Economic Security in Proof-of-Stake

Economic security is the primary metric for evaluating the security of proof-of-stake blockchains. Unlike proof-of-work (where security is measured in hash rate and electricity cost), PoS security is denominated in staked capital.

**The attack cost framework:**

In Ethereum's Casper PoS: - A 51% attack (controlling block production) requires 51% of staked ETH - A 33% attack (preventing finality but not controlling the chain) requires 33% of staked ETH - An attack that could corrupt finality requires ~66% of staked ETH

With ~30M+ ETH staked (as of 2024–2025), acquiring 33% would require ~$50B+ at market prices — plus the market impact of the acquisition itself (buying that much ETH would dramatically increase the price before the attacker held enough).

**Economic security vs. cryptographic security:** Cryptographic security is binary: a hash function is secure or broken. Economic security is continuous: the attack becomes more or less expensive as the staked amount and token price change. A bear market that drops ETH 80% also reduces economic security 80% — a vulnerability that pure cryptographic approaches don't have.

**Slashing as a security multiplier:** In Ethereum, validators who behave maliciously (double-signing, equivocating) lose their staked ETH through slashing. This means an attacker not only needs to acquire billions in ETH but risks losing much of it in the attack. The staked capital is at risk, creating a 'skin in the game' deterrent.

**The validator count dimension:** More validators = more decentralized security, but only if validators are independent. Correlated validator failures (same cloud provider, same client software) can compromise economic security without requiring a direct stake attack. Client diversity and geographic distribution are key non-capital security factors.

**Restaking and security dilution:** EigenLayer and similar restaking protocols allow staked ETH to secure multiple services simultaneously. This leverages ETH's economic security across more applications — but if stakers overcommit (slashing conditions on multiple services), cascading slashing could threaten the base layer's security.