I
Ishan Pandey
Guest
Nine in ten professional developers now use AI coding agents at work every week, according to JetBrains' 2026 survey of more than 15,000 of them. Total stablecoin supply has also grown to about $316 billion, which means a rising share of the software that writes code and sends money runs through agents. AEREDIUM, a digital asset infrastructure company based in Austin, Texas, launched two products on 13 October 2026 for that situation: The Chair, which governs AI coding agents, plus MCP Connector, which governs agents that pay and trade. Both place a policy engine between the agent and the system it wants to touch, so that every proposed action is approved, denied or held for a human to review before anything runs.
Stablecoin supply stood at about $27 billion at the end of 2020, reached about $163 billion a year later and finished 2025 near $308 billion, according to DefiLlama figures compiled by Transak, which puts the total at roughly $316 billion on 12 June 2026. Tether's USDT accounts for about 59% of that supply and Circle's USDC for about 24%, while PayPal's PYUSD and the euro-denominated EURC are smaller tokens that sit in the same family of assets. These are the tokens that AEREDIUM's MCP Connector lets an agent send, which connects a market that grew nearly twelvefold in six years with the next group of users who will move it, namely software that acts on a person's behalf.
JetBrains found that 90% of professional developers use AI coding agents at work at least weekly and 68% use them daily, based on responses collected between May and July 2026. The mix of tools shifted quickly during the year. Claude Code rose from 18% to 39% of developers since January, Codex went from 3% to 16%. GitHub Copilot slipped from 29% a year earlier to 21%. A tool that executes commands, edits files and pushes changes inside a company's repositories is a new kind of actor inside a software team. The survey shows that most teams already have one working beside them.
The Chair sits between an AI agent and an organization's development environment and evaluates each tool call before it runs, covering actions such as executing commands, modifying files, merging code and pushing changes. Requests covered by an existing rule are approved or denied automatically. Requests outside the rules are held in a control panel where authorized users approve or reject them. Every decision produces a signed receipt that documents what the agent attempted, which policy applied and who authorized the action. It supports Claude Code through a dedicated plugin and provides a governed shell for other coding agents including those from OpenAI. Organizations write their policies by answering a questionnaire so they never have to code the rules by hand.
Dadon said that AI agents are no longer limited to answering questions, since they write code, move money and make decisions with real consequences. He added that the challenge is making sure they operate within the rules their owners set. He described the two products as a way for organizations to put AI to work without giving up control over what it can do. The Chair is priced at $10 per chair per month or $80 per year after a 30-day free trial. MCP Connector starts at $49 a month across four packages, also with a 30-day trial.
MCP Connector applies the same approach to money by combining a policy engine with a digital asset wallet, so that an agent can execute approved payments and trades while staying inside limits its owner has set. Each agent operates through its own wallet, policy and designated owner, who defines the agent's role, approves its mandate with a passkey and can halt or delete the agent at any time.
A Payer can send USDC, USDT, PYUSD and EURC to addresses its owner has approved and has no permission to trade, while a Trader can trade on designated exchanges within its own wallet and has no permission to send funds to outside recipients. The wallet uses threshold signatures generated inside secure enclaves, which means no complete private key is ever created and AEREDIUM holds no key to the owner's funds. The product supports Ethereum, Arbitrum and Base, keeps transaction records for seven years and is designed for Claude with openness to OpenAI and other AI frameworks.
Both products build on the threshold-signature technology and policy engine that AEREDIUM uses for institutional digital asset security, where a transaction has to satisfy predefined rules before it can be signed and every decision is recorded for later verification. On 2 September 2026 the company launched AERSeal, which uses its AERKey infrastructure to keep key shares in separate hardware-attested enclaves with the CGGMP24 threshold signing protocol so that no full private key exists in one place. Organizations set an M-of-N approval policy that signatories approve with passkeys.
The company says it holds United States Patent 12,694,407, granted on 28 July 2026, for a method in which AI models monitor stablecoins to assess whether each coin is backed by the reserves it claims. It operates from Melbourne, Tel Aviv and Austin, Texas.
The market is moving in the direction these products address. Gartner forecasts that 33% of enterprise software applications will include agentic AI by 2028, up from less than 1% in 2024. It also forecasts that 15% of day-to-day work decisions will be made autonomously by then, up from 0%.
The same forecast expects more than 40% of agentic AI projects to be canceled by the end of 2027 because of rising costs, unclear business value or inadequate risk controls. That last item is the one a policy engine with signed receipts is built to answer. AEREDIUM's launch extends its offering from digital asset custody and settlement into the governance of AI-driven financial and software operations. The pricing, the trial periods and the support for several coding agents give teams a low-cost way to test whether rules written in advance can keep pace with the agents they already run on aeredium.ai.
Don’t forget to like and share the story!
Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR.
Stablecoins Give AI Agents Money to Move
Stablecoin supply stood at about $27 billion at the end of 2020, reached about $163 billion a year later and finished 2025 near $308 billion, according to DefiLlama figures compiled by Transak, which puts the total at roughly $316 billion on 12 June 2026. Tether's USDT accounts for about 59% of that supply and Circle's USDC for about 24%, while PayPal's PYUSD and the euro-denominated EURC are smaller tokens that sit in the same family of assets. These are the tokens that AEREDIUM's MCP Connector lets an agent send, which connects a market that grew nearly twelvefold in six years with the next group of users who will move it, namely software that acts on a person's behalf.
Coding Agents Are Now a Daily Tool for Most Developers
JetBrains found that 90% of professional developers use AI coding agents at work at least weekly and 68% use them daily, based on responses collected between May and July 2026. The mix of tools shifted quickly during the year. Claude Code rose from 18% to 39% of developers since January, Codex went from 3% to 16%. GitHub Copilot slipped from 29% a year earlier to 21%. A tool that executes commands, edits files and pushes changes inside a company's repositories is a new kind of actor inside a software team. The survey shows that most teams already have one working beside them.
The Chair Checks Every Tool Call Before It Runs
The Chair sits between an AI agent and an organization's development environment and evaluates each tool call before it runs, covering actions such as executing commands, modifying files, merging code and pushing changes. Requests covered by an existing rule are approved or denied automatically. Requests outside the rules are held in a control panel where authorized users approve or reject them. Every decision produces a signed receipt that documents what the agent attempted, which policy applied and who authorized the action. It supports Claude Code through a dedicated plugin and provides a governed shell for other coding agents including those from OpenAI. Organizations write their policies by answering a questionnaire so they never have to code the rules by hand.
Dadon said that AI agents are no longer limited to answering questions, since they write code, move money and make decisions with real consequences. He added that the challenge is making sure they operate within the rules their owners set. He described the two products as a way for organizations to put AI to work without giving up control over what it can do. The Chair is priced at $10 per chair per month or $80 per year after a 30-day free trial. MCP Connector starts at $49 a month across four packages, also with a 30-day trial.
MCP Connector Gives Agents Wallets With Limits
MCP Connector applies the same approach to money by combining a policy engine with a digital asset wallet, so that an agent can execute approved payments and trades while staying inside limits its owner has set. Each agent operates through its own wallet, policy and designated owner, who defines the agent's role, approves its mandate with a passkey and can halt or delete the agent at any time.
A Payer can send USDC, USDT, PYUSD and EURC to addresses its owner has approved and has no permission to trade, while a Trader can trade on designated exchanges within its own wallet and has no permission to send funds to outside recipients. The wallet uses threshold signatures generated inside secure enclaves, which means no complete private key is ever created and AEREDIUM holds no key to the owner's funds. The product supports Ethereum, Arbitrum and Base, keeps transaction records for seven years and is designed for Claude with openness to OpenAI and other AI frameworks.
An Institutional Security Background Behind the Products
Both products build on the threshold-signature technology and policy engine that AEREDIUM uses for institutional digital asset security, where a transaction has to satisfy predefined rules before it can be signed and every decision is recorded for later verification. On 2 September 2026 the company launched AERSeal, which uses its AERKey infrastructure to keep key shares in separate hardware-attested enclaves with the CGGMP24 threshold signing protocol so that no full private key exists in one place. Organizations set an M-of-N approval policy that signatories approve with passkeys.
The company says it holds United States Patent 12,694,407, granted on 28 July 2026, for a method in which AI models monitor stablecoins to assess whether each coin is backed by the reserves it claims. It operates from Melbourne, Tel Aviv and Austin, Texas.
The Outlook for Governing Autonomous Agents
The market is moving in the direction these products address. Gartner forecasts that 33% of enterprise software applications will include agentic AI by 2028, up from less than 1% in 2024. It also forecasts that 15% of day-to-day work decisions will be made autonomously by then, up from 0%.
The same forecast expects more than 40% of agentic AI projects to be canceled by the end of 2027 because of rising costs, unclear business value or inadequate risk controls. That last item is the one a policy engine with signed receipts is built to answer. AEREDIUM's launch extends its offering from digital asset custody and settlement into the governance of AI-driven financial and software operations. The pricing, the trial periods and the support for several coding agents give teams a low-cost way to test whether rules written in advance can keep pace with the agents they already run on aeredium.ai.
Don’t forget to like and share the story!
Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR.