Base MCP, The Next Step for x402

marsbitPublicado em 2026-05-28Última atualização em 2026-05-28

Resumo

Base has officially launched Base MCP, allowing users to connect their Base Account to AI Agents to perform actions like swaps, transfers, portfolio tracking, and transaction history queries through conversational commands. This move aligns with Base's strategic focus on AI, driven by the broader competition in the emerging Agent-to-Agent payment sector. The evolution of Agent payments has accelerated. In late 2024, the primary method involved insecure browser automation. By 2025, solutions like Coinbase's x402 (providing crypto wallets for Agents), Google's AP2, and Visa's token-based system emerged. x402 has since processed 176 million transactions totaling over $70 million, with a median value between $0.01 and $0.10. Stablecoins, particularly USDC, dominate these settlements due to their negligible transaction costs compared to traditional payment fees, which are prohibitive for micro-payments. Coinbase faces competition from Stripe, which has built a comparable infrastructure for Agent payments with its Tempo blockchain, Privy wallets, Bridge routing (acquired for $1.1B), and the recently launched MPP protocol. Both companies are now competing at the application layer. The core reason AI is central to Base's strategy is to expand the scenarios for Agent payments, ensuring more transactions occur on its network. By securing a dominant position and scale advantage in this nascent field, Coinbase aims to capture the future commercial potential of Agent-driven payments. T...

Yesterday, Base officially launched Base MCP. By connecting Base Account to AI Agent through Base MCP, you can use colloquial language, like chatting, to have the Agent perform operations such as Swap, transfers, position tracking, and transaction history queries.

Players familiar with Base know that the main focus on the Base chain right now is AI, so such an update from Base doesn't come as a surprise. Some players even anticipate new gameplay on the Base chain, similar to the AI meme coin $SHIT on Ethereum before, where they could directly use Base MCP to let an Agent participate in chain-based token launches through chat.

But if we step back from the perspective of a chain degen and look at the competition in Agent-to-Agent payments, we might find a new answer as to why AI has become the main focus for Base.

Rapidly Developing Agent Payments

Let's rewind to September 2024. Back then, if you wanted an AI Agent to complete a payment for something, humans basically had only one choice: use browser automation tools (like headless browsers such as Playwright, Selenium) to let the AI Agent simulate human actions and complete the checkout process on a webpage.

Since this required providing payment credentials (like full credit/debit card numbers, CVV, expiration dates, etc.) to the AI Agent, this sole option was not secure.

By May 2025, Coinbase launched x402, providing an AI Agent with a crypto wallet and solving this problem in a crypto-native way. But Coinbase wasn't the only one to recognize this as a potential market, and the solutions weren't limited to just crypto-native ones. In 2025, Google launched AP2, allowing users to authorize spending permissions to an Agent. Visa expanded its existing card payment channels, launching Visa Intelligent Commerce, which doesn't give the Agent sensitive information like credit card numbers or CVV, but instead provides the Agent with specific, limited tokens to complete payments.

Today, x402 has processed 176 million transactions from AI Agents, with a total transaction volume exceeding $70 million. This amount might not seem huge, but neither Coinbase nor the traditional giants are taking this emerging payment competition lightly:

- On January 22, 2026, Capital One, the sixth-largest bank in the US with $470 billion in assets, $330 billion in deposits, and the third-largest credit card issuer in the nation, announced the acquisition of Brex for $5.15 billion to enhance AI payment capabilities.

- In March 2026, Mastercard acquired the stablecoin infrastructure company BVNK for $1.8 billion.

- In February 2025, Stripe acquired the stablecoin payment platform Bridge for $1.1 billion.

While they haven't explicitly stated it, acquiring stablecoin-related companies is likely a move to prepare for the upcoming era of Agent payments. Stablecoins are indeed crucial for Agent payments.

Why Are Stablecoins Important for Agent Payments?

According to data statistics from Keyrock, the median transaction amount for Agent transactions processed on x402 so far is between $0.01 and $0.10, with 76% of transactions being under $0.30.

$0.30 is the most common flat fee per transaction in the US and many major markets. This fee is like a wall, making micropayments under $1 very uneconomical. For example, for a 3-cent API call, a $0.30 fee is 10 times the call cost; if an Agent pays with a credit card, the cumulative cost would be prohibitively high.

Blockchain solves this problem well. On Base, the transaction settlement cost is $0.0001. With this immense advantage, stablecoins have almost naturally won the competition against traditional payment giants in the realm of Agent payments.

Of the 176 million Agent transactions processed by x402, 98.6% were settled in USDC. Given the close relationship between Coinbase and Circle, it's fair to say Coinbase is also a big winner at the settlement layer.

But the settlement layer is just one layer in Agent payments. In the track of solving Agent payments via crypto-native means, Coinbase has a competitor—Stripe.

The Challenge from Stripe

This March, Stripe launched the Agent Payment Protocol MPP, which has brought Stripe's architecture map for Agent payments almost on par with Coinbase's.

- From the settlement layer: Coinbase has Base, Stripe has Tempo.

- From the wallet layer: Coinbase has Agent Wallet, Stripe has Privy.

- From the routing layer: Coinbase has built-in routing facilities, Stripe has Bridge, acquired for $1.1 billion.

- From the payment protocol: Coinbase has x402, Stripe has MPP.

Now let's return to the Base MCP mentioned at the beginning of the article. Since both competing sides now have these four layers of supporting infrastructure, the next battlefront is naturally the application layer.

This is the core reason why AI can become Base's main focus—Base needs to ensure that AI (at least in the cryptocurrency space) happens on Base. This is not actually about providing a perspective for degens on the Base chain, but about broadening the scenarios for Agent payments, enabling more Agents to conduct more transactions for more applications, thereby securing its leading position in the Agent payment track.

Once a dominant scale advantage is established, when Agent payments enter the commercial domain in the future, Coinbase stands to win even bigger.

Looking at the launch of Base MCP from this angle, one can sense that this is just a small step in Coinbase's grand ambition.

Perguntas relacionadas

QWhat is Base MCP and what key functionality does it enable for AI Agents?

ABase MCP is a protocol recently launched by Base. It allows users to connect their Base Account to an AI Agent. Once connected, users can use natural, conversational language to instruct the Agent to perform actions such as executing token swaps, making transfers, tracking portfolio positions, and querying transaction history.

QAccording to the article, why are stablecoins considered crucial for Agent payments?

AStablecoins are crucial for Agent payments because they enable micro-transactions economically. Traditional payment methods often have fixed fees around $0.30 per transaction, which makes small payments (e.g., a few cents for an API call) prohibitively expensive. In contrast, settlement costs on Base are around $0.0001, allowing stablecoins like USDC to facilitate vast numbers of small-value transactions efficiently, as evidenced by 98.6% of x402 transactions being settled in USDC.

QHow does the article contrast Coinbase's and Stripe's approaches to the Agent payment infrastructure?

AThe article contrasts their approaches by outlining the similar four-layer infrastructure both companies have built: Settlement Layer (Coinbase has Base, Stripe has Tempo), Wallet Layer (Coinbase has Agent Wallet, Stripe has Privy), Routing Layer (Coinbase has internal routing, Stripe acquired Bridge), and Payment Protocol Layer (Coinbase has x402, Stripe has MPP). This positions them as direct competitors in the Agent payment space.

QWhat strategic reason does the article suggest is behind Base making AI its main narrative?

AThe article suggests that Base's strategic focus on AI is not primarily for on-chain 'degen' activities but to expand the scenarios for Agent payments. By ensuring more AI activity (at least in the crypto space) happens on Base, Coinbase aims to generate more Agent-driven transactions across more applications. This builds a dominant scale advantage, securing Coinbase's leading position in the emerging Agent payment competition for future commercial applications.

QWhat does the median transaction amount on x402 indicate about the nature of Agent payments?

AThe median transaction amount on x402 falls between $0.01 and $0.10, with 76% of transactions being under $0.30. This data indicates that Agent payments are predominantly micro-transactions, involving very small sums of money per transaction.

Leituras Relacionadas

Circle Blacklists Zama Protocol Address, Freezing $12.6M In User Funds – Details

Circle has blacklisted a smart contract address for the Zama protocol's Confidential USDC (cUSDC), freezing approximately $12.6 million in user funds. The action, identified by on-chain investigator ZachXBT, targeted a publicly documented contract on Ethereum. Investigations suggest the frozen funds may be indirectly connected to ongoing controversy and legal issues surrounding Overnight Finance, an asset-management protocol. A wallet linked to Overnight Finance had deposited a large sum into the Zama contract amidst governance disputes and allegations of a possible rug pull. This move by Circle, reportedly made without prior warning to Zama, has reignited criticism over transparency and unilateral enforcement by centralized stablecoin issuers. It highlights concerns about how external legal and governance risks can impact unrelated users in interconnected DeFi systems. Circle has not provided an official explanation for the freeze.

bitcoinistHá 19m

Circle Blacklists Zama Protocol Address, Freezing $12.6M In User Funds – Details

bitcoinistHá 19m

Why More AI Agents Does Not Equal Higher Productivity?

Editor's Note: As AI Agents become cheaper and easier to use, a new constraint emerges: the cost isn't in launching more Agents, but in the human attention required to manage, judge, and integrate their outputs. This hidden cost is called the "orchestration tax." The article argues that a developer's cognitive bandwidth is the key bottleneck—a serial, non-parallelizable resource akin to a Global Interpreter Lock (GIL). While many Agents can run concurrently, their results ultimately require human judgment for review, conflict resolution, and final integration. Therefore, more Agents don't automatically mean higher productivity; they can simply create longer queues, lead to cognitive fatigue, and create the illusion of busyness without real output. The core solution is to design workflows around this scarce human attention. Key strategies include: scaling the number of Agents to match review capacity (not UI capacity), categorizing tasks (delegating independent ones, keeping complex judgment-heavy ones serial), batch reviewing results to minimize context-switching costs, automating verifiable checks to reserve human judgment for critical decisions, and protecting focused, uninterrupted thinking time. Ultimately, the critical skill is not launching many Agents, but architecting systems that respect the fundamental limit of human attention. Unpaid "orchestration tax" accumulates as both technical and cognitive debt, undermining system understanding and quality. True productivity comes from thoughtfully managing the single-threaded resource—your focus.

marsbitHá 1h

Why More AI Agents Does Not Equal Higher Productivity?

marsbitHá 1h

Three Years Later: Looking Back at My Predictions About ChatGPT in 2023

Three Years Later: Revisiting My 2023 Predictions on ChatGPT In March 2023, shortly after ChatGPT's launch, I made 20 predictions about its future. Now, in mid-2026, I've used AI agents to fact-check each one against the latest data. Overall, most major directional forecasts were correct, with only one outright error (incorrectly stating GPT-4 had 100 trillion parameters). Key successes included predicting that RAG and retrieval architectures would become the standard for handling knowledge and hallucinations, that natural language interfaces (LUI) would create a massive new industry layer beyond the models themselves, and that China would develop viable large language models, significantly closing the performance gap with Western counterparts within about three years. Predictions about the absence of mass unemployment, the rise of a new "robot network" for agent communication, and ChatGPT not possessing consciousness also held true in their core arguments. However, the "devil was in the details." Errors frequently involved specific numbers, timelines, or overlooking distributional effects. I tended to overestimate the speed of adoption (e.g., for agent networks) while underestimating the ultimate scale of capabilities or costs (e.g., AI winning IMO gold without tools, or the extreme capital required for frontier models). Other misjudgments included: underestimating how AI would reinforce, not dissolve, information filter bubbles; incorrectly assuming AI-generated content would easily circumvent copyright (it has instead triggered record-breaking settlements); and misidentifying where value would be captured (it accrued overwhelmingly to the compute layer, like Nvidia, not just the application or model layers). Key lessons from reviewing these predictions are: 1) Directional and mechanistic insights are far more reliable than precise numbers or absolute statements. 2) There's a consistent bias to overestimate short-term speed but underestimate long-term magnitude. 3) Errors often lie in missing distributional impacts within a generally correct aggregate trend. 4) Predictions phrased with nuance and caveats aged the best. 5) Some fundamental debates (e.g., on machine consciousness or the ultimate value chain) remain unresolved even after three years. This exercise is less about scoring the past and more about establishing rules for clearer thinking about the next three years of AI.

marsbitHá 7h

Three Years Later: Looking Back at My Predictions About ChatGPT in 2023

marsbitHá 7h

Three Years Later: Looking Back on My 2023 Predictions for ChatGPT

Looking Back After Three Years: Revisiting My 2023 Predictions on ChatGPT In March 2023, shortly after ChatGPT's debut and before GPT-4's release, I made over twenty predictions about AI's future based on limited information and intuition. Now, in May 2026, I revisited those forecasts using an AI-driven analysis with 41 Opus 4.8 agents to cross-reference them with the latest data. The assessment used symbols: ✅ Correct, 🟢 Mostly Correct, 🟡 Partially Correct, ❌ Incorrect. Overall, the directional judgments held up well, with only one major factual error regarding GPT-4's rumored parameter size (incorrectly cited as 100T). However, nuances and degrees of accuracy revealed more. **What Was Largely Correct:** Predictions about mechanisms and directions proved accurate. The rise of RAG (Retrieval-Augmented Generation) as the standard architecture for combating AI hallucination was confirmed, as was the transformative potential of LUI (Language User Interface) in creating a new industry layer atop GUIs. The emergence of "robot networks" (agent-to-agent communication protocols) and China's rapid catch-up in developing capable large models (closing the performance gap with top models to ~2.7%) were also on point. The analysis affirmed that LLMs lack consciousness and that the Turing Test merely measures perceived intelligence. **What Was Off Target:** Errors often involved specific numbers, over-optimistic timelines, or misjudged distributions. The prediction that value would primarily accrue to the application layer was half-right but missed NVIDIA's dominance as the profitable infrastructure layer. Forecasts about AI circumventing copyright issues and fostering a "global common ground" by averaging human viewpoints were incorrect; instead, major copyright settlements occurred and AI personalization is increasing. Estimates for model training costs ("$5-10 billion cap") were significantly off, underestimating frontier costs and overestimating replication costs. The notion that LLMs could never do complex math without tools was disproven by later models winning IMO gold. **Key Patterns from the Review:** 1. **Direction over precision:** Judgments about mechanisms and trends were more reliable than specific numbers or definitive statements. 2. **Timing bias:** There was a tendency to overestimate short-term speed but underestimate long-term magnitude and transformation. 3. **The distribution blind spot:** Aggregate-level correctness often masked uneven impacts (e.g., on young professionals' employment). 4. **The value of qualifiers:** Predictions framed with caution (e.g., "reportedly," "for now," "prototype in 2-3 years") aged better. 5. **Some debates continue:** Issues like the nature of "emergent abilities" or machine consciousness remain unresolved. This three-year review highlights that while seeing the big picture is crucial, humility regarding specifics, timelines, and disparate impacts is essential for future forecasting.

链捕手Há 10h

Three Years Later: Looking Back on My 2023 Predictions for ChatGPT

链捕手Há 10h

AI Bubble Warning: AI Investments Are Negative Returns for Most Tech Giants

The article issues a stark warning about a potential AI investment bubble. It notes that while the AI boom shares similarities with the TMT bubble of the late 1990s, its scale is vastly larger, currently driving 93% of U.S. GDP growth. Major hyperscale cloud providers like Microsoft, Alphabet, Amazon, Meta, and Oracle are planning to invest trillions in AI data centers over the coming years. However, calculations based on analyst projections for 2025-2030 reveal a concerning math problem: expected capital expenditure growth far outpaces projected revenue growth. Even under an extremely optimistic scenario of zero costs, the implied return on investment for most of these tech giants (except Amazon) is deeply negative. This suggests that the current trajectory could lead to one of history's largest shareholder value destruction events. The piece outlines two potential escapes: AI generating vastly more revenue than currently anticipated—a near-impossible task—or a significant cutback in the planned investment splurge. The latter scenario could trigger a domino effect, severely impacting the entire tech supply chain (from Nvidia to TSMC), potentially pushing the U.S. economy into recession, and causing a major stock market downturn. The author suggests upcoming high-profile IPOs by companies like OpenAI and Anthropic might represent a transfer of risk from early investors to public market participants. While the peak of the hype cycle might sustain investment through 2026, the fundamental financial dilemma remains unresolved, setting the stage for a potential market correction in 2027 or 2028, similar to the years following Alan Greenspan's "irrational exuberance" warning.

marsbitHá 11h