News Heat Trend
Project Overview
MindCP (Model + Intent + Decentralized Context Protocol) is a new standard for building intelligent, context-aware AI systems. MindCP is designed for the future of decentralized applications, where users have control over their data.
AI security company Neo announces completion of $100 million funding round, led by Andreessen Horowitz and Bessemer Venture Partners
According to the official press release, AI security company Neo officially announced today the completion of $100 million in financing, led by Andreessen Horowitz and Bessemer Venture Partners, with participation from Craft Ventures and Merlin Ventures. Neo was co-founded by former executives from SentinelOne, Wiz, and Palo Alto Networks, positioned as an "Intelligent Software Control Company", aiming to provide enterprises with a real-time control layer, helping SecOps teams conduct comprehensive inventory, risk assessment, behavior attribution, and policy enforcement on AI agents, AI applications, browser extensions, MCP servers, and traditional software. According to Gartner's forecast, only 5% of enterprise applications will have agent capabilities in 2025, while by the end of 2026 this proportion will rise to 40%.
AI financial data infrastructure company Daloopa completes $47 million Series C funding round, led by Brighton Park Capital
: AI financial data infrastructure company Daloopa announced the completion of a $47 million Series C funding round, led by Brighton Park Capital, with participation from Squarepoint Capital, Touring Capital, and Nexus Venture Partners. The funds will be used to accelerate platform expansion and expand the engineering, product, and marketing teams.Daloopa provides structured and traceable financial data for financial institutions. It has currently expanded data access through MCP connectors with OpenAI ChatGPT, Anthropic Claude, Perplexity, and Rogo, and offers APIs, cloud delivery, and partner APIs, enabling third-party developers to use its high-quality financial data in their own products. (PRNewswire)
Dune: Lays Off 25% to Optimize AI Efficiency and Restructure Core Business
Dune co-founder and CEO Fredrik Haga stated that Dune is laying off 25% of its staff this week as part of a restructuring aimed at further concentrating efforts on developing the industry's core data products.Haga noted that the company is increasing investment in AI-driven tools, using Dune MCP to enable teams and agents to build dashboards and workflows without requiring knowledge of SQL or data infrastructure. Additionally, as assets like currencies, stocks, and bonds accelerate their migration on-chain, the company is heavily investing in the data layer and providing services to institutions. Dune currently has ample financial reserves, having completed a $69.42 million Series B funding round led by Coatue in 2022. (theblock)
Claw Intelligence closes $3 million seed funding round, with participation from Castrum Istanbul and others
According to official announcements, Web3 intelligence-layer project Claw Intelligence has raised $3 million in seed funding. Investors include Castrum Istanbul, Titans Ventures, Super Labs, and Genesis Capital. Claw Intelligence is simplifying user interaction with the Web3 ecosystem and underlying computing resources through its unified intelligence layer, lowering operational barriers. The platform leverages an encryption-native Model Context Protocol (MCP) service to transform fragmented data endpoints into conversational workflows. New features include a secure, isolated large language model (LLM)-powered code execution sandbox that safely runs code directly within the chat interface—enabling real-time computation, data processing, and script prototyping; and LLM-driven cross-device/multi-computer control, allowing users to centrally manage multiple devices and servers via natural-language commands.
Etherscan launches AI tool suite, supporting AI agents to access data from over 60 EVM chains
Odaily News: The suite covers over 60 EVM-compatible chains and provides AI agents with structured, verified, real-time on-chain data access through three methods: MCP, CLI, and Skills. Etherscan MCP supports programming agents like Claude and Codex in querying on-chain information via natural language; the CLI tool outputs results in JSON, table, or CSV format in the terminal and comes with an interactive interface; the Skills module includes built-in features such as Orchestrator, contract auditing, transaction debugging, and workflow orchestration. Etherscan stated that developers can refer to the official documentation to use these tools and can submit feedback or share applications built on top of them.
Coinbase Builds AI "Digital Teammate" CEEcil: Long-Term Memory, Proactive Follow-Ups, and Autonomous Judgment
Odaily News - Coinbase has disclosed the architecture and application of its internal AI system "CEEcil." This system is not a simple chatbot but is designed as an AI agent that functions like a "team member." It can be mentioned in Slack, and possesses long-term memory, proactive follow-up, and judgment capabilities.CEEcil retains team context through a three-tier memory mechanism: in the background, it continuously extracts information such as decisions, blocked items, and owners from authorized channels; each night, it consolidates short-term observations into long-term summaries; and when answering questions, it retrieves historical memories, recent conversations, and real-time data on demand. The system also adopts a tiered model architecture, where simple queries are prioritized for processing via APIs or knowledge bases, and only complex, multi-step tasks are delegated to more powerful AI models, thereby reducing cost and latency.In practical applications, CEEcil can proactively join relevant Slack discussions, send emoji reactions, and automatically follow up when issues go unaddressed. Coinbase states that the system also once refused to submit operational documents containing customer identity information to Git, suggesting the use of redacted versions or compliant storage methods instead—demonstrating a certain level of security judgment.Architecturally, CEEcil consists of a Go service and an AI agent runtime, and connects to tools such as memory retrieval, knowledge bases, real-time queries, and Slack via MCP. Coinbase also has set up a real-time "kill switch," call limits and spending caps, audit mechanisms, and requires human review and merging of the code it generates.Coinbase says CEEcil is now able to help teams handle production incidents and answer expert knowledge that previously would have required waiting hours to obtain. In the future, the company hopes to expand this model into multiple AI agents tailored to specific teams, enabling AI not only to "answer questions" but also to continuously remember, proactively participate, and complete actual work.
Gate Founder and CEO Dr. Han Featured in The Economist Interview: Gate Accelerates Construction of Multi-Asset Financial Infrastructure
Odaily News – According to the latest report by The Economist Enterprise, after 13 years of development and multiple strategic upgrades, Gate is gradually evolving from a traditional crypto asset trading platform into a comprehensive financial infrastructure connecting digital assets with traditional financial markets. The report notes that as more traditional financial products — including stocks, ETFs, tokenized assets, forex, and metals — enter the digital asset ecosystem, Gate is leveraging its multi-asset product layout and global infrastructure development to further bridge the gap between crypto finance and TradFi. In the interview, Gate Founder and CEO Dr. Han stated that as the industry evolves, digital asset platforms are no longer facing only technical challenges, but also risk management, user protection, and industry regulation.In terms of TradFi integration, Gate has established a multi-layered framework spanning tokenized assets, derivatives, and native stock trading, successively launching products and services such as xStocks, Ondo, Gate TradFi, Pre-IPOs, Direct IPO Access, and Gate Stocks. Among these, Gate Stocks now supports stock trading in the U.S., Hong Kong, and South Korea, significantly lowering the barrier for global users to participate in traditional financial markets. Meanwhile, Gate continues to enhance its 24/7 trading and liquidity infrastructure to meet the cross-market asset allocation needs of users worldwide.The Economist Enterprise also pointed out that as the digital asset market moves further toward institutionalization, compliance, transparency, and infrastructure capabilities are becoming key components of platform competitiveness. Gate continues to advance its compliance framework across multiple global jurisdictions and has provided third-party audits and open-source Proof of Reserves since 2020. In addition, Gate is introducing AI infrastructure into Web3, connecting AI with trading, wallets, and more services through products such as Gate AI, Gate MCP, and GateClaw, driving the platform's evolution from a traditional trading venue into a comprehensive financial infrastructure that connects digital assets, traditional finance, and AI applications.
Base lead: Base is striving to become the underlying network for global financial activities
Base lead Jesse Pollak introduced Base's latest development direction at the Coinbase System Update: Take Control conference held today. Base is committed to becoming a platform that supports global financial activities.Base announced support for tokenized stocks launched by Coinbase and introduced the B20 native token standard. The Base App will support Solana, Bitcoin, and multiple EVM networks, and will also launch a web version. Base introduced Base MCP, allowing AI Agents to perform on-chain activities such as transfers, transactions, swaps, and lending with user authorization.Base stated that it has processed over $19 trillion in stablecoin payment volume this year and supports more than 25 local currency stablecoins. Base has also launched private transactions and a new ledger architecture to meet corporate compliance and auditing requirements.
a16z Crypto explores the fundamental tension between “censorship resistance” and “low latency” in blockchains
a16z Crypto published an article exploring the fundamental contradiction between blockchain “censorship resistance” and “low latency,” pointing out that any Byzantine Fault Tolerant (BFT) blockchain protocol with censorship resistance, where more than one-fifth of validators could be malicious, requires at least 5 rounds of communication for its optimal good-case latency, whereas traditional BFT consensus only requires a minimum of 3 rounds.The article notes that in traditional BFT protocols, the block proposer holds the power for both block construction and consensus progression, allowing them to censor by excluding specific transactions. This is also a major source of the MEV problem. To address this issue, Ethereum is researching FOCIL / EIP-7805, while Solana is exploring mechanisms like Constellation and MCP. The core idea behind these approaches is to have validators proactively collect "Inclusion Lists" of transactions that cannot be ignored before a block is formally proposed.a16z Crypto states that achieving censorship resistance requires two additional rounds of communication: first, user transactions must be broadcast to all validators, and then the validators need to confirm and write these into an inclusion list before the consensus process can begin. Therefore, in a partially synchronous network environment, there is no protocol design that can achieve both BFT and censorship resistance in just 4 rounds; 5 rounds represent the mathematical lower limit.The article emphasizes that while a censorship-resistant mechanism increases protocol latency, it can significantly reduce the "effective latency" users actually experience. In a system lacking censorship resistance, transactions may be indefinitely delayed due to validator censorship. Conversely, in a system with censorship resistance guarantees, transactions will be included in a block within a maximum of 5 rounds of communication, making transaction confirmation times more predictable.
PrimePiper Launches Prime Broker Dedicated to AI Agents, Enabling Multi-Exchange Connectivity, Cross-Venue Reconciliation, and Risk Control & Audit Capabilities
PrimePiper has launched an enterprise-grade prime broker platform for AI agents, designed to address challenges including fragmented account management, inadequate risk control, inability to reconcile across venues, and insufficient compliance auditing in AI-driven automated trading. According to the company, its infrastructure supports unified connectivity to multiple trading venues—including Hyperliquid, OKX, Tiger Brokers, and Interactive Brokers (IBKR). For risk control, PrimePiper offers enterprise-grade API key management, spending limits, and circuit-breaker mechanisms to constrain AI agent trading behavior. At the execution layer, it enables automated strategy execution via SDK or the Model Context Protocol (MCP). For compliance and auditing, it provides audit-grade reporting capabilities tailored for funds and traders. PrimePiper has been selected for the latest cohort of Founders Inc’s accelerator program; its product is currently in the Alpha stage. Team members hail from Galois Capital, Kraken, DRW, and AWS.
Expected to hit a record high, South Korea's June storage semiconductor exports reach 38 to 42 billion USD
: Citrini analyst jukan posted on the X platform that driven by AI demand and supply shortages, South Korea's storage semiconductor export value and unit price rose simultaneously in June. According to trade statistics from the Korea Customs Service, as of June 20, 2026, the total export value of major storage products exceeded 23 billion USD, already surpassing 60% of the 37.16 billion USD total export value for the entire month of May.Based on the recent monthly growth of HBM, NAND, and SSD, the total storage semiconductor export value in June is expected to reach 38 to 42 billion USD, surpassing last month's record and reaching an all-time high. In June, both the export value and unit price of various storage semiconductors rose simultaneously, with the export value of MCP (i.e., HBM) increasing by 51% month-over-month.Influenced by continued investment in AI data centers by global tech giants like NVIDIA, the supply shortage of HBM3E and HBM4, centered around SK Hynix, persists. As wafers are concentrated on HBM production, the supply of general DRAM has relatively decreased, pushing the unit price of general DRAM to approximately 2 to 3 times that of the same period last year. Demand for NAND flash and SSD is growing alongside the expansion of AI inference server construction, with both product categories seeing a month-over-month increase of 25% to 28%. The share of storage products in total semiconductor exports has risen from the previous 70% range to 90% this month. The combined export value of storage and system semiconductors reached 25.5 billion USD, with the total semiconductor export value for June expected to be between 42 and 46 billion USD.Park Junyoung, a researcher at Hanwha Investment & Securities, stated that South Korea's storage industry is currently overcoming weaknesses by relying on two tools: long-term supply agreements and HBM. Even if the profitability of the storage industry declines in the future, the drop in operating profit is unlikely to be as severe as in the past.
South Korea's preliminary export data for the first 20 days of June shows a significant year-on-year surge in memory chips, with SSD exports up 405% year-on-year.
Odaily News: Citrini analyst jukan posted on platform X that South Korea's preliminary export data for June 1-20 shows DRAM (including modules) up 342% year-on-year and 3% month-on-month; NAND flash up 336% year-on-year and 28% month-on-month; MCP (HBM) up 209% year-on-year and 51% month-on-month; SSD up 405% year-on-year and 25% month-on-month.
AWS partners with Coinbase and Stripe to launch AI agent stablecoin payment functionality, enabling microtransactions using USDC
According to The Block, Amazon Web Services (AWS) has partnered with Coinbase and Stripe to launch Amazon Bedrock AgentCore Payments, enabling AI agents to conduct transactions using stablecoins. Coinbase stated that developers can build “agent-based payment” solutions using the x402 protocol, allowing AI agents to make micro-payments in USDC. This feature enables AI agents to instantly pay for web content, APIs, MCP servers, and other agents. AWS noted that developers can choose between Coinbase and Stripe wallets and fund those wallets using either stablecoins or fiat currency.
OKX Agent Trade Kit Launches "Smart Money Signal Suite," Opening Trader Data Capabilities
according to official sources, OKX Agent Trade Kit has launched the "Smart Money Signal Suite," which integrates the real-time positions, win rates, and profit/loss data of over 1,000 popular traders from OKX Star and packages them as signal tools callable by AI Agents. This suite supports long/short market analysis and allows for multi-dimensional analysis including capital weighting, trader screening, and signal quality evaluation. Additionally, the system can track trend changes and sentiment shifts, providing auxiliary support for trading decisions. These features have been integrated into the Agent Trade Kit, and users can access them after upgrading to the latest version.It is reported that OKX Agent Trade Kit is an open-source exchange MCP toolset designed for AI Agents and professional traders. It previously launched suites such as "Skill Square," "Market Screening and Open Interest Analysis," and "Sentiment Radar."
PrimePiper Launches Prime Broker Dedicated to AI Agents, Enabling Multi-Exchange Connectivity, Cross-Venue Reconciliation, and Risk Control & Audit Capabilities
PrimePiper has launched an enterprise-grade prime broker platform for AI agents, designed to address challenges including fragmented account management, inadequate risk control, inability to reconcile across venues, and insufficient compliance auditing in AI-driven automated trading. According to the company, its infrastructure supports unified connectivity to multiple trading venues—including Hyperliquid, OKX, Tiger Brokers, and Interactive Brokers (IBKR). For risk control, PrimePiper offers enterprise-grade API key management, spending limits, and circuit-breaker mechanisms to constrain AI agent trading behavior. At the execution layer, it enables automated strategy execution via SDK or the Model Context Protocol (MCP). For compliance and auditing, it provides audit-grade reporting capabilities tailored for funds and traders. PrimePiper has been selected for the latest cohort of Founders Inc’s accelerator program; its product is currently in the Alpha stage. Team members hail from Galois Capital, Kraken, DRW, and AWS.
Aave Launches Official MCP Server, Enabling AI Agents to Access Protocol Data and Prepare Transactions
According to The Block, Aave has officially launched its official MCP server, allowing AI agents to access real-time protocol data from Aave V3 and V4 and prepare transactions. The server supports integration with MCP-compatible tools such as ChatGPT and Claude, enabling users to independently control transaction signing permissions.
AI Agent's Next Stop: From "Generating" to "Executing" – X-Agent Explores Automation of Real-World Tasks
Odaily News – While "Vibe Coding" has enabled more people to generate applications using natural language, the next competitive frontier for AI Agents is shifting toward a harder challenge: how to truly accomplish a real-world task that spans multiple platforms and services, involving budgets and authorizations.In X-Agent's latest Medium article, "Beyond Vibe Coding, Toward Agentic Execution," the team uses "booking a trip to Tokyo" as an example to outline its exploration of Agentic Execution. Users would no longer need to switch back and forth between flight, hotel, and payment pages; instead, they would initiate the task with a single request, and the AI would coordinate search, comparison, confirmation, and subsequent execution. Throughout the process, the user retains control over destination, budget, terms, and final authorization.However, the gap between "recommending options" and "delivering reliable execution" still involves a whole set of infrastructure challenges. Price fluctuations, login verifications, order timeouts, and partially successful bookings all require the Agent to preserve task state, distinguish between confirmed and unconfirmed results, and pause to re-request user authorization when needed—rather than blindly retrying or spending automatically.To address this direction, X-Agent proposes decoupling and coordinating four types of capabilities—conversation, task, execution, and events—while exploring different execution pathways such as WebMCP, APIs, MCP tools, browser automation, and Computer Use. The core principle is: tools can help Agents act, but they must not bypass user permissions, budget limits, or result verification.X-Agent notes that the article currently describes a product direction still under exploration and validation, rather than a general-purpose booking capability that has already been launched. The goal is to advance AI from "understanding and generating" toward "completing tasks under user control," gradually making natural language a new gateway connecting real-world services with digital execution.
Etherscan launches AI tool suite, supporting AI agents to access data from over 60 EVM chains
Odaily News: The suite covers over 60 EVM-compatible chains and provides AI agents with structured, verified, real-time on-chain data access through three methods: MCP, CLI, and Skills. Etherscan MCP supports programming agents like Claude and Codex in querying on-chain information via natural language; the CLI tool outputs results in JSON, table, or CSV format in the terminal and comes with an interactive interface; the Skills module includes built-in features such as Orchestrator, contract auditing, transaction debugging, and workflow orchestration. Etherscan stated that developers can refer to the official documentation to use these tools and can submit feedback or share applications built on top of them.
Coinbase Builds AI "Digital Teammate" CEEcil: Long-Term Memory, Proactive Follow-Ups, and Autonomous Judgment
Odaily News - Coinbase has disclosed the architecture and application of its internal AI system "CEEcil." This system is not a simple chatbot but is designed as an AI agent that functions like a "team member." It can be mentioned in Slack, and possesses long-term memory, proactive follow-up, and judgment capabilities.CEEcil retains team context through a three-tier memory mechanism: in the background, it continuously extracts information such as decisions, blocked items, and owners from authorized channels; each night, it consolidates short-term observations into long-term summaries; and when answering questions, it retrieves historical memories, recent conversations, and real-time data on demand. The system also adopts a tiered model architecture, where simple queries are prioritized for processing via APIs or knowledge bases, and only complex, multi-step tasks are delegated to more powerful AI models, thereby reducing cost and latency.In practical applications, CEEcil can proactively join relevant Slack discussions, send emoji reactions, and automatically follow up when issues go unaddressed. Coinbase states that the system also once refused to submit operational documents containing customer identity information to Git, suggesting the use of redacted versions or compliant storage methods instead—demonstrating a certain level of security judgment.Architecturally, CEEcil consists of a Go service and an AI agent runtime, and connects to tools such as memory retrieval, knowledge bases, real-time queries, and Slack via MCP. Coinbase also has set up a real-time "kill switch," call limits and spending caps, audit mechanisms, and requires human review and merging of the code it generates.Coinbase says CEEcil is now able to help teams handle production incidents and answer expert knowledge that previously would have required waiting hours to obtain. In the future, the company hopes to expand this model into multiple AI agents tailored to specific teams, enabling AI not only to "answer questions" but also to continuously remember, proactively participate, and complete actual work.
x402 has processed over 205 million transactions with a volume of $53 million, Coinbase unveils AiFi financial system
Odaily News, Coinbase stated on the X platform that the financial layer emerging around AI agents is called agentic finance, namely AiFi, covering agent transactions and agent commerce. AI agents can research information, make decisions, and execute actions on behalf of individuals and businesses, while also needing to perform fundamental financial activities such as making payments, receiving funds, storing value, and transferring money. Coinbase stated that it is building tools for AI agents to enable them to continuously execute financial operations within preset goals, permissions, limits, and security boundaries.Coinbase stated that Coinbase Advisor, Coinbase for Agents, x402, and Coinbase Business have formed the AiFi infrastructure. Coinbase Advisor is an SEC-registered investment advisor integrated into the app for Coinbase One users, providing portfolio analysis and guidance through a chat interface. Coinbase for Agents connects tools such as Claude, ChatGPT, and Cursor to Coinbase accounts via MCP, allowing users to set operating conditions and security boundaries for agents, as well as create separate sandbox accounts for agent use.Coinbase stated that x402 is an open payment standard initiated by Coinbase and now managed by an independent foundation, supporting AI agents in paying fees to other agents, applications, and services. To date, x402 has processed over 205 million transactions with a volume of $53 million, covering 200,000 sellers, with Coinbase facilitating 67% of this activity. Individual payments can be used for API calls, datasets, research materials, or computing resources.Coinbase stated that Coinbase Business supports enterprises in natively receiving USDC payments from AI agents, while Coinbiz helps merchants sell services and content to software without relying on card credentials, manual invoicing, or payment processes designed around human customers. Eligible businesses can earn a floating 3.35% reward on idle USDC balances. USDC currently accounts for 99% of on-chain agent commerce transactions, x402 accounts for 97% of agent finance protocol usage, and the related activities run on Base.
From one person to one system: X-Agent redefines the business model for OPCs
Odaily News, X-Agent recently published a technical article titled "An OPC Is a Stack, Not a Headcount Story," arguing that as AI significantly reduces R&D and operational costs, the competitiveness of a One Person Company (OPC) should no longer be measured by how much work a single person can complete.The article points out that an OPC with true commercial capability needs to establish a complete closed loop spanning demand discovery, AI production, capability infrastructure, distribution, economic systems, and market feedback.And this is precisely the problem X-Agent aims to solve.At the capability layer, X-Agent leverages MCP to help individual developers productize their capabilities, transforming a project into a service that can be discovered, invoked, and verified by Agents. At the commercial layer, X-Agent further connects Agent distribution and payment infrastructure, enabling services to enter real Agent workflows, with usage-based billing, budget management, payment, and settlement.Meanwhile, the X-Agent MCP Hackathon is also driving this model into practice: developers do not need to start by building an "MCP server" from scratch. Instead, they first prove that a real capability can solve a practical problem, then convert it into a production-grade Agent service through standardization, deployment verification, and security checks.Therefore, what X-Agent is building is not just a development tool, but a framework that helps OPCs complete:Demand discovery → Capability building → MCP productization → Agent distribution → Commercial settlement → Market feedbackUltimately forming a self-reinforcing AI business closed loop.The OPC of the future is not one person taking on all the work, but one person using AI and Agents to organize a complete set of company-level capabilities.
Related news
Aave Launches Official MCP Server, Enabling AI Agents to Access Protocol Data and Prepare Transactions
According to The Block, Aave has officially launched its official MCP server, allowing AI agents to access real-time protocol data from Aave V3 and V4 and prepare transactions. The server supports integration with MCP-compatible tools such as ChatGPT and Claude, enabling users to independently control transaction signing permissions.
AI Agent's Next Stop: From "Generating" to "Executing" – X-Agent Explores Automation of Real-World Tasks
Odaily News – While "Vibe Coding" has enabled more people to generate applications using natural language, the next competitive frontier for AI Agents is shifting toward a harder challenge: how to truly accomplish a real-world task that spans multiple platforms and services, involving budgets and authorizations.In X-Agent's latest Medium article, "Beyond Vibe Coding, Toward Agentic Execution," the team uses "booking a trip to Tokyo" as an example to outline its exploration of Agentic Execution. Users would no longer need to switch back and forth between flight, hotel, and payment pages; instead, they would initiate the task with a single request, and the AI would coordinate search, comparison, confirmation, and subsequent execution. Throughout the process, the user retains control over destination, budget, terms, and final authorization.However, the gap between "recommending options" and "delivering reliable execution" still involves a whole set of infrastructure challenges. Price fluctuations, login verifications, order timeouts, and partially successful bookings all require the Agent to preserve task state, distinguish between confirmed and unconfirmed results, and pause to re-request user authorization when needed—rather than blindly retrying or spending automatically.To address this direction, X-Agent proposes decoupling and coordinating four types of capabilities—conversation, task, execution, and events—while exploring different execution pathways such as WebMCP, APIs, MCP tools, browser automation, and Computer Use. The core principle is: tools can help Agents act, but they must not bypass user permissions, budget limits, or result verification.X-Agent notes that the article currently describes a product direction still under exploration and validation, rather than a general-purpose booking capability that has already been launched. The goal is to advance AI from "understanding and generating" toward "completing tasks under user control," gradually making natural language a new gateway connecting real-world services with digital execution.
B.AI cumulative users surpass 2.3 million, continuously building infrastructure for the AI agent era.
One-stop AGI infrastructure platform B.AI announced that its cumulative user base has surpassed 2.3 million, maintaining strong growth momentum. As an intelligent global settlement layer connecting models, Agents, and value flows, B.AI continuously optimizes compute costs, aggregates cutting-edge large models, and employs intelligent routing and scheduling to bring top-tier AI capabilities to broader real-world application scenarios at a lower entry threshold and with higher cost-effectiveness. From LLM Services, BAIclaw, and MCP to Agent payments, B.AI is building an end-to-end infrastructure that covers model invocation, Agent execution, and value settlement—positioned above all models and below all Agents, where every single call is routed, metered, and settled.
Etherscan launches AI tool suite, supporting AI agents to access data from over 60 EVM chains
Odaily News: The suite covers over 60 EVM-compatible chains and provides AI agents with structured, verified, real-time on-chain data access through three methods: MCP, CLI, and Skills. Etherscan MCP supports programming agents like Claude and Codex in querying on-chain information via natural language; the CLI tool outputs results in JSON, table, or CSV format in the terminal and comes with an interactive interface; the Skills module includes built-in features such as Orchestrator, contract auditing, transaction debugging, and workflow orchestration. Etherscan stated that developers can refer to the official documentation to use these tools and can submit feedback or share applications built on top of them.
Coinbase Builds AI "Digital Teammate" CEEcil: Long-Term Memory, Proactive Follow-Ups, and Autonomous Judgment
Odaily News - Coinbase has disclosed the architecture and application of its internal AI system "CEEcil." This system is not a simple chatbot but is designed as an AI agent that functions like a "team member." It can be mentioned in Slack, and possesses long-term memory, proactive follow-up, and judgment capabilities.CEEcil retains team context through a three-tier memory mechanism: in the background, it continuously extracts information such as decisions, blocked items, and owners from authorized channels; each night, it consolidates short-term observations into long-term summaries; and when answering questions, it retrieves historical memories, recent conversations, and real-time data on demand. The system also adopts a tiered model architecture, where simple queries are prioritized for processing via APIs or knowledge bases, and only complex, multi-step tasks are delegated to more powerful AI models, thereby reducing cost and latency.In practical applications, CEEcil can proactively join relevant Slack discussions, send emoji reactions, and automatically follow up when issues go unaddressed. Coinbase states that the system also once refused to submit operational documents containing customer identity information to Git, suggesting the use of redacted versions or compliant storage methods instead—demonstrating a certain level of security judgment.Architecturally, CEEcil consists of a Go service and an AI agent runtime, and connects to tools such as memory retrieval, knowledge bases, real-time queries, and Slack via MCP. Coinbase also has set up a real-time "kill switch," call limits and spending caps, audit mechanisms, and requires human review and merging of the code it generates.Coinbase says CEEcil is now able to help teams handle production incidents and answer expert knowledge that previously would have required waiting hours to obtain. In the future, the company hopes to expand this model into multiple AI agents tailored to specific teams, enabling AI not only to "answer questions" but also to continuously remember, proactively participate, and complete actual work.
x402 has processed over 205 million transactions with a volume of $53 million, Coinbase unveils AiFi financial system
Odaily News, Coinbase stated on the X platform that the financial layer emerging around AI agents is called agentic finance, namely AiFi, covering agent transactions and agent commerce. AI agents can research information, make decisions, and execute actions on behalf of individuals and businesses, while also needing to perform fundamental financial activities such as making payments, receiving funds, storing value, and transferring money. Coinbase stated that it is building tools for AI agents to enable them to continuously execute financial operations within preset goals, permissions, limits, and security boundaries.Coinbase stated that Coinbase Advisor, Coinbase for Agents, x402, and Coinbase Business have formed the AiFi infrastructure. Coinbase Advisor is an SEC-registered investment advisor integrated into the app for Coinbase One users, providing portfolio analysis and guidance through a chat interface. Coinbase for Agents connects tools such as Claude, ChatGPT, and Cursor to Coinbase accounts via MCP, allowing users to set operating conditions and security boundaries for agents, as well as create separate sandbox accounts for agent use.Coinbase stated that x402 is an open payment standard initiated by Coinbase and now managed by an independent foundation, supporting AI agents in paying fees to other agents, applications, and services. To date, x402 has processed over 205 million transactions with a volume of $53 million, covering 200,000 sellers, with Coinbase facilitating 67% of this activity. Individual payments can be used for API calls, datasets, research materials, or computing resources.Coinbase stated that Coinbase Business supports enterprises in natively receiving USDC payments from AI agents, while Coinbiz helps merchants sell services and content to software without relying on card credentials, manual invoicing, or payment processes designed around human customers. Eligible businesses can earn a floating 3.35% reward on idle USDC balances. USDC currently accounts for 99% of on-chain agent commerce transactions, x402 accounts for 97% of agent finance protocol usage, and the related activities run on Base.