Widening the Design Space of AMMs with Solana
summary
Proposes a power-function AMM with a tunable parameter s that concentrates liquidity near mark price, implemented via iterative solvers (Newton-style) because the curve is nonlinear; notes a single power/log call normally consumes ~10% of a Solana tx compute budget (~$1 on mainnet) and introduces the Tachon math Library which reduces those costs ~10x by uploading precomputed tables and doing interpolation on-chain. Also outlines interconnected markets that share on‑chain average prices (zero‑copy accounts) to shrink arbitrage windows and lower impermanent loss.
AI-written summary · about
related talks
28mToo Much Liquidity? The State of Automated Market-Makers in DeFi
24mReimagining the AMM with 1inch Aqua
32mMaking DeFi Truly Cross-Chain
09mReflections on Solana From an EVM Builder
32mLiquidity: AMMs on Stellar
05mPerena: Infrastructure for Money
06mBuilding a Mutual Fund on Solana
22mSmart Markets in DeFi