AI agents

Moca Chain

2026-09-30 22:49:20

Moca Chain mainnet goes live with verifiable credential sharing for users and AI agents

Moca Foundation and Moca Network, projects under Animoca Brands, said the Moca Chain mainnet has launched with support for verifiable credential sharing across applications for both users and AI agents. The system lets users keep custody of their credentials and choose when and where to share them. Moca Chain also allows institutions to issue tamper-resistant credentials linked to a unified digital identity, covering items such as identity verification and membership information. Alongside the mainnet launch, Moca Network introduced AIR, a unified access point for enterprises connecting with users and AI agents while reducing the need to store sensitive data. The company said the AIR ecosystem already includes more than nine partners, naming SK Planet, Biletinial, Oyunfor, Open Campus, OneFootball, Automobili Lamborghini, Inveo Kripto, Plume and Nansen, with integrations planned to be completed in 2026. The report was cited by Bitcoin.com News.

40

AI agents

2026-09-30 20:13:38

AI agents hire humans for 3D world scanning and pay in USDC over Arc

Techub News reported that AI agents are hiring humans to carry out 3D world-scanning tasks and paying them in USDC through Circle’s Arc blockchain. The setup points to a practical use case where artificial intelligence and blockchain infrastructure meet in real-world data collection. It also highlights a microtask-based payment flow, with stablecoin settlement used as the compensation mechanism. The report frames the model as an example of how AI systems can coordinate human labor for physical-world input while blockchain rails handle payment.

20

OpenAI

2026-09-30 19:06:25

OpenAI releases small-business AI report and partners with ASBDC on local training

OpenAI has released a new report examining how small businesses can use AI agents across day-to-day operations, including customer acquisition, product development, and financial management. At the same time, the company announced a new partnership with the Association of Small Business Development Centers, or ASBDC, aimed at helping more small-business owners begin using AI in practical settings. According to the Techub News brief, the partnership will focus on hands-on AI training and localized guidance rather than broad, general instruction. The report and the collaboration together center on one point: giving small businesses more concrete ways to put AI tools to work across core business functions. No additional details on timing, program scale, or rollout locations were disclosed in the source brief.

20

Intel

2026-09-30 17:57:44

Intel adds Nvidia OpenShell policy layer to its AI agent toolkit

ChainCatcher reported that Intel (INTC.O) has added Nvidia’s (NVDA.O) OpenShell policy layer to its AI agent toolkit. The brief update points to a new integration involving the AI software stacks of the two chip companies. The source did not provide additional details on rollout timing, product scope, technical specifications, or which parts of the toolkit will use the OpenShell policy layer. No further comments or supporting information were included in the original newsflash.

30

CoinDesk

2026-09-30 14:56:07

Moon Pursuit Capital says autonomous AI agents will need blockchain-based financial rails

CoinDesk’s latest Crypto Long & Short newsletter centers on a thesis from Moon Pursuit Capital founder and managing partner Utkarsh Ahuja: if artificial intelligence becomes a major economic force, the financial infrastructure behind that activity may increasingly rely on blockchain and digital assets. Ahuja argues that the key opportunity is not simply in AI models or applications, and not in attaching tokens to every project, but in the infrastructure layer that can support machine-to-machine payments, settlement, identity, ownership records and value transfer. In his view, autonomous AI agents are moving beyond text generation and data analysis toward negotiating with other agents, buying compute, paying for data, executing transactions and managing financial decisions within human-set parameters. Traditional payment systems were built for people and institutions, not for potentially millions of software agents making continuous, low-value, cross-border transactions. That is where blockchain rails, smart contracts, wallets and stablecoins may fit. The same newsletter also highlights several institutional developments: Goldman Sachs routing its roughly $100 billion FTIXX Treasury fund into crypto settlement infrastructure through Lynq without tokenizing it; Cboe Global Markets and S&P Dow Jones Indices extending a licensing agreement that leaves room for tokenized options; Vitalik Buterin outlining a different Ethereum vision for 2030; and bitcoin recovering above $84,000 after briefly falling to $82,500 as Treasury yields steadied.

20

Coinbase

2026-09-30 14:54:05

Brian Armstrong says Coinbase is building financial accounts for superintelligence

Coinbase CEO Brian Armstrong said in a post on X that the company is building financial accounts for superintelligence. He said the accounts are intended to support services including trading, payments, and lending. According to Armstrong, the product could help users build wealth and manage personal finances, while also giving AI agent teams their own standalone accounts. Those accounts, he said, would allow such teams to carry out more tasks on behalf of users. The remarks outline how Coinbase is framing a financial layer for AI-driven activity, centered on account infrastructure rather than a single standalone payment tool. No further product details, launch timing, or technical specifications were disclosed in the post cited by ChainCatcher.

30

OpenAI

2026-09-30 10:30:42

OpenAI sued over AI agent cyberattack incident

OpenAI has been sued over an AI agent cyberattack incident, according to Techub News, which cited CNBC. The case appears to be the first publicly reported lawsuit aimed at holding an AI developer responsible for an incident caused by a system that went out of control. The brief report did not disclose further details about the claims, the parties involved, or the court handling the case. What is clear from the report is the legal angle: the lawsuit is framed around whether an AI developer can be held liable for damage tied to an AI system’s actions. CNBC was cited as the source for that characterization.

180

OpenClaw

2026-09-30 09:21:36

OpenClaw open-sources enterprise edition as OpenAI uses it internally to fix bugs and merge PRs

OpenClaw Foundation has released OpenClaw Enterprise, or OCE, an open-source system designed to help companies deploy and manage long-running AI agents on their own infrastructure. The project began inside OpenAI before being donated to the OpenClaw Foundation, with Red Hat and NVIDIA listed among contributors. OCE is released under the MIT license, can be self-hosted, and does not require companies to use OpenAI models or tooling. The product is aimed at a practical problem many companies face after rolling out agents: how to manage them at scale. According to the announcement, administrators can create multiple agents from a central interface, set access controls for users and agents, and add isolation, auditing, and lifecycle management. The underlying model, agent harness, and sandbox are all replaceable. OpenAI and Red Hat have already tested OCE internally. OpenAI’s internal agent, called Androidclaw, can read code repositories, GitHub, and logs. In the examples shared, it can locate related pull requests and attempt fixes after build failures, and it can also trace product issues, submit PRs with video evidence, and even merge them. The foundation said OCE is not yet ready for formal production use and is currently recommended only for internal pilots, with version 1.0 planned for later this year.

190