Flying Tulip Weekly Update - May 29 This week: ftUSD became available as collateral on Lend, BTC borrow costs were reduced on Sonic and Ethereum, and Sonic Lend borrowing parameters were updated to make borrowing more accessible. ftUSD supply and Margin Lending TVL continued to grow, with ftUSD APY on Sonic increasing to 4.57% — all without incentives, points, or leverage-driven farming, and with instant redemptions. More stats in reply 👇

In Web3, reliability is non-negotiable. As a Premier @GoogleCloud Partner, BCW Group worked with @0xSonicLabs to architect a "secure-by-design" foundation that puts AI at the center of development. The Highlights: 🔹 Spawn App Builder: Concept to dApp in 15 mins via Gemini models. 🔹 Total Reliability: High-availability operations powered by Google Cloud. 🔹 Data at Scale: Real-time analytics with BigQuery & GKE. 🔹 Internal Efficiency: Gemini 1.5 Pro automating tech support workflows. The future of dApp growth is AI-augmented. 🚀 Full Story: cloud.google.com/customers/s… @gcloudpartners

Respect for the transparency. You’re hitting on challenges we’ve seen businesses at your scale run into when real-world traffic meets the blockchain. Our team has been heads-down with leaders across the space on how an L1 can fine-tune the execution environment around what serious apps actually need: settlement certainty, predictable costs, and business-specific tooling. Polymarket is the clearest example of a system that needs web2 reliability on web3 rails. Quick thoughts from the chain side: Latency: Sonic finalizes in ~500ms, deterministically. No probabilistic confirms, no reorgs. Your “near-instant” target is already reality on an L1 stack. ~10k TPS sustained, sub-second single-slot finality, ~$0.0001 average fees. Chain migration: Doors are open, and we’re actively making bespoke improvements for serious teams. Most chains slow down as throughput climbs; Sonic is designed to get faster under high load. Cancelled transactions: At your scale, this is a throughput and execution problem. Sonic’s headroom removes the congestion vector. Deterministic finality removes the reorg vector. Your CLOB shouldn’t be dropping orders because the chain can’t keep up. Cheap gas, instant finality, battle-tested uptime, and native USDC are default. Prediction markets need reserved capacity for market makers, gasless user flows, protocol-native atomic batching, and app-critical transactions that still land when volume spikes. Would love to compare notes with your tech team on what Polymarket’s real-world throughput has taught you, and where we can build to support it. DMs open 🫡