution ~100M + Data ~900M), never a billion composable transactions over one global state. The concrete network these phases start from is described in The Testnet.

The last word: if you can verify it on the live network, it is real; if it is written in the future tense, it is the vision We are building toward — out in the open, one honest milestone at a time. Come check Our work, and come build.