The global payment system is undergoing structural reconstruction. The explosive growth of stablecoin scale and the rise of the AI agent economy have jointly created an urgent demand for next-generation payment infrastructure.
When executing autonomous tasks, the payment behavior of AI agents (Autonomous AI Agents) is fundamentally different from traditional human payments. The following five core needs constitute the basic requirements of the AI agent economy for payment infrastructure:
Traditional SWIFT payment networks and general-purpose blockchains have difficulty fully meeting the aforementioned payment needs of the AI agent economy. Thus, Tempo emerged.
II. Tempo: A Blockchain Built for the AI Era
As a payment-native blockchain launched by Commonware, Tempo uses Simplex BFT pipelined consensus to achieve sub-second finality, ensures payment priority through dedicated block space and a stablecoin-native Gas mechanism, and provides end-to-end payment capabilities without human intervention for AI agents through the MPP protocol.
III. Tempo Blockchain Technical Architecture
3.1 Overall Architecture Overview
Tempo adopts a specialized Layer-1 architecture. Its design philosophy is "payment-first"—technical decisions at every layer of the chain are aimed at optimizing payment scenarios, rather than the general-purpose design of universal smart contract platforms.
3.2 Simplex BFT Pipelined Consensus
Tempo's consensus layer is based on the Simplex BFT protocol (ePrint 2023/463). Through a pipelined design, this protocol converges the confirmation delay per round to a single network round-trip time (1Δ).
Three-Phase Consensus Process
A single round of Simplex BFT consensus consists of three sequential phases:
Timing Comparison: Traditional BFT vs. Simplex Pipeline
The figure below shows the latency difference between traditional three-phase BFT and the Simplex pipeline. The vertical axis is the consensus round, and the horizontal axis is the network time step (Δ).
Key Performance Improvement: In pipeline mode, the Propose phase of B2 overlaps with the Vote phase of B1. Each round only needs to wait 1Δ to proceed to the proposal of the next block, whereas traditional BFT requires a full 3Δ of serial waiting per round.
View-Change Optimization
View-Change is triggered in two situations: (1) The current Leader fails to broadcast a valid proposal within the specified timeout period; (2) A node detects abnormal Leader behavior (such as duplicate proposals or illegal message formats).
3.3 BLS Aggregate Signatures
Uses the BLS (Boneh-Lynn-Shacham) scheme to aggregate N validator signatures into a single signature, requiring only two elliptic curve pairing operations for verification, significantly reducing bandwidth and computational overhead. This is particularly important for high-frequency micro-payment scenarios, effectively lowering the computational and bandwidth costs per transaction.
BLS Signature Principle
Aggregate Signature Process Visualization
3.4 Parallel Transaction Execution Mechanism
Tempo's parallel transaction execution capability comes from two technically documented designs:
1. EIP-2718 Custom Transaction Type (Transaction Type 0x76)
Tempo's defined Crypto-Native Transaction format extends three types of native capabilities on top of standard EVM transactions:
- Batch Execution (Batch): Atomically execute multiple instructions within a single transaction.
- Scheduled Execution (Scheduled): Specify execution to be triggered in a future block.
- Parallel Execution (Parallel): Declare no state dependencies, allowing concurrent processing with other transactions.
2. Expiring Nonce System
Traditional EVM's strictly increasing Nonce forces all transactions from the same account to be executed serially. Tempo changes the Nonce to an "effective block range," requiring only that the Nonce is unique within its validity period. Multiple independent transactions from the same account can be submitted and executed concurrently, eliminating account-level serial bottlenecks.
3. Dedicated Payment Lanes
Payment Lanes are block space specifically reserved at the protocol level for TIP-20 payment transactions on Tempo. Unlike Ethereum where all transactions compete for the same gas pool, Tempo splits the block gas budget into multiple independent lanes, ensuring payment transactions are not interfered with by "noisy neighbors" like DeFi operations, NFT minting, or high-frequency contract calls.
Block Gas Partition Structure
The Tempo block header carries independent gas limit fields, dividing the 500M total gas budget into three non-interfering regions:
3.5 Stablecoin-Native Design
Tempo treats stablecoins as first-class citizens within the protocol, redesigning the entire stack—from Gas fees, on-chain exchange, to Token standards—with stablecoins at the core.
IV. Machine Payments Protocol (MPP)
4.1 Protocol Positioning and Core Concept
MPP (Machine Payments Protocol) is an open payment standard jointly designed by Stripe and Tempo, referred to by the industry as the "OAuth for payments." Its core goal is to provide standardized, human-intervention-free payment capabilities for autonomous AI agents.
4.2 MPP Complete Interaction Flow
JWT Payload Structure
4.3 Session Mechanism
The Session mechanism is one of the core innovations of the MPP protocol, solving the payment efficiency problem when AI agents continuously consume resources over long periods:
This design means that during long-running tasks, there's no need to trigger on-chain confirmation for every interaction, significantly improving payment efficiency.
4.4 Cross-Rail Payment Routing
The core design of MPP is to completely decouple the protocol from the payment rail. The core layer only defines the HTTP challenge-response flow, error handling, and security model, without binding to any specific payment network. Therefore, adding a new payment method only requires registering a method identifier and publishing the corresponding Schema and validation logic, without changing the protocol itself. During payment, the agent does not need to care about the underlying rail; the server declares acceptable methods in the 402 response, and the client matches accordingly. This is the key difference between MPP and single-chain or single-network solutions.
Payment Rails Currently Supported by MPP
V. Application Scenario Analysis
Scenario 1: Cross-Border Corporate Payments
Traditional cross-border payments typically involve multiple steps: the paying bank, the SWIFT messaging network, correspondent banks, and the receiving bank. This often takes 3 to 5 business days, with fees usually between 0.5% and 3%, and does not support real-time processing on weekends and holidays.
In contrast, Tempo aims to provide an alternative path: if both payer and payee use stablecoins for settlement, according to current testnet design goals, a USDC-to-USDC cross-border payment could theoretically be completed in about 0.5 seconds, with a single transaction fee of approximately $0.001.
Scenario 2: 7×24 Hour Settlement for Tokenized Deposits
Tokenized deposits are digital financial assets representing bank deposit liabilities on a blockchain. These assets face a practical obstacle: the Fed's Fedwire has fixed operating hours and cannot handle settlements on non-business days or at night.
However, blockchains can naturally support 7×24 hour, year-round operation. Furthermore, Tempo's built-in exchange module can support protocol-level conversions between different tokenized deposits, making round-the-clock settlement possible.
Scenario 3: High-Frequency, Low-Value Automated Payments
Credit card processing fees typically include a fixed fee of about $0.20 per transaction plus a percentage fee of 1.5% to 3%, making transactions below $1 commercially unviable—this is the fundamental reason for the long-standing gap in the "micropayment" market. With a target transaction fee design of approximately $0.001, Tempo makes the following scenarios commercially feasible for the first time:
Scenario 4: Autonomous Payments by AI Agents
As AI agents are increasingly used to execute complex commercial tasks (booking resources, procuring materials, calling external services), these agents generate genuine payment needs. Tempo's EVM-compatible architecture and dedicated payment interfaces enable agents to autonomously trigger payments via smart contracts, without requiring manual approval for each transaction.
VI. Competitive Landscape Analysis
The payment-specific chain赛道 witnessed intensive entry in 2025–2026. This chapter provides a horizontal comparison of three types of competitors from a technical architecture perspective.
6.1 Payment-Specific Chains: Tempo vs. Circle Arc vs. Stable
All three are payment-specific L1s, but their underlying technical routes differ significantly. The following breaks down their respective technical choices from three dimensions: consensus engine, fee mechanism, and core architectural innovation.
Competitive Positioning Matrix
The three chains are highly convergent in performance metrics; the real differentiation lies in target customers, stablecoin binding strategy, core bets, and known risks.
6.2 Comparison with General-Purpose Blockchains: Ethereum L2 and Solana
Ethereum L2 and Solana are two types of general-purpose chains currently widely used in payment scenarios. The core gap compared to payment-specific chains is reflected in the following dimensions:
VII. Conclusion
The value proposition of payment-specific chains has never been about being "faster" than Ethereum or "cheaper" than Solana, but rather about whether they can internalize payment semantics as design constraints of the protocol itself.
The core judgment of Tempo and MPP is: when handling payment scenarios, general-purpose blockchains are not lacking in functionality, but are wrong in their level of abstraction—they treat "asset transfer" as the entirety of payment, but overlook authorization, sessions, routing, and reconciliation, elements that have been deeply engineered in traditional finance.
The AI agent economy has injected a new sense of urgency into this赛道. When software agents begin to代替 humans in completing economic behaviors like procurement, subscriptions, and service calls, the authorization model of the traditional payment system—built on the real-name authentication and manual confirmation of human subjects—will face systemic structural mismatch. The MPP protocol attempts to solve precisely this layer of "agent sovereignty" problem: who is qualified to initiate payments, within what scope, for how long, and how can it be revoked. This is highly isomorphic to the logic by which OAuth solves API authorization.
However, it must be pointed out that the large-scale adoption of autonomous payments by AI agents前提是 the legal status, liability attribution, and anti-money laundering compliance path for agent identity are clarified. The challenges facing Tempo are structural, not merely executional. First, regulatory uncertainty remains a core variable: a stablecoin-native design means Tempo must engage in direct dialogue with monetary regulators worldwide, rather than hiding behind the narrative of "neutral infrastructure"; second, the tension with EVM compatibility has not been resolved—abandoning EVM could换来 a cleaner design space, but also means giving up years of accumulated developer inertia and toolchain support from the Ethereum ecosystem; third, the partnership with Stripe赋予 the MPP protocol rare commercial endorsement, but this strong dependency is also a source of fragility. There is an inherent tension between the protocol's openness and the利益 boundaries of commercial partners, which requires long-term observation.
For industry practitioners, the most值得研究的 aspect of Tempo/MPP might not be whether it will ultimately become the "winning payment public chain," but rather the question it raises itself: After on-chain payment infrastructure enters the era of professional division of labor, how should the competitiveness of protocol design ultimately be evaluated? Beyond performance benchmarks, the precision of payment semantic expression, plug-and-play compliance, and the agent authorization model might be the true differentiators of the next generation of payment infrastructure.
References
- Tempo Official Website: https://tempo.xyz
- Tempo Mainnet Launch Blog: https://tempo.xyz/blog/mainnet/
- MPP Protocol Technical Specification: https://docs.tempo.xyz/mpp
- Fortune: Stripe-backed Tempo releases AI payments protocol (2026.03.18)
- The Block: Tempo Mainnet goes live with Machine Payments Protocol for agents
- Privy Blog: Building on Privy with Tempo's Machine Payments Protocol (MPP)
- Medium (jrodthoughts): The Architecture of Autonomous Wealth — Inside Tempo's MPP
- McKinsey & Artemis Analytics: 2025 Stablecoins in Payments Report
- CoinGecko Stablecoins Market Data
- DeFiLlama On-chain Stablecoins Data
Связанные с этим вопросы
QWhat are the five core requirements of AI agent economy for payment infrastructure as mentioned in the article?
AThe article states that AI agents have five core requirements for payment infrastructure: 1) High-frequency micropayments capability, 2) Sub-second finality for transaction confirmation, 3) Autonomous payment execution without human intervention, 4) Predictable and ultra-low transaction fees, and 5) Robust session management for continuous resource consumption.
QHow does Tempo's Simplex BFT consensus achieve sub-second finality compared to traditional BFT protocols?
ATempo's Simplex BFT uses a pipelined design where the Propose phase for the next block (B2) overlaps with the Vote phase of the current block (B1). This reduces the per-block confirmation latency to just one network round-trip time (1Δ), whereas traditional three-phase BFT requires a full 3Δ of serial waiting, making it significantly slower.
QWhat is the primary innovation of the MPP (Machine Payments Protocol) Session mechanism?
AThe primary innovation of the MPP Session mechanism is that it allows an AI agent to pre-authorize a spending limit for a specific service and duration. This creates a session where the service can bill the agent incrementally without requiring a separate on-chain transaction confirmation for each micro-payment, drastically improving payment efficiency for long-running tasks.
QHow does Tempo's 'Payment Lanes' feature address the 'noisy neighbor' problem in blockchain transactions?
ATempo's 'Payment Lanes' are dedicated portions of block space specifically reserved for TIP-20 payment transactions at the protocol level. Unlike Ethereum where all transactions compete for a single gas pool, Tempo partitions the total block gas budget into multiple independent channels. This ensures that payment transactions are not delayed or impacted by other high-gas activities like DeFi operations, NFT mints, or smart contract calls.
QAccording to the competitive analysis, what is a key differentiator between Tempo and other payment-specific chains like Circle Arc and Stable?
AA key differentiator, as shown in the competitive positioning matrix, is Tempo's target customer focus. While Circle Arc targets 'Institutional & Merchant Services' and Stable focuses on 'Consumers & Retail Payments', Tempo is positioned for 'Autonomous Agents & B2B Microservices'. Its core bet is on enabling machine-to-machine (M2M) payments and AI agent economies, which is a distinct market segment.







极速>



















