Stripe and Paradigm launch Tempo’s mainnet and the Machine Payment Protocol, targeting high-speed, stablecoin-based payments for AI agents and global enterprises.
Summary
- Stripe- and Paradigm-incubated Tempo launches its payments-focused Layer 1 mainnet with no native gas token, settling fees in major stablecoins instead.
- The new Machine Payment Protocol aims to standardize machine-to-machine and AI agent payments, positioning Tempo as settlement rail for an “AI-native” economy.
- Tempo targets cross-border payments and B2B flows with ISO 20022 compliance, EVM compatibility, and early interest from firms like Klarna, Visa, Nubank, and Shopify.
After months of anticipation following a public testnet deployment in December 2025, Tempo — the payments-focused Layer 1 blockchain incubated by payments giant Stripe and crypto venture firm Paradigm — officially launched its mainnet on Wednesday. The announcement, made via official channels, was accompanied by the simultaneous release of the Machine Payment Protocol (MPP), an open standard for autonomous machine-to-machine transactions co-developed by Stripe and Tempo. The dual launch marks one of the most significant entries of a traditional fintech heavyweight into blockchain infrastructure to date.
Tempo has been positioned from inception as a purpose-built alternative to general-purpose chains like Ethereum or Solana, targeting the specific demands of high-frequency, real-world payments. According to Paradigm’s own documentation, the chain is designed to process tens of thousands of transactions per second with sub-second deterministic finality — performance comparable to, or exceeding, traditional card networks. Unlike most blockchains, Tempo does not require a native token to pay gas fees; instead, users can settle transaction costs in any major stablecoin via an integrated AMM, using the TIP-20 standard. No token is being issued at launch, with the team citing the need for greater regulatory clarity before any such move.
The MPP’s release is arguably the more forward-looking element of Wednesday’s announcement. Developed jointly with Stripe, the protocol establishes an open standard for payments between machines — software agents, AI systems, and automated processes — without requiring human intermediaries. As AI agents increasingly execute real-world commercial tasks autonomously, proponents argue that a dedicated payment rail becomes essential infrastructure. Tempo’s architecture was explicitly designed with this use case in mind, with Stripe’s CEO previously describing the chain as a “decentralized, internet-scale SWIFT” for next-generation settlement.
The practical scope of Tempo’s ambitions is substantial. Stripe processed $1.9 trillion in total payment volume in 2025, a 34% year-on-year increase, while global stablecoin volumes doubled over the same period to $400 billion, with 60% now attributable to B2B activity. Tempo targets the $190 trillion annual cross-border payment market, where traditional correspondent banking can impose settlement delays of one to three days and unpredictable fees. The chain’s ISO 20022 compliance — the international financial messaging standard used by banks — is designed to allow enterprises to integrate with existing reconciliation systems without wholesale infrastructure overhauls.
Early ecosystem commitments have been notable. Klarna announced plans to launch a stablecoin on Tempo’s mainnet, while Visa, Nubank, and Shopify were cited among early adopters during the testnet phase. Developers can build on Tempo through public RPC endpoints, with the chain’s EVM compatibility lowering the barrier for teams already familiar with Ethereum tooling.
The mainnet launch arrives at a moment of acute market turbulence, with crypto and risk assets broadly under pressure from geopolitical tensions and resurgent inflation. Yet for Tempo, the timing may be immaterial — its proposition is structural rather than speculative, betting that the next wave of blockchain adoption will be driven not by token appreciation, but by settlement infrastructure that actually works at scale.
