Coinbase is expanding its AI agent trading platform with a new 24/7 feature powered by Deribit, enabling automated trading across equities, crypto, and derivatives. Eligible users can connect their accounts to AI platforms and access perpetual contracts, stock indexes, precious metals, and energy contracts. CEO Brian Armstrong said the company is investing in infrastructure for automated trading, with agentic trading volume growing roughly 20% weekly. Spot trading volume, however, fell 10% quarter over quarter in the third quarter of 2026.
Chainalysis disclosed a data breach earlier this month after detecting an external attacker that used AI-driven agentic techniques to access its systems. The breach exposed user IDs, full names, email addresses, and organization role labels. Chainalysis said investigation content, transaction results, passwords, and authentication secrets were not accessed. The company is patching vulnerabilities and working with external cybersecurity experts and large language models to review and strengthen its defenses.
DoorDash CEO Tony Xu said AI agents have generated several hundred orders through third-party integrations, but cautioned against overhyping the technology at this stage. The company is exploring new tools and ways to collaborate with AI-driven ordering systems. Xu warned against being delusional about AI capabilities after only a few hundred orders, signaling that adoption remains in its early phases on the DoorDash platform.
Tencent launched an AI-powered retail robot at a 7-Eleven store in Shenzhen, powered by its Hunyuan AI models. The robot greets customers, recommends products and promotional bundles, and helps place orders through conversation. Customers complete payment by scanning a code. The system uses Tencent's audio technology to recognize speech in noisy environments. WeChat Pay said the robot handles routine interactions so employees can focus on restocking and assisting shoppers.
Pennymac now uses AI agents to automate about 40% to 45% of its backend fulfillment operations, with another 15% to 20% in development. The company previously achieved 30% to 50% efficiency gains from basic automation before adding AI agents. The technology has cut conditional approval timelines in half. Pennymac trains its AI agents like new hires, starting with full human oversight and daily deployments, and aims to reach over 80% automation in fulfillment.
Investing.com released a new set of AI-picked stocks for November, evaluating thousands of stocks using over 15 years of financial data across more than 150 quantitative models. Previous strategies posted strong returns, including Mid-Cap Movers at +115.17% and Energy Elite at +81.83%. The AI explains its reasoning behind each pick, and monthly strategies are rebalanced with new opportunities while removing stocks that no longer meet criteria.
Research shows hedge funds using generative AI earned 2 to 4 percentage points more per year in abnormal returns, though non-hedge firms saw no similar benefit. Some AI trading strategies failed due to contaminated data or hidden costs like price impact. Nasdaq's Dynamic M-ELO order type showed better execution results by weighing over 140 parameters. The findings suggest that trading with real money still requires validated data beyond what automated tools alone can provide.
Cisco CEO Chuck Robbins urged partners to take a leading role in the AI shift during the Cisco Partner Summit 2026. He said partners help customers manage complex IT, budget limits, and infrastructure refreshes. The Cisco 360 program rewards partners for delivering value and building recurring revenue. Robbins highlighted networking, security, silicon, optics, and Splunk as key strengths and encouraged partners to represent full architectures to customers.
Qodo built a plugin for Codex that integrates organizational knowledge and codebase standards into AI code review. The tool helps developers catch critical P0 bugs across many repositories, even when services depend on each other. It uses OpenAI models including GPT-6 Astra to gather context, review code, and monitor quality. Qodo also introduced a Decisions API to choose which coding rules apply at each development stage, aiming to automate review while keeping human judgment on unresolved issues.
Key Takeaways
- Coinbase launched 24/7 AI agent trading via Deribit, with agentic trading volume growing ~20% weekly while spot trading fell 10% QoQ in Q3 2026.
- Chainalysis suffered an AI-driven cyberattack exposing user IDs, names, emails, and role labels, but passwords and authentication secrets were not compromised.
- DoorDash CEO Tony Xu said AI agents have generated several hundred orders and the company is exploring new tools, but adoption remains in early stages.
- Tencent deployed an AI retail robot powered by Hunyuan models at a 7-Eleven in Shenzhen during the opening of Guangdong's 2,000th location on October 29.
- Pennymac uses AI agents to automate 40-45% of backend fulfillment, cutting conditional approval timelines in half, with a goal of over 80% automation.
- Investing.com's AI stock picks for November use 150+ quantitative models on 15+ years of data, with past strategies returning up to +115.17%.
- Hedge funds using generative AI earned 2-4 percentage points more annually in abnormal returns, while non-hedge firms saw no benefit from the same technology.
- Cisco CEO Chuck Robbins urged partners to lead in the AI era at the Partner Summit 2026, highlighting networking, security, and Splunk as key strengths.
- Qodo built a Codex plugin using GPT-6 Astra to catch critical P0 bugs across repositories and introduced a Decisions API for coding rules at each development stage.
- Chainalysis recommends users enable multi-factor authentication while the company patches vulnerabilities and works with external experts to strengthen defenses.
Coinbase launches 24/7 AI agent trading with Deribit
Coinbase introduced 24/7 trading for AI agents powered by Deribit. Eligible users can connect their accounts to AI platforms for advanced trading strategies. The feature supports equities, crypto, and derivatives trading. Users can also make micropayments for on-demand data and trading strategies.
Coinbase AI trading expands with Deribit derivatives
Coinbase now lets eligible traders use AI agents for global derivatives through Deribit. CEO Brian Armstrong said the company is investing in infrastructure for automated trading across multiple asset classes. Agentic trading volume is growing approximately 20% weekly. Spot trading volume fell 10% quarter over quarter during the third quarter of 2026.
Deribit opens perpetual contracts to AI agents via Coinbase
Deribit announced that its perpetual contracts are now accessible to AI agents through the Coinbase for Agents platform. Eligible users can link their accounts to supported AI agents for research, structuring, and executing trades. The integration also covers stock indexes, precious metals, and energy contracts. Deribit spot execution connects to Coinbase Exchange liquidity for supported instruments.
Chainalysis investigates AI-driven cyberattack on its systems
Chainalysis detected an external attacker using AI-driven agentic techniques to access its systems this month. The attacker reached user IDs, full names, email addresses, and organization role labels. Investigation content, transaction results, passwords, and authentication secrets were not accessed. Chainalysis is patching vulnerabilities and working with external cybersecurity experts and large language models to review its security.
Chainalysis reveals data breach from AI-driven cyberattack
Chainalysis disclosed a data breach caused by an AI-driven cyberattack this month. The attack exposed user IDs, full names, email addresses, and organization role labels. The company confirmed that investigation content, transaction results, and passwords were not compromised. Chainalysis is working with external experts and large language models to strengthen its defenses and recommends users enable multi-factor authentication.
Investing.com launches new AI stock picks for November
Investing.com released a new list of AI-picked stocks for November. The AI system evaluates thousands of stocks using over 15 years of financial data across more than 150 quantitative models. Previous strategies showed strong returns, including Mid-Cap Movers at +115.17% and Energy Elite at +81.83%. The AI also explains its reasoning behind each stock selection to help members make decisions. The monthly strategies get rebalanced with new opportunities and removed stocks that no longer meet criteria.
DoorDash CEO says AI orders still limited but company explores new tools
DoorDash CEO Tony Xu said the company has seen several hundred orders from third-party AI agents. He warned against being delusional about AI capabilities after just a few hundred orders. The company is exploring new tools and ways to work with AI-driven ordering systems. AI agents are still in early stages of generating orders on the DoorDash platform.
Tencent deploys AI sales robot at 7-Eleven store in Shenzhen
Tencent's WeChat Pay launched an AI-powered retail robot at a 7-Eleven store in Shenzhen. The robot debuted during the opening of Guangdong's 2,000th 7-Eleven location on October 29. Powered by Tencent's Hunyuan AI models, the robot greets customers, recommends products and promotional bundles, and helps place orders through conversation. Customers scan a code to complete payment themselves. The system uses Tencent's audio technology to recognize speech in noisy environments. WeChat Pay said the robot handles routine interactions so employees can focus on restocking and assisting shoppers.
Pennymac uses AI agents to automate 40% of backend operations
Pennymac now uses AI agents to handle about 40% to 45% of its backend fulfillment operations. Another 15% to 20% is in development. The company first saw 30% to 50% efficiency gains from basic automation before AI agents were added. AI agents have cut conditional approval timelines in half. Pennymac trains AI agents like new hires, starting with 100% human oversight and daily deployments. The company aims to reach over 80% automation in fulfillment while moving employees to more complex tasks.
How to let AI supervise a trading book without picking trades
This article describes a system where AI supervises a trading book but does not choose trades. The trading rules are fixed in advance and stored in a public GitHub repository. The AI model only watches the orders and can halt the book if it detects data errors. The system uses an Interactive Brokers paper account so no real money is at risk. It runs on a Mac using IB Gateway kept alive by IBC and open-source tools. The agent logs in automatically each morning and operates on its own schedule.
AI gives hedge funds a small trading edge but data risks remain
Research shows hedge funds using generative AI earned 2 to 4 percentage points more per year in abnormal returns. Non-hedge firms saw no benefit from the same technology. Some AI trading strategies failed because of contaminated data or hidden costs like price impact. Nasdaq's Dynamic M-ELO order type showed better execution results by weighing over 140 parameters. Trading with your own money still requires real data and validation that automated tools alone cannot provide.
Cisco CEO urges partners to lead in AI era
Cisco CEO Chuck Robbins spoke at the Cisco Partner Summit 2026 about the importance of partners in the AI shift. He said partners help customers manage complex IT, budget limits, and infrastructure refreshes. The Cisco 360 program rewards partners for delivering value and building recurring revenue. Robbins highlighted networking, security, silicon, optics, and Splunk as key strengths. He encouraged partners to stay agile and represent full architectures to customers.
Qodo adds codebase context to AI review with Codex
Qodo built a plugin for Codex that brings organizational knowledge and codebase standards into AI code review. The tool helps developers catch critical P0 bugs across many repositories, even when services depend on each other. It uses OpenAI models including GPT-6 Astra to gather context, review code, and monitor quality. Qodo also introduced a Decisions API to choose which coding rules apply at each development stage. The goal is to automate review while keeping human judgment on unresolved issues.
Sources
- Coinbase Unveils 24/7 Trading for AI Agents, Driving User
- Coinbase AI Trading Brings 24/7 Derivatives via Deribit
- Deribit announces AI agent access to perpetual products through Coinbase for Agents
- Chainalysis Investigates Security Incident Involving AI-Driven Attack Techniques
- Chainalysis Discloses Data Breach Following AI-Driven Cyberattack
- +115, +82%, 70% and counting: New AI-picked stocks for November are NOW LIVE
- DoorDash sees limited AI orders, but explores new tools
- Tencent’s WeChat Pay puts AI sales robot to work in 7-Eleven store · TechNode
- How Pennymac uses AI agents to automate 40% of its backend fulfillment operations
- Don’t Let the AI Pick the Trades. Let It Run the Book.
- Can Artificial Intelligence Improve Trading Decisions?
- Cisco CEO Chuck Robbins discusses partner opportunities in the AI era
- Qodo builds context into AI code review with Codex
Comments
Please log in to post a comment.