GMX Protocol Deep Dive - DeFi Perpetuals Analysis

This was the major project for FINS5547 - the same course I later taught as a Casual Academic at UNSW. Writing the report as a student and then teaching the same material a year later gave me a completely different perspective on what actually matters when explaining DeFi mechanics.

The biggest thing I discovered: understanding a protocol's tokenomics is the same skill as understanding a company's capital structure. The GLP liquidity model, the esGMX staking mechanism, the fee split between LPs and token holders - these aren't "crypto" concepts. They're incentive design problems with the same underlying logic as any equity/debt structure. That realisation changed how I read both.

The V1-to-V2 upgrade was a useful case study in protocol risk management. V1's zero-price-impact design benefited large traders but created pool imbalances that became exploitable. V2's isolated GM pools fixed that - and the audit firm count jumping from 2 to 5 between versions was a clean signal of how much the risk surface had grown.

What I took from this: DeFi composability is a strength until it becomes an attack surface. Oracle dependency, regulatory exposure from anonymous founders, LP counterparty risk in bear markets - these aren't edge cases. They're structural features of how on-chain finance works, and identifying them is the real analytical work.