Builder Spotlight: @Spiko_finance Spiko has raised a $90M Series B led by NEA to bring cash yield onchain. The third billion in funds arrived in under four months. The team is building regulated cash funds onchain for businesses and fintechs. Base keeps access fast and cheap. ■ $3.1B in AUM, growing 13.7% month over month ■ 6,500+ businesses keep cash in these funds ■ Close to $10B has moved through the platform ■ Funds in EUR, USD, GBP and CHF Onchain cash management is scaling fast on Base.

Builder Spotlight: @longbowlend Your stocks can borrow for you on Robinhood Chain. @longbowlend is the credit layer for Robinhood Chain. Users lock tokenized stocks, crypto or RWAs and borrow USDG without selling. 63 markets are live, with $4.3M supplied and 74.6% utilization. 15 Ondo stocks just joined, from NVDAon to SPYon. How it works: > Curated USDG and ETH vaults earn interest from borrowers > Isolated markets settle on @Morpho > Fixed and variable rate loans, including multi-collateral fixed credit > Short on Longbow lets users borrow ETH against USDG > An MCP server lets any AI agent build Longbow transactions Longbow Cards powered by @Stripe are coming next, so borrowed USDG goes straight to spending.

Builders landing in Floripa for @islanddao deserve a proper Brazilian welcome 🇧🇷🔥 @contavcbr, Nora & @cloak_ag are bringing everyone together for a churrasco: great food, good company, and conversations beyond the timeline. See you by the grill:

Builder Culture is built through tangible contributions A working version that users can experience A piece of reusable code Documentation that helps newcomers get started A bug report with steps to reproduce the issue An update that responds to user needs These concrete efforts form the foundation for the continuous growth of the ecosystem

Builder’s Pick this week: context-mode 👀 An open-source context optimization tool for AI coding agents. Instead of dumping every file read, shell command, GitHub issue, or MCP result directly into the model’s context window, context-mode keeps the raw output outside the conversation and only brings back what the agent actually needs. It also keeps a searchable local record of session activity, so long-running coding sessions can recover relevant context without reloading everything again. The project currently supports tools including Claude Code, Cursor, Codex, Copilot and other agent environments. What makes it interesting is that it tackles a very unglamorous problem that becomes obvious once you actually use coding agents for a while: the agent may be getting smarter, but the context window still fills up. GitHub: #Builders# #OpenSource# #AICoding# #DevTools# #OpenBuild#