Solana Launches 'Payment Channels' for AI Payments, Competing with Ethereum's Next-Gen Infrastructure

[Block Media Reporter Ham Ji-hyun] Solana has launched a new payment technology called 'Payment Channels' aimed at large-scale micropayments for artificial intelligence (AI) agents. Instead of approving and settling payments on the blockchain every time an AI agent calls APIs, it sets an initial spending limit and settles the actual usage amount all at once at the end.

As Ethereum focuses on building an infrastructure for the 'agentic economy' centered around x402 and ERC-8004 (a standard for on-chain identity of AI agents), Solana is also accelerating its entry into the AI agent market by emphasizing ultra-fast and low-cost payments.

On the 3rd of this month (local time), the Solana Foundation unveiled the 'Payment Channels' for more efficient AI agent payments.

The Solana Foundation explained, "AI does not wait at the payment screen," stating that a different structure is needed for AI agents to engage in economic activities autonomously compared to traditional internet payments.

Currently, in the AI agent payment method, payments must be approved and settled every time a paid API or tool is called. While this is not a significant issue for one or two payments, it becomes problematic when an AI calls hundreds of low-cost APIs in a short time or uses services where costs vary based on usage.

A typical example is large language models (LLMs) that charge based on token usage or services where costs are determined by data processing volume and computing time. If each token generated by the AI incurs a cost or if hundreds of micropayments occur in a short time, settling each one on-chain leads to accumulated costs and delays.

Payment Channels solve this in a manner similar to a bar tab or prepaid meter.

First, the user deposits the maximum spending amount allowed for the AI agent into an on-chain escrow. The funds are stored in a Solana program, maintaining a non-custodial structure.

The AI agent does not generate on-chain transactions every time it uses a service. Instead, it accumulates the usage amount off-chain using signed messages. Once the work is completed, the actual usage amount is recorded in a single on-chain transaction and paid to the service provider. Any unused deposit is returned to the original wallet.

In other words, even if the AI calls 1,000 APIs, the blockchain can compress on-chain processing into just the tasks of opening the channel and final settlement.

Another feature is that individual calls do not incur blockchain transaction fees.

This also differs from the existing prepaid credit method. Prepaid credits often involve a custodial structure where the user transfers funds to the service provider first, and the provider manages the balance in their database. In contrast, in Payment Channels, the funds are held in the on-chain program until settlement.

Ultimately, Payment Channels allow for the settlement of millions of 'machine-to-machine (M2M)' micropayments that may occur during the AI's data purchases and API and computing usage without recording each one on the blockchain, only settling the final results.

The Solana Foundation emphasized that this can simultaneously address issues such as repetitive payment approvals by humans, settlement costs and delays incurred with each payment, and custodial issues with prepaid credits.

During the testing phase, they connected 100,000 unique wallets to the Payment Channel proxy, processing over 1 million payment requests per second.

This translates to a throughput capable of supporting over 80 billion payments in a day. The estimated processing cost per payment, according to Solana, is just $0.000000000776. However, this result is possible because the actual on-chain work has been reduced to channel opening and final settlement.

Solana has made the program and technical specifications of Payment Channels open-source, allowing anyone to reproduce the same benchmarks directly.

Real service applications have also begun.

Alibaba Cloud started supporting related API endpoints simultaneously with the launch of Payment Channels. AI agents can repeatedly use Alibaba Cloud's APIs after approving a spending limit once, without needing to open a separate account or approve payments each time they call an API.

It is also noteworthy that rather than creating a new payment standard for readers, it is designed to connect with existing AI payment standards. Payment Channels currently support both x402 and MPP (Machine Payments Protocol) used in Solana.

Solana's 'Payment Performance' vs. Ethereum's 'Trust Infrastructure'

Ethereum is also viewing AI agents as the next generation of blockchain users and is working on building related infrastructure.

However, there are differences in the approaches that have emerged so far.

While Solana has delved into the question of "how quickly and cheaply can AI handle large-scale payments" through this Payment Channel, the Ethereum ecosystem is focusing on building an economic infrastructure that combines payment with the identity, reputation, and authority of AI agents.

Key technologies include x402 and ERC-8004.

x402 is an open payment standard that utilizes the HTTP status code '402 Payment Required' to enable AI agents to automatically pay with stablecoins while using APIs, data, content, and computing. ERC-8004 is an Ethereum standard for managing the identity, reputation, and verification information of AI agents on-chain.

By using smart accounts, spending limits, allowlists, and session keys can be set for AI agents. This allows programming the authority to spend only a certain amount on specific services while granting the AI a wallet. The goal is to create a foundation for AI agents to verify who they are, what their reputation is based on past performance, and whether the results can be trusted when seeking other AIs to delegate tasks.

Recently, the Ethereum-based wallet service MetaMask unveiled the 'MetaMask Agent Wallet,' designed to support the autonomous economic activities of AI agents by incorporating spending limits, allowlists, and risk profiles.

Currently, the AI strategies of the two ecosystems are showcasing different strengths.

Solana has directly targeted areas where usage is difficult to predict, such as AI inference or repeated service calls in the hundreds or thousands, through Payment Channels. It has focused on solving the cost and speed issues of high-frequency, low-value payments.

Rishin Sharma, head of AI growth at the Solana Foundation, stated in an interview last month that existing API services rely on subscription or prepaid credit models, and there is a need to change the billing system for AI agents. He said, "In an agentic environment, payment per request is much more suitable."

On the other hand, Ethereum is expanding its ecosystem in a direction that connects the identity, reputation, authority, and payment systems necessary for AI agents to become independent economic entities on-chain.

Shan Agrawal, Chief Business Officer of Coinbase, recently described the agentic payment market as "still in its early stages, perhaps just beginning to warm up," emphasizing that widespread payment acceptance and infrastructure to verify the identities of AI agents are needed for market expansion.

Simplifying this, Solana focuses on "how AI can spend money faster and cheaper," while Ethereum emphasizes "how AI can operate as a trustworthy economic entity." The strategies of both sides are not entirely opposed. Open standards like x402 are used across multiple chains, indicating the potential for the AI agent economy to develop in a multi-chain manner.

Ultimately, the core of the competition lies not in who owns a specific payment standard but in which blockchain AI agents choose when buying and selling data, APIs, AI inferences, and computing resources.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.