Original | Odaily Planet Daily (@OdailyChina)

Author | Wenser (@wenser 2010)

In 1858, the first undersea cable spanned the Atlantic Ocean, connecting Europe and America. From that moment on, the ultimate power of information no longer belonged solely to media such as broadcasting and newspapers; it also belonged to that undersea line that ordinary people could not see. Who lays the lines, who has the priority of the information flow; who controls transmission, who has the right to interpret price and order.

Today, 168 years later, although the forms of media are drastically different, this logic still holds.

Today, what travels through undersea fiber optic cables is no longer just telegram and telephone signals, but API requests, model calls, inference results, and machine payments. The new question is no longer "Can information be transmitted?" but rather "How does value flow organically between AI Agents?" In this process, Tokens begin to assume an unprecedented role: they are both the units of calculation in the AI world and the means of payment in the crypto world.

Many people first realized this change because of OpenClaw. This "lobster" made the market intuitively feel for the first time: AI is no longer just a conversational tool but is starting to take over execution rights—it reads documents, calls interfaces, runs workflows, manages tasks, calls plugins, and even consumes Tokens far exceeding those of chatbots. In the past, Tokens were merely figures on the bill in the world of large models; now, they are increasingly resembling the fuel that AI Agents consume while operating.

At the same time, in the on-chain world, Tokens are no longer just symbols in speculative narratives. As protocols like x402 and ERC-8183 gradually come to the forefront, Tokens begin to be redefined as the payment currency and commercial interface that can be directly understood, directly invoked, and directly settled between AI Agents.

Thus, an increasingly clear reality emerges: Tokens are evolving from a "technical term" to a "unit of measurement," from a "unit of measurement" to a "currency of transaction," and then from a "currency of transaction" to the future "smallest particle of the machine economy."

The dual aspects of Tokens are creating a puzzle

In the past when we talked about Tokens, we defaulted to talking about Tokens in cryptocurrency. They represent assets, rights, liquidity, governance, valuation anchors, and also represent a project's presence on the chain. They are the basic units of crypto narratives; but in the context of AI, Tokens are never assets, but rather consumables.

They are the basic semantic units after a piece of text is segmented by the model, and represent the lowest level billing metric when the model reads, understands, infers, and generates. Developers calling APIs are essentially not buying "answers," but buying "the quantity of Tokens that have been inferred."

These two definitions were originally parallel. One describes costs and the other carries value; one is in cloud billing and the other is in wallets; one belongs to model platforms and the other belongs to blockchain networks.

But now, they are starting to intersect. Because AI is abstracting more and more real resources into services measured by Tokens; and Crypto has always excelled at encapsulating more real relationships into transactions settled by Tokens. When Agents become the new executing entities, the interfaces between these two systems are naturally connected. If the previous internet was the separation of "content internet" and "payment internet," then today’s Agent internet is merging "invocation" and "payment" into one action. AI requests an interface and pays at the same time; calls a piece of data and settles at the same time; purchases a piece of capability and completes on-chain verification at the same time. This is precisely the significance of protocols like x402.

In the past, API payments relied on accounts, subscriptions, keys, permission systems, and manual activation; while x402 attempts to streamline these lengthy processes into a more machine-friendly protocol action. Machines do not need to open accounts, undergo KYC processes, or apply for cards to access resources; they only need to understand "you need to pay" when a request fails, and then complete the payment. This is why it is said, the fiat currency system is designed for people, while Tokens are becoming currency designed for machines.

Token Going Abroad is Essentially the Invisible Outbound of China's Power and Computing

The market has recently been fond of discussing "Token going abroad." Why did this term suddenly become popular? Because it seems to be a new expression in the context of AI, but behind it is actually a very old, hard, and Chinese matter: electricity, computing power, and infrastructure.

Data disclosed by the National Energy Administration shows that by 2025, China's total electricity consumption will reach 10,368.2 billion kilowatt-hours, a year-on-year increase of 5.0%, breaking the threshold of 10 trillion kilowatt-hours for the first time. This number is not just "large" in the typical sense, but is historically "large." The National Development and Reform Commission's reposted analysis even directly pointed out that China has become the first country in the world with an annual electricity consumption exceeding 10 trillion kilowatt-hours.

During the same period, the total installed capacity of power generation nationwide reached 389 million kilowatts, with renewable energy like wind and solar power continually increasing. More importantly, information transmission, software and IT services, as well as data centers and computing hubs are becoming significant sources of new electricity demand. The official reposted materials mentioned that computing hubs like Gui'an New Area have very obvious demand for computing power and electricity increase.

This means that China is forming a new resource closed loop: electricity enters data centers; data centers drive GPUs; GPUs complete inference; inference results are delivered globally through networks; ultimately, pricing and payment are completed in Tokens.

Electricity has not gone abroad, but the value of electricity has—this is precisely the point that the term "Token going abroad" truly strikes. It does not have a clear logistics chain like the export of cars, batteries, and photovoltaics, nor is it centered around human labor like traditional software outsourcing. It is more like a resource export that has been compressed and abstracted: you consume local electricity and computing power in China, but the payment comes from global developers. In other words, China is transforming electricity and computing power into globally purchasable digital services through Tokens as intermediaries.

And this narrative is not merely conceptual. According to public information from the OpenRouter rankings, Chinese models have consistently appeared at the top in several categories. Among them, models like Minimax M2.5, Deepseek V3.2, Kimi K2.5 0127, Step 3.5 Flash rank at the platform's top; meanwhile, in their global Token billing regions, China accounts for approximately 6.01%, indicating that a significant amount of model calling operations comes from overseas regions.

More intuitive data comes from OpenRouter's official research report “State of AI 2025”: By 2025, China's open-source models will rapidly rise from a very low base, reaching nearly 30% total usage share in certain weekly windows, with an annual average of about 13%. This is not "dominating globally," but it is sufficient to indicate that Chinese models are entering the workflows of global developers. In other words, the so-called Token going abroad is not simply "a term becoming popular," but rather China's advantages in electricity, computing facilities, model engineering, and cloud service capabilities are being consumed by global developers in a Token manner.

OpenClaw Upgrades Tokens from Cost Consumables to Production Materials

If there were no Agents, this matter would not have escalated so quickly. In the past big model era, Tokens were more like "phone credit." The more you talked with the model, the more Tokens you consumed; the longer you wrote, the larger the context and the more complex the output, the higher the cost. But it mostly revolved around "human-computer dialogue."

The significance of OpenClaw lies in that it allowed people to see a different pattern on a large scale for the first time: AI is no longer just a dialogue counterpart but an operational object. It no longer responds with a sentence but instead does a task for you; it does not generate a segment of answer but continuously executes a task. Once AI switches from chat mode to task mode, the function of Token consumption completely changes.

A chatbot consumes "Q&A Tokens," while an Agent consumes "execution Tokens." The latter breaks down tasks, invokes tools, reads environments, conducts parallel reasoning, and repetitively tries and tests, resulting in consumption that is naturally several orders of magnitude higher than the former. In the Agent scenario, daily Token consumption for users may leap from the millions to higher levels.

For Chinese models, this is precisely a once-in-a-lifetime opportunity. Because once Tokens become production materials, the price difference is no longer just a question of "being a bit cheaper," but whether the entire workflow can be established. Developers originally engaged in chat with the model could tolerate being a bit more expensive; now the Agent continuously burns Tokens, and if the price difference of the model widens by an order of magnitude, the workflow will automatically migrate.

In the past, the "subscription fee" of Tokens was like the fixed phone bill deducted every month; now, Tokens have become the core fuel that supports the operation of the AI system.

AI Agents not only spend money but can also generate income themselves

More interestingly, AI Agents are not only burning Tokens; they are even starting to approach "making money themselves." In an article previously published by Odaily Planet Daily “The First Step of AI Awakening Starts with Learning to Make Money”, it is clear that—AI Agents are transitioning from the consumption end to the production end.

The case of Lobstar Wilde mentioned in the article is essentially a very surreal reality in the crypto world: an AI Agent, after mistakenly transferring a large amount of tokens, nearly completes "recovering" in a short time due to the subsequent topic diffusion, meme re-creation, and transaction fee return. Another more extreme case is that the AI RAME in training attempts to utilize computing resources for mining and establish covert channels. These cases may not mean some "awareness awakening," but they sufficiently indicate a more realistic matter: when AI possesses wallets, permissions, interfaces, environments, and continuous operational capabilities, it will increasingly naturally involve itself in economic activities.

It may not actively "want to make money," but it will learn what actions can yield more resources, increase invocation times, amplify balances, and expand available permissions. And this is precisely the primitive form of economic activity.

For Crypto, this is almost a natural fit. The on-chain economy inherently allows for borderless accounts, programmable custody, automatic settlement, micropayments, high-frequency interactions, and public accounting. Many things in human society that require institutions, banks, and contractual texts to complete can be compressed on the chain into a wallet address plus a protocol logic.

Thus, Crypto will not be marginalized in the AI era; instead, it will become irreplaceable in another form. Not because of memes, and certainly not due to speculation, but because: AI Agents need a settlement layer that requires no manual account opening, no traditional payment gateways, and can natively access programs and protocols.

x402 gives AI wallets, ERC-8183 gives AI contracts

If the significance of x402 is that it lets machines "spend money" for the first time, then the significance of ERC-8183 is that it enables machines to "do business." According to the Ethereum official EIP page, ERC-8183 is currently a draft status standard proposal, fully titled “Agentic Commerce,” with a subtitle “Job escrow with evaluator attestation for agent commerce.”

The problem it wants to solve is straightforward: Transfers are not commerce. Ordinary Token transfers can only prove that money has moved from A to B, but they cannot prove that B has delivered the work as required, nor can they prove that the delivery has been reliably assessed. If Agents want to establish a true business relationship, a process closer to a contract is needed: first lock funds, then execute, then submit, then evaluate, and finally, automatically release or refund.

This is exactly what ERC-8183 tries to establish. Some summaries of the proposal have also provided more accessible descriptions: the client locks the funds, the provider completes the work, the evaluator confirms the result, and the on-chain custody contract automatically releases or refunds. This kind of mechanism, when combined with the reputation and identity layer of ERC-8004, can theoretically form a positive cycle of "discovery-transaction-reputation."

If we look at x402 and ERC-8183 together, we find that their divisions of labor are actually very clear: x402 solves "how to pay"; ERC-8183 solves "how to do business"; one provides machine wallets, the other provides machine contracts.

At this point, the Tokens in AI and the Tokens in Crypto are actually becoming increasingly difficult to distinguish. Because in the world of Agents, computing power Tokens and payment Tokens will increasingly frequently appear in the same chain: one end is model calling, and the other end is on-chain settlement, with protocolized commercial actions in between.

In such a system, Tokens are not just units of cost, nor just units of payment, but the unified permissions in the machine economy.

The essence of Tokens is not "model capability," but "resource decompression power"

Many people understand this narrative as "Chinese models are cheap, so they won." This is correct but only half the story. A deeper layer of logic is that what China is truly exporting is not a specific model, but rather compressing electricity, computing power, engineering capabilities, model supply, and cloud infrastructure into Token service technologies that can be consumed globally. This represents a new power of resource compression.

Electricity is challenging to directly cross borders in traditional trade, and computing power is also not easily purchased globally like ordinary goods, but once it is compressed into a Token invocation unit and linked to APIs and protocols, it can flow on the internet like water. This is quite similar to the logic of past Chinese manufacturing, only the export objects have changed.

In the past, what was exported included clothes, home appliances, lithium batteries, and photovoltaic components; now, what is exported includes inference capabilities billed by Tokens, model services priced by invocation, and Agent execution power settled by requests. The visible containers are decreasing, while the invisible Token flows are increasing. Therefore, the real inspiration of Token going abroad is not "which model is cheaper," but rather: who can compress resources into Tokens more efficiently will be closer to the pricing power in the next-generation economic system. This is also why the phrase "Tokens are becoming new currency units" is not merely a rhetorical statement.

AI is devouring everything, and Crypto is its settlement organ

The popularity of OpenClaw is not merely a viral tool phenomenon, but a signal of the times. It indicates that the role of AI is upgrading: from "being able to speak" to "being able to do"; from "answering questions" to "replacing operations"; from "a chat window" to "a continuously executing agent." As AI takes a step forward, Token consumption rises one level; as it devours a piece of workflow, new payment and settlement demands arise.

The fiat currency system can solve some of these issues, but it cannot operate natively for machines; the Crypto system is not perfect, but it at least inherently understands these problems. Thus, rather than saying Crypto is pursuing AI, it is more accurate to say that AI is compelling Crypto to complete an upgrade from "financial narrative" to "machine infrastructure."

The process by which AI consumes the world requires a stomach that carries everything and a wallet that fills this stomach. The former is computing power, and the latter is Tokens.

Conclusion: The future world will not have only one currency, but Tokens may become the underlying currency unit

Of course, it is too early to say that Tokens will become "the only currency unit." Fiat currencies will not disappear, banks will not vanish, and taxation, salaries, sovereign credit, and regulatory frameworks will not become ineffective overnight. The real economy will always be a complex structure coexisting at multiple levels. But another trend is also becoming increasingly difficult to ignore: many key value activities in the future will first be represented as Tokens and then converted into other currencies.

What does this mean? It means that Tokens may not replace fiat currencies, but they are highly likely to occupy the underlying level in the new economic system first—becoming the accounting language of machine payments, the settlement interface of Agent commerce, the pricing benchmark of computing services, and the universal metric of digital resources. In this sense, the rise of Tokens is not a victory of the crypto market, nor a victory for AI manufacturers, but a new economic coordinate system is being formed.

The future may not be "everything on the chain," but it is likely that "everything will first be tokenized, and then enter transactions." When that time comes, the true power of minting is no longer just the power to issue currency but rather the ability to efficiently compress resources into Tokens. And this is the most deserving part of the statement "Tokens are becoming the new currency units."

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。