Short, dated observations. I write these down when a number or a claim changes, so the entities above can be corrected instead of quietly rotting.
Arc has been the loudest stablechain of the past year and it still has no public mainnet; the date on the calendar is 16 September 2026, and every Arc figure quoted today is testnet or private mainnet. Tempo, which said much less, has been live since 18 March 2026. When reading this category, the useful question is not who has the better architecture diagram but who has validators producing blocks.
Stable collected more than two billion dollars in pre-deposits before launching in December 2025. Checking DefiLlama today it shows roughly thirty-four million in chain TVL. I do not know whether those deposits ever migrated on chain or whether coverage is incomplete, so I have not written the two billion into the entity. A launch-week number and a current number are different claims and should never be merged.
The headline transaction counts are real but they measure facilitator activity, and both Chainalysis and the trackers warn that tests, infrastructure traffic and self-directed calls are included. Artemis put roughly half of March 2026 volume in that bucket. The honest version of the x402 story is that the plumbing works and the demand is still being manufactured.
tempo.xyz/developers for benchmark numbers and the fee AMM
circle.com/pressroom and arc.io for Arc testnet and mainnet dates
plasma.org/docs for PlasmaBFT and the paymaster design
docs.stable.xyz for the USDT0 gas model
x402.org and docs.x402.org for the facilitator spec
mpp.dev and stripe.com/blog/machine-payments-protocol
chainalysis.com for independent adoption measurement
defillama.com for TVL and on-chain stablecoin supply