Robinhood Launches AI Trading Protocol with 150K Users

Key Takeaways

• Robinhood launched Robinhood MCP, enabling AI agents to execute trades, with 150,000+ agentic trading accounts and 30 million daily tool calls
• Amazon Marketplace uses Amazon SageMaker AI to review A+ Content, driving 8-20% sales gains for brands with high-quality listings
• Coherent reported 79% revenue growth to $1.95 billion in datacenter communications, with capex surging 7.5x to $775.5 million
• OpenAI backed a data center policy framework and committed a $4.8 million grant in Georgia and a $6 million childcare investment in Ohio
• Brainbase Labs improved a Domino's pricing agent from 41% to 97% accuracy across 11 versions, cutting runtime from 129 to 71 minutes
• Canva restructured, cutting 1,271 jobs (22% of workforce), but denied AI is replacing employees
• Aina unveiled an AI ring running a Context Aware Operating System that predicts users' next three actions
• Capgemini was chosen to lead an EU AI security consortium modeling cyberthreats under the EU's Artificial Intelligence Act
• Nous Research released Hermes Agent, a self-improving AI agent, while OpenAI promoted Nemotron 3.5 Lightning for local AI use
• AI-generated audio was used to fake comments from Indian Lieutenant General Girish Kalia, detected by Hiya voice tool

Robinhood launches AI agent trading and blockchain stock tokens

Robinhood launched Robinhood MCP, a protocol that lets AI agents execute trades for customers. More than 150,000 customers have opened agentic trading accounts, generating 30 million tool calls per day. The company also relaunched stock tokens alongside Robinhood Chain, a blockchain for real-world assets, now live in over 120 countries. Robinhood was selected as the sole initial trustee for Trump Accounts, which will seed $1,000 into an S&P 500 ETF for every child born from January 2025 onward.

Nous CEO warns AI labs will trap or replace users

Nous CEO Dillon Rolnick said AI labs will either trap users or replace them by owning the user layer. He noted that token prices are falling while spending surges, forcing labs to build their own solutions or control the market. Rolnick described agents as operating systems combining code and data with reasoning. Nous Research released Hermes Agent, a self-improving AI agent listed on GitHub, and the company has 116 repositories.

Aina reveals AI ring that predicts your next actions

Aina, a San Francisco startup, unveiled an AI ring on November 5 after 15 months in stealth. The ring runs a Context Aware Operating System that predicts your three most likely next actions based on data from your phone, Mac, and agents. It works in three modes: Voice for agent prompts, Cues for proactive predictions, and Remote for live controls like music or meeting tools. The ring includes a screen, mic, speaker, and joystick, and pre-orders opened on November 5.

Figma finance team uses AI to replace manual data tasks

Figma's finance team moved away from 400 to 500 line SQL queries and Excel macros to use modern AI tools for data orchestration. The team compressed weeks of analysis into about an hour, catching anomalies early and saving hundreds or thousands of dollars. Finance staff now focus on judgment work like running experiments and pricing studies instead of filing tickets. One growth-stage team cited in the article spun up 40 to 50 agents in one to two months, though 90 percent are no longer used.

Brainbase Labs boosts Domino's pricing agent to 97 percent accuracy

Brainbase Labs improved a Domino's pricing agent from 41 percent accuracy to 97 percent across 11 versions. The agent monitors competitor prices across Domino's 20,000 plus stores in 90 plus markets. At version 1 the agent ran 129 minutes per run at $2 per source, and at version 11 it ran 71 minutes at $1 per source. Brainbase Labs built a six-step process including kickoff, benchmarks, iteration, and handoff, where analysts review only exceptions. The company plans to expand to additional Domino's markets.

Delaying AI builds may beat rushing, experts say

Blackstone CTO John Stecher and Norm AI CTO Scott Worland say waiting six months to build can be better than building now. They recommend investing in evaluations, data plumbing, and human approval gates first. Norm AI learned this lesson when earlier code became obsolete after agentic approaches arrived. Both leaders argue for flexible conduits rather than fixed plans.

OpenAI proposes new rules for data center communities

OpenAI backed a new policy framework shared with POLITICO that gives communities more say in data center development. The proposal asks utility providers to make large customers cover added grid costs and allows throttling power during high demand. Developers would face public reporting on water use, energy use, noise, and job openings.

OpenAI sets responsible AI infrastructure standards

OpenAI published a compact and policy blueprint on November 5, 2026 for responsible data center development. Future campuses will have local compacts with yearly public reporting on energy and water use. OpenAI commits to paying full energy costs, protecting local resources, and creating local job opportunities. Examples include a $4.8 million grant in Georgia and a $6 million childcare commitment in Ohio.

AI audio faked Indian general's border criticism

After an Indian official accused China of challenging border claims in October 2026, a fake video circulated with AI-generated audio. The video falsely claimed Lieutenant General Girish Kalia criticized India's border response. AFP found the footage was from an older press conference about a different topic. The Hiya voice detection tool flagged the audio as synthetic.

Fake AI audio attributed to Indian general

A manipulated video showed Lieutenant General Girish Kalia appearing to criticize India's handling of border tensions with China. His actual remarks were about an operation in Jammu and Kashmir from years earlier. The audio sounded unnaturally flat and did not match mouth movements. An AI detection tool confirmed the soundtrack was synthetic content.

Capgemini leads EU AI security and cyberthreat consortium

Capgemini was chosen by the European Commission's AI Office to lead a consortium focused on AI security. The project will model cyberthreats and assess risks from advanced AI models under the EU's Artificial Intelligence Act. Capgemini will work with Wavestone and French institution LNE to build assessment tools and compliance methods. The goal is to help Europe monitor and reduce systemic risks linked to AI-driven cyberattacks.

Canva says AI is not replacing employees despite global cuts

Canva announced a global restructuring affecting 1,271 workers, or 22 percent of its workforce. The company denied that AI is replacing employees who are leaving. A spokesperson said AI is changing the products and skills needed, not substituting for workers. The changes follow slower growth and a shift in how Canva organizes teams, including its India content operations.

Amazon uses AI to judge product content quality

Amazon Marketplace uses Amazon SageMaker AI to evaluate A+ Content on product pages. Selling partners with high-quality content see sales gains of 8 to 20 percent. The system uses Vision-Language Models to review the full page as one piece rather than checking images and text separately. It then gives brands actionable feedback to improve their product storytelling.

Canadian tech stocks fall as AI sell-off spreads

Canada's tech sector shows a split as a global retreat from AI stocks hits chip and hardware companies. Celestica fell 11.43 percent in one day and is down about 54.9 percent from its 52-week high. Software and telecom stocks like BCE held up better as investors moved toward steadier businesses. The shift reflects concern over high AI spending and risky valuations.

Video games may help train robots

A new idea suggests video games could be a cheap way to train robots. Pim de Witte says game data provides billions of action examples each year, helping robots learn fast adjustments. The approach uses gamer controllers and pixel-to-action learning instead of rare human teleoperation. KraneShares AI ETF AGIX has invested in this method, though some experts note it still lacks full physical safety controls.

Open AI Model Nemotron 3.5 Lightning Lets Anyone Run AI Locally

Alexiuk, known as llm_wizard and Joe Nemotron, is promoting Nemotron 3.5 Lightning as an open AI model anyone can run. He believes openness in AI means providing access to weights, data, and training recipes. Nemotron 3.5 Lightning is designed for repetitive, high-frequency tasks, while Nemotron Ultra focuses on long-running agents and research. The model emphasizes local use, putting pressure on hardware to handle memory and latency without relying on the cloud.

Coherent Capex Surges 7.5x as Datacenter Sales Jump 79%

Coherent reported that its Datacenter and Communications revenue rose 79% to $1.95 billion. Its capital expenditure also surged 7.5 times to $775.5 million, which is about a third of its revenue. The company, which makes optical transceivers for AI data centers, shared these results in early November 2026. This spending growth highlights the strong demand for connectivity equipment in AI data centers.

Sources

• How Robinhood Wants Agents to Trade For You
• Nous CEO says labs will trap you or replace you
• Aina Built a Ring to Approve What AI Guesses
• Figma finance team stops filing tickets, builds
• Brainbase Labs pushed Domino's agent to 97% accuracy
• Waiting for the next model is the strategy
• OpenAI lays out new policy framework for data centers
• Setting a new standard for responsible AI infrastructure
• AI-generated audio misrepresented as Indian general's criticism over border spat with China
• AI-generated audio misrepresented as Indian general's criticism over border spat with China
• Capgemini selected by Brussels for AI security project
• Canva denies AI is replacing departing employees as global restructuring impacts India content operations
• How Amazon Marketplace evaluates content quality with Amazon SageMaker AI
• This Canadian tech stock is getting crushed as AI sell-off spreads across the mark
• Video Games Are About to Train Your Robot
• One wizard is selling Nemotron 3.5 Lightning as open AI anyone can run
• Coherent Capex Hit $776M, 7.5x, as Datacenter Sales Rose 79%