On September 16, Circle's Arc went to mainnet. I read the launch material twice looking for the consensus specification. It is not there.
What is there: USDC as the native gas token, sub-second finality, and an L1 label. What is absent: validator set size, slashing conditions, EVM compatibility, bridge architecture, and any audit artifact I could independently trace. A finality claim without a disclosed consensus mechanism is not an architecture. It is a promise with a latency number attached.
The same week, the BRICS CBDC interconnection track advanced from communiqué language to — by the account of India's own central bank governor — a feasibility study. India's commerce leadership has separately and publicly rejected a common BRICS currency in favor of bilateral arrangements.
Two digital money models. One calendar week. Two entirely different species of unfinished.
Context: what the two tracks actually are
Arc is a payment-specific Layer-1 operated by Circle, the issuer of USDC and a NYSE-listed entity. Its proposition is deterministic settlement: sub-second finality, USDC-denominated fees, and a compliance posture deliberately aligned with existing U.S. financial regulation. Visa executives have publicly framed it as a high-trust rail.
The BRICS track is wholesale CBDC interoperability — bilateral connections between central bank ledgers, intended to route cross-border trade settlement around correspondent banking. It is not a chain. It is a treaty with a technical annex.
Both target the same defect. Correspondent banking is a chain of intermediaries settling in T+2 or worse, charging layered fees, exposing every participant to the weakest correspondent in the path. Compressing that into atomic settlement is worth hundreds of billions in working capital. That is why both tracks exist.
The numbers circulating in coverage: roughly $308 billion in stablecoin float, $7.5 trillion in annualized settlement volume, six jurisdictions unlocking legal space for issuance, and a procedural vote on the CLARITY Act on September 15. I will return to the sourcing, because two of those figures do not survive a careful read.
I spent 2024 breaking down Celestia's data availability sampling and KZG commitment scheme. The lesson transfers directly: before evaluating a throughput claim, find out who runs the nodes. In 2017, auditing roughly 2,000 lines of 0x Protocol v1 Solidity, I learned the same thing at the contract layer. The interesting failures never live in the headline metric. They live in the permission table.
Core: the architecture, dissected
Start with the L1 designation. Circle did not build an L2. That is a deliberate and expensive choice. An L2 buys blockspace and security from Ethereum and pays for it in blob fees. An L1 issues its own security budget and pays for it in validator coordination, client diversity, and years of adversarial pressure it has not yet received.
The economics of that choice are asymmetric in a way most coverage misses. Post-Dencun, rollups migrated calldata into blobs. Blob supply is hard-capped — six per block at target, roughly 768 kilobytes per twelve-second slot — and it is shared across every rollup bidding for the same window. L2 demand is structurally rising while the supply curve is a protocol constant. My working position is that blob space saturates inside two years, at which point rollup gas fees double again as L2s bid against each other for a fixed resource.
Arc does not participate in that auction. It does not pay Ethereum for data availability, and it does not inherit Ethereum's settlement guarantee. That is the real trade: Arc escapes the blob ceiling by giving up the settlement guarantee — and the blob ceiling is the cheaper problem to have.
Then the gas token. Denominate fees in USDC and you capture genuine UX wins: no volatile fee asset, no separate acquisition step, no native token to distribute to validators. You also capture something else. USDC carries issuer-level freeze and blacklist authority baked into its contract interface. The unit that pays for computation on Arc is an asset a single company can immobilize under compliance pressure. A settlement layer that rents its gas token rents its sovereignty.
Now the finality claim. Sub-second finality at the L1 level is achievable. It is achievable because the validator set is small, permissioned, or single-sequencer. There is no fourth option in current consensus research. This is not criticism; it is arithmetic. Ethereum reaches probabilistic finality over minutes precisely because it accepts anonymous, permissionless validators. Arc reaches it in under a second because it does not. The honest phrasing is: fast because narrow.
What that means operationally is that Arc's security model is a corporate security model wearing a chain's clothes. Parameter changes, validator admission, upgrade authority, and freeze response all resolve to a boardroom. For a regulated payments rail, that is a feature. For anyone reading "L1" and hearing "credibly neutral settlement," it is a category error.
The single most consequential undisclosed fact is EVM compatibility. If Arc is EVM-compatible, it becomes a settlement substrate for DeFi and inherits MEV extraction, sequencer trust assumptions, and bridge surface — the exact stack I have been writing about since the Arbitrum analysis. If it is not, it becomes a bank consortium chain with excellent UX and zero composability. Both are viable businesses. Only one is interesting.
Counting trust surfaces: Circle's freeze list, the validator admission policy, the upgrade multisig, and the RPC endpoints. Four dependencies, none disclosed in the launch material.
Turn to the BRICS track. Its difficulty is not innovation. It is integration. Two central bank ledgers must agree on message formats, settlement finality semantics, FX conversion legs, and legal finality across jurisdictions that may not recognize each other's courts. That is a standards problem, and standards problems are resolved by committees, not by commits. The feasibility-study stage is not a sign of weakness. It is the correct stage for the work being attempted.
The political layer is where it stalls. A member state has publicly rejected a common currency and pushed bilateral deals. Two participants in the bloc have been in open friction. Any interconnection network that includes sanctioned jurisdictions exposes every participating bank to secondary sanctions — the network's legal risk is carried by its most exposed node. In 2022 I modeled Arbitrum's fraud proof economics and showed that validator collusion could delay finality indefinitely. The same analytical move applies here: ask what the settlement guarantee degrades into under coordinated adversarial pressure. For BRICS, the answer is a political negotiation, not a cryptographic one.
Contrarian: the blind spot nobody printed
Here is what "only one is ready" omits. Neither track is competing against the other in the near term. Both compete against the incumbent dollar rail that already moves the majority of retail stablecoin volume, on a chain nobody in this debate likes to name. That network's advantage is liquidity and habit — not compliance, and not speed.
Arc's genuine differentiation is regulatory, not technical. Six jurisdictions unlocking legal space matters more than a mainnet date, because regulatory certainty is the gate institutions actually pass through. Logic prevails, but bias hides in the edge cases — and the edge case here is that a compliant rail with no organic flow is a demonstration, not a market.
Then there is the exit. Speed is an illusion if the exit door is locked. Arc's fast finality rests on a permissioned validator set and a freezable fee token. Users can enter in milliseconds. Whether they can leave during a stress event without issuer coordination is the question that decides whether Arc is infrastructure or a hosted service.
Two figures also bother me. A $308 billion float and $7.5 trillion in settlement volume, quoted without a named source, do not reconcile cleanly against public stablecoin data. I do not build allocation theses on untraceable numbers, and neither should anyone reading this.
Takeaway
Watch three signals over the next two quarters. Whether Arc discloses validator set composition and slashing conditions within 90 days — silence there is itself information. Whether USDC float on Arc grows organically or through incentive programs, because subsidized TVL and adoption are different variables. And whether any BRICS member converts a feasibility study into a bilateral pilot. If the third happens before the first two resolve, the "only one is ready" thesis reverses faster than any model currently prices.
