Daily 9 for November 11, 2026 covers nine developments across AI, fintech, crypto, software and tech, announced between November 9 and November 10. Google opened Gemini 4 Carbon to paying developers and AI Ultra subscribers, the Git project fixed six security flaws, and Coinbase’s national trust company received its final federal charter.
Each item links to the company, project or organization that published it.
1. Google releases Gemini 4 Carbon
Google made Gemini 4 Carbon, which it calls its strongest model, available after first giving it to a group of cybersecurity defenders through its Fairwind Program two weeks earlier. Google says Gemini 4 Argon remains the right choice for most coding and knowledge work, while Carbon spends more tokens on long, multi-step tasks. Paying developers can use it in Google AI Studio and Gemini for business at the same introductory price as Argon: $2 per million input tokens and $10 per million output tokens. It is also rolling out to Google AI Ultra subscribers in Antigravity, the Gemini app and AI Mode in Search. Google’s own benchmark claims include first place on CursorBench 4.0. Google, November 10.
2. OpenAI opens Businesses in ChatGPT in beta
OpenAI launched Businesses in ChatGPT in beta. Through a tool called Business Hub, a company can claim its listing, verify it, update key details and add product feeds with pricing and availability. Any eligible business located where ChatGPT is available can claim a listing, and Shopify merchants can connect their store directly. Verified businesses can also build a Business Agent that answers shoppers’ questions inside ChatGPT, starting with eligible e-commerce and online service businesses in the U.S. OpenAI says it plans to add more markets and categories later. OpenAI, November 10.
3. Rain puts funded vaults inside AI agents
Rain introduced Agent Vaults, which give an AI agent its own balance, spending rules and a Rain-issued card for each purchase. The owner funds the vault by bank transfer, card or stablecoins, and the agent can only spend what is already in it. Each purchase uses a scoped card that is retired afterward. Rain says the cards are accepted at 175 million merchant locations in 200 countries and territories, and that agents can pay out to local bank accounts in more than 80 countries. The product supports Visa Intelligent Commerce and Mastercard Agent Pay. Rain, November 10.
4. BioCatch flags AI agents in banking sessions
BioCatch launched AiRE (Agentic AI Risk Essentials), a feature that tells banks when an AI agent rather than a person is using an account, before any money moves. It covers agents acting for an accountholder and agents run by fraudsters. The signals include unusual interaction patterns such as synthetic inputs, the browser’s permission profile, agent infrastructure and whether the agent declares itself. Banks get it through their existing BioCatch integration. The company says close to half of the agent activity it detected in customer sessions happened at account opening. BioCatch, November 10.
5. Coinbase’s trust company gets its final OCC charter
The Office of the Comptroller of the Currency’s filing record lists Coinbase National Trust Company as consummated and effective on November 6, 2026, after preliminary approval on April 2. Coinbase announced the final approval on November 9. The charter lets it run its crypto custody business under a single federal framework. Coinbase Custody Trust Company has held a New York limited purpose trust charter since 2018, and Coinbase says the new charter builds on that work with the NYDFS. Coinbase, November 9, with the OCC filing record. Background: what a crypto wallet stores.
6. Helius launches Solana Privacy Rings on mainnet
Helius released Solana Privacy Rings, shielded pools that keep balances encrypted on Solana itself rather than on a separate network. In the permissionless Ring now live, the asset and the amount of each transfer are hidden, while wallet addresses stay visible, and only the user can decrypt a balance with a viewing key. Rings support SOL, SPL and Token-2022 tokens. Helius says three firms audited the programs and circuits: Certora, ZK Security and OtterSec. Decrypt, Ore and Aurora Intents are live on it. Custom Rings for institutions, with their own compliance rules, are still in audit. Helius, November 10, with the Solana Foundation’s note.
7. Git fixes six security flaws
The Git project released Git 2.56.1, with backports to 2.55.1, 2.54.1, 2.53.1, 2.52.1, 2.51.3 and 2.50.2. The most serious flaw, CVE-2026-111760, lets a malicious repository run code during a recursive clone: an integer truncation in index handling lets an overlong path pass a safety check and then be written under a shorter name. The other five cover path traversal through a server’s advertised HEAD, memory corruption in signature stripping and large merges, binary patches and fast-import streams. GitHub has patched github.com, and GitHub Enterprise Server fixes are in 3.19.15, 3.20.11, 3.21.9 and 3.22.4. GitHub, November 10.
8. Gemini Intelligence reaches the Pixel 10
Google’s November Pixel Drop brings Gemini Intelligence to the Pixel 10 series. It includes Rambler, a voice-typing mode that removes filler words and formats what you say into clean text. Gemini’s task automation, which can order food, book rides or find restaurant reservations across apps, expands to the UK, India and Japan. Create My Widget, which builds a home-screen widget from a written description using Gmail, Calendar and Search, rolls out over the next few weeks. Motorola’s first Gemini Intelligence phone, the signature 27, opens for U.S. pre-order on November 19. Google, November 10, with the Pixel Drop post. Background: on-device vs. cloud processing.
9. MemryX unveils the MX3+ AI accelerator
MemryX announced MX3+, the next version of its MX3 inference chip. The company says it will deliver up to 10 times the peak AI compute and nearly 10 times the on-chip memory of MX3 within the same low-power envelope. Like MX3, several chips can work as one device, and a 16-chip configuration is designed for more than one peta-operation per second and 2.16 GB of distributed on-chip memory. MemryX targets larger vision, transformer and small language models at the edge. Engineering samples are targeted for late 2027 and volume production for mid-2028. MemryX, November 10.







