Arc vs Tempo: the 2028 clock that Circle signed with the SEC

Circle sold 807.5 million ARC tokens for $242.2 million and signed a commitment with the SEC to return that money if Arc, its Layer 1 blockchain, has not completed the transition to Proof-of-Stake (validators who put a deposit at stake and can lose it) by May 8, 2028, at the latest. The network launched its public mainnet on September 16, 2026, under Proof-of-Authority (identified and authorized validators, one by one, with no deposit to lose) and that same week minted the 10 billion ARC of the initial supply with a phrase from its own note: "this initial mint is not a commitment to launch ARC publicly." As of 06:44 UTC on September 18, the Arc RPC showed 17 addresses proposing blocks in uniform rotation and a base fee of 20 Gwei—0.001 USDC per 50,000 gas units—with $339.0 million in TVL (capital deposited in its protocols, according to DefiLlama at 06:45 UTC). Tempo, the payments chain incubated by Stripe and the venture fund Paradigm, had been on mainnet for 184 days with $60.9 million in TVL and has no equivalent clock because its documentation states "Tempo has no native token" and there is nothing sold to refund. What follows separates what that deadline mandates and what it does not.

What did Circle sell before launching Arc?

The ARC presale was closed in two tranches, both documented in 8-K forms, the material event notice that a U.S. public company files with the SEC. The first, signed on May 8, 2026, and registered on the 11th: 740 million ARC tokens at $0.30, approximately $222.0 million gross. The second, signed on June 29 and 30 and registered on July 2: 67.5 million tokens at the same price, approximately $20.25 million.

The Q2 10-Q (quarterly report to the SEC), filed on August 5, consolidates both sales: 807.5 million tokens and $242.2 million gross, of which $222.0 million was collected within the quarter. That document stated that "the total initial supply of ARC tokens is expected to be 10 billion," subject to increase by the programmatic operation of the protocol. On September 16, the launch note took the step: Circle "completed the genesis mint of the ARC token in the United States this week, creating the full initial supply of 10 billion tokens." The 807.5 million sold represent 8.075% of the minted initial supply, and the implied fully diluted valuation of $3 billion is now calculated on a closed supply.

The body of the May 8-K (Item 8.01) only states that the placement was "led by a16z crypto." The thirteen buyers are listed in Exhibit 99.1 of the same 8-K, the Q1 earnings release provided under Item 2.02: "a16z crypto, Apollo Funds, ARK Invest, BlackRock, Bullish, General Catalyst, Haun Ventures, Intercontinental Exchange, IDG Capital, Janus Henderson Investors, Marshall Wace, SBI Group, and Standard Chartered Ventures." The 10-Q adds the amount for one of them: an entity linked to IDG Capital, holding more than 5% of Circle's capital, purchased 83.3 million tokens for $25.0 million, 10.3% of the entire presale, declared as a related-party transaction.

What exactly does the May 8, 2028, deadline mandate?

The clause is written identically in both 8-Ks and repeated in the 10-Q. The agreements "provide for redemption rights in certain circumstances, including if the ARC tokens are not delivered or if the Arc network has not completed the transition to a Proof-of-Stake or delegated Proof-of-Stake consensus mechanism on or before May 8, 2028, or if certain buyer-specific legal, regulatory, or compliance conditions are not satisfied." Added to this is a lock-up period of at least one year starting from the date of the transition—not from the purchase—and up to four years of additional transfer restrictions.

The 10-Q describes the starting point: "Arc is expected to initially operate under a Proof-of-Authority consensus model. Over time, it could transition to a Proof-of-Stake or delegated Proof-of-Stake mechanism. If such a transition occurs, the network would introduce a native token." The token that Circle has already sold and minted is born, according to its own filing, from a consensus change described in conditional terms. The September 16 note attaches a year to it for the first time, using the same verb: the mint is "a significant technical milestone on the Arc roadmap as the network explores a future transition from Proof of Authority toward Proof of Stake in 2027." "Explores" is not a commitment, but 2027 falls before May 8, 2028: if the goal is met, there are 129 days of margin between the last day of 2027 and the contractual expiration.

The destination of the change is described in Arc's technical documentation with an adjective that defines the scope: the roadmap contemplates "a potential transition from Proof-of-Authority to permissioned Proof-of-Stake (each validator still requires authorization) to expand validator participation while maintaining compliance requirements." The contract requires a consensus mechanism and a date; it does not require that anyone can validate. Meeting the May 8, 2028, deadline with a set of validators still authorized one by one (SOC 2 certification, geographic distribution, and service level agreements, according to the consensus documentation), now with an ARC deposit, satisfies the letter of the agreement, extinguishes the redemption right, and starts the one-year lock-up clock without opening the network. The engine is Malachite, an implementation of the Tendermint BFT protocol (the consensus of the Cosmos family) with deterministic finality.

How does Circle account for the $242.2 million from the ARC presale?

Circle does not recognize the $242.2 million as revenue. The 10-Q explains that "funds received from the ARC token presale are recognized as deferred revenue until the company's performance obligations under the token purchase agreements are satisfied," and only then will they move to the Other Income line. Table 11.2 of the document shows the effect: the deferred revenue balance went from $11.512 million on December 31, 2025, to $246.203 million on June 30, 2026, meaning 98.4% of the balance at the end of the semester is the ARC presale (own calculation based on the two 10-Q figures). The other current liabilities item grew by $237.6 million in the same period due to that operation.

The Q1 10-Q, filed on May 11, is the most explicit regarding its risk factors: "there can be no assurance that the ARC token will ultimately be launched"; Circle "could decide not to proceed with the launch of the ARC token, or delay it for an extended period," and in that case could face "reputational damage, disputes with presale buyers, and/or increased regulatory or litigation risk." The phrase in the September 16 note—the mint "is not a commitment to launch ARC publicly"—is the press version of that risk factor. As of September 18, 2026 (EDGAR consulted at 06:43 UTC), Circle has filed only one 8-K with new agreements since the August 10-Q, the September 8 filing for the stock purchase of Tazapay (the Singaporean payment gateway), and it does not touch the token purchase agreements: the clock has not moved in any filing.

Why doesn't Tempo have an equivalent clock?

Tempo reached mainnet on March 18, 2026, according to its own FAQ page: 182 days before Arc launched its own. Its fee documentation opens with the phrase that explains the entire asymmetry: "Tempo has no native token." Fees—the base and priority tip—are paid in stablecoins that meet three conditions: being denominated in dollars, being issued as a TIP-20 contract (Tempo's token standard, equivalent to ERC-20), and having sufficient liquidity in the Fee AMM (the network's internal market that swaps the stablecoin for the fee). The distribution is direct: "all fees accrue to the validator that proposes the block."

Without a token, there is no presale; without a presale, there are no buyers with redemption rights; and without redemption rights, there is no date. Nor is there a written instrument with which to demand an opening date from Tempo: its documentation is limited to saying that "Tempo is continuously expanding its active validator set on a path toward permissionless validation" (anyone could validate), without a deadline.

The gas unit is the other side of the same asymmetry. Arc "denominates all transaction fees in USDC," and USDC is a liability of Circle: every unit in circulation is a redemption obligation of the company that builds the network and has sold 8.075% of its token. Anyone settling on Arc assumes counterparty risk from the issuer of the unit of account, operational risk from the set of authorized validators, and, when it arrives, governance risk from the token that will decide the parameters—and all three point to the same group as long as Proof-of-Authority lasts. Arc's fee design documentation states who keeps the amount: "both the base fee and the priority tip are credited to the block beneficiary," with no burn. On Tempo, the payer chooses the stablecoin, its issuer remains outside the perimeter of the chain operator, and the gas price depends on an internal market whose liquidity is the piece to watch: at 06:46 UTC on September 18, DefiLlama measured $44,143 of TVL in the Fee AMM. In balances, at that same hour, Arc had $634.7 million in USDC as the only dollar stablecoin, plus $7.0 million in EURC and $0.5 million in the tokenized pound GBPA; Tempo had $214.5 million in dollars spread across USDC, USDB, pathUSD, USDT, and DLUSD, with USDC at 39.0%. The previous day, at 15:54 UTC, Tempo had $148.8 million with USDC at 42.9%: $65.7 million more in one day, almost all in USDB and USDC. These dependencies, when the stablecoin is the center of the infrastructure, are addressed in the analysis of the stablechains that want to replace Tron and in the stablecoin risks guide.

Who validates on Arc and Tempo and how many nodes propose blocks?

The founding cohort of Arc, announced on August 5, 2026, consists of eleven names, listed in the table. The launch note clarifies that "along with Circle, Arc will have a phased rollout from the founding validator cohort," without stating how many nodes were active on day one. Tempo lists four named validators: MoneyGram, Stripe, Visa, and Zodia Custody by Standard Chartered.

Since neither publishes the size of the active set, we measure it from the outside: each block carries the address of the one who proposed it, and 500 consecutive blocks read by each network's public RPC tell us how many distinct addresses propose and with what distribution. On Arc, at 15:56 UTC on September 17, 17 distinct addresses proposed the 500 blocks, each between 29 and 30 times: uniform rotation, as expected in Tendermint with equal weights. These are more than the eleven named institutions; the correspondence between address and institution is not public, and the difference could be Circle and additional nodes from the same entities. On Tempo, at 15:57 UTC, 13 distinct addresses proposed the 500 blocks with unequal distribution (from 109 blocks for the most frequent to 25 for the least); one of them is the address 0x0000…0001, which may be a sentinel value and is counted separately: at least 12 operational addresses compared to four named validators. The measurement repeated at 06:44 and 06:45 UTC on September 18 over another 500 blocks gave the same result: 17 addresses on Arc (30 blocks for seven of them and 29 for the other ten) and 13 on Tempo (from 122 to 24 blocks).

Standard Chartered appears in three roles: a node under its own name on Arc, Zodia Custody—a subsidiary with its own licenses and risk perimeter—on Tempo, and Standard Chartered Ventures among the presale buyers according to Exhibit 99.1. DTCC, an Arc validator, has its own analysis as a depository in the piece on the DTCC token and its collateral value.

How much does a 50,000 gas transaction cost on Arc and Tempo?

Both networks document their fees with different transaction sizes. Tempo gives its cap and floor for the base fee for a standard transfer of 50,000 gas units: "the base fee cap is approximately $0.0006, with a floor during quiet periods of around $0.00003." Arc publishes its base fee in Gwei (billionths of USDC per gas unit)—minimum of 20, maximum of 20,000—and a design goal of "approximately $0.001 per ERC-20 transfer."

Arc's documentation retains the "(testnet)" label on the minimum base fee and warns that parameters "may change prior to mainnet launch"; as of September 17, the page had not been updated after the start. The public mainnet RPC does respond: at 15:56 UTC, the base fee for the last 100 blocks was 20.0 Gwei in all of them, with 74 transactions per second, and arc.etherscan.io showed a median of 20.166 Gwei. The testnet floor is the value observed on mainnet, and since the base fee did not move from the minimum in the measured window, the $0.001 goal per transfer and the 20 Gwei floor describe the same thing as of September 17, 2026. On Tempo, the base fee for the last 100 blocks at 15:57 UTC was 0.6 Gwei, $0.00003 for 50,000 gas, also at its floor. The reading on September 18 at 06:44-06:45 UTC repeated both values, 20.0 and 0.6 Gwei, with Arc at 34.7 transactions per second.

With these numbers, Arc's floor is 33 times higher than Tempo's floor and 1.67 times higher than its ceiling; Arc's ceiling is 1,667 times higher than Tempo's ceiling (own calculation based on the table parameters). In the September 17 window, an average transaction consumed 156,134 gas on Arc and 112,716 on Tempo: the actual cost exceeded that of the standard transfer on both networks, and the proportion remains.

What does Arc's first day show compared to six months of Tempo?

The figures for the two networks do not measure the same thing, and the table separates them by metric and source. A payments network can move a lot with little TVL, and a newly launched network can be filled with launch TVL that is withdrawn a week later: the DefiLlama series for Arc goes from $51,071 on September 16, 2026, to $339.3 million on the 17th.

The TVL for both networks is provided by the same protocol, Morpho, and on Arc, two lending markets—Morpho and Aave V4—concentrate 90.4% of the capital in the September 17 snapshot and 90.9% in the September 18 snapshot. Within that flat total ($339.3 vs $339.0 million), the distribution shifted overnight: Aave V4 went from $78.5 million at 15:52 UTC on the 17th to $126.5 million at 06:46 UTC on the 18th, and Morpho from $228.3 to $181.5; Aave V4 debuted one of its first markets on Arc, and its architecture is described in the analysis of the Aave v4 unified liquidity layer. Day one trading volume depends on who measures it: on September 16, DefiLlama only had Uniswap V2 and V3 adapters on Arc and did not see the launchpad curves (by 06:46 UTC on the 18th, it already totaled eleven, including Uniswap V4, SushiSwap V3, and Virtuals, and $129.6 million for September 17); the Dune dashboard cited by the press includes those curves, with $336.3 million in memecoin launchpads and $202.4 million in Arguspad (memecoin launchpad) alone, which also created 83,751 of the more than 97,000 tokens created that day. These are different perimeters and are not averaged. DexScreener gave a portrait of that traffic at 15:56 UTC: the highest volume pair on Arc, DUKE/USDC, had moved $17.1 million in 24 hours on $165,813 of liquidity, 103 times its pool. Tempo does not publish DEX volume or cumulative transactions; its most cited figure, the more than $1 billion in transfers over 30 days, was published by Crypto Briefing on September 2 without declaring an underlying source, and subsequent press attributes it to Tempo itself without a document. It is cited as press and is not compared with Arc's TVL or DEX volume.

The most comparable data in the table is the smallest: transactions in 500 consecutive blocks, measured at the same time by the same method. In the September 17 snapshot, Arc processed 79 times more than Tempo; in the September 18 snapshot—Arc at 34.7 transactions per second and Tempo at 1.10—29 times more. The 82% of memecoins on Arc compared to the payroll transfers on Tempo (Deel, the international payroll platform, issues its own stablecoin, DLUSD, on the network) explains why this transaction ratio says nothing about payments.

What signals will indicate if Arc moves to Proof-of-Stake before May 8, 2028?

From September 16, 2026, to May 8, 2028, there are exactly 600 days. Four verifiable signals without privileged access will tell if the commitment is being met:

1. A Circle 8-K or 10-Q that modifies the Token Purchase Agreements and moves or removes the date. Where to verify: EDGAR; the next opportunity is the Q3 10-Q, expected in November 2026.
2. An announcement of Arc's transition, and whether the resulting Proof-of-Stake retains the "permissioned" adjective from docs.arc.io: whether the change opens the network or only changes the selection mechanics depends on this.
3. The number of addresses proposing blocks, measured by public RPC (17 on September 17 and 18, 2026), as an indicator of whether the "phased rollout" is progressing.
4. A native token on Tempo, which would equalize the incentive structure and render the asymmetry described here moot. Where to verify: Tempo's fee documentation, which states "Tempo has no native token."

For those who must decide where to settle dollars, the operational question is which dependency they prefer to assume, because neither network is permissionless as of September 18, 2026, and both state so in writing: a gas unit mandated and issued by the chain operator, with a consensus commitment dated before the SEC and a self-imposed 2027 goal, or a basket of stablecoins selectable by the payer, with a declared path to opening and no deadline. Comparison with the same instrument requires weeks of Arc mainnet with real fees, and the next scheduled snapshot is September 24, 2026.

Sources and links: SEC — Circle 8-K, May-11-2026 (Item 8.01, first closing) · SEC — Exhibit 99.1 of the same 8-K (Q1 earnings release with the thirteen buyers) · SEC — Circle 8-K, Jul-2-2026 (second closing) · SEC — Circle 10-Q, Second Quarter 2026 · SEC — Circle 10-Q, First Quarter 2026 (ARC risk factors) · SEC — Circle 8-K, Sep-8-2026 (Tazapay purchase) · Circle — Arc founding validator cohort, Aug-5-2026 · Circle — Arc mainnet launch note, Sep-16-2026 · Arc — Launch blog post, Sep-16-2026 · Arc Docs — Consensus layer (Malachite, permissioned PoS) · Arc Docs — Gas and fees · Arc Docs — Stable fee design (base fee destination) · Arc Docs — Mainnet RPC endpoints · Tempo Docs — Transaction fees · Tempo — FAQ (mainnet date and validators) · Tempo — Ondo's USDY on Tempo, Aug-24-2026 · Visa Investor Relations — Validator node on Tempo, Apr-14-2026 · MoneyGram — Remittance validator on Tempo, May-20-2026 · DefiLlama — Arc chain (TVL and protocols) · DefiLlama — Tempo chain (TVL and protocols) · arc.etherscan.io — Cumulative transactions and Arc mainnet gas · Cryptopolitan — Arc day one DEX volume according to Dune (adam_tehc) · Crypto Briefing — Stablecoin transfers on Tempo, Sep-2-2026