The Impact of Maximal Extractable Value Strategies and Attacks in Proof of Stake Blockchains

Maximal Extractable Value (MEV) is often introduced as a purely technical artifact of transaction ordering, yet its real-world significance is determined elsewhere, in the execution pipeline that decides whether an opportunity is merely observable or can be reliably monetized. This dissertation studies MEV as an empirical, measurable phenomenon, and as an organizational outcome shaped by protocol rules, off-chain coordination, and an increasingly professional infrastructure that mediates how transactions reach blocks. The thesis addresses a simple but consequential question: what changed when Ethereum block production shifted from proof of work to proof of stake, and when proposer builder separation (PBS) and private order flow became commonplace. The central claim is that these shifts not only relocate power among roles, they also increase conversion efficiency, meaning that a larger fraction of theoretical opportunities becomes realized revenue, and that the identity of the beneficiaries becomes more legible in the data. To test this claim, the dissertation develops measurement pipelines across eras and links monetary flows to the economic relations among searchers, builders, validators, and users. On Ethereum, it quantifies how relay-mediated block building reshapes reward composition and the observed concentration of extraction, then it moves from prevalence to behavior, asking whether ordering under MEV infrastructure departs from a fee-based “neutral” baseline. By contrasting what priority should look like under simple price-time logic with what blocks actually exhibit, we surface systematic patterns consistent with preferential inclusion for specific submission channels and explicit payment paths. Beyond Ethereum, the dissertation connects these findings to the broader proof of stake landscape, using Solana as a motivating reference point for high-throughput execution and specialized MEV routing infrastructure, and framing cross-domain extraction as a sequencelevel phenomenon. The cross-chain component introduces an 𝑛-hop detector for sequencedependent arbitrage across swaps and bridge transactions, and empirically bounds multihop extraction over proof of stake systems and rollups, where bridging latency, confirmation rules, and fee schedules act as first-order feasibility constraints. A network perspective reframes MEV from isolated transactions to persistent relations. The dissertation constructs MEV interaction graphs and uses community detection algorithms to identify stable, role-structured clusters that emerge as statistical evidence of organized extraction without assuming the existence of “cartels” a priori. Overall, the dissertation offers a cross-era and cross-chain view of realized MEV, clarifying how modern proof of stake execution architectures can simultaneously improve efficiency and sharpen trade-offs among decentralization, fairness, and governance.