On August 19, the 'Encrypt the Mempool' call revealed a sobering truth: no known cryptographic primitive can satisfy all requirements for on-chain privacy at Ethereum scale. The admission came from the authors of EIP-8184, also known as LUCID, Ethereum's latest attempt to shield pending transactions from MEV bots.
This is not a breakthrough. It is a confession of a gap.
Context: The Public Mempool Problem
MEV extraction relies on visibility. Every pending transaction sits in the public mempool, exposed to bots that front-run, sandwich, or back-run. Current mitigations—private relays, Flashbots, VPN-like channels—work but introduce a trusted intermediary. The relay operator sees the transaction, can censor it, or leak it. The Ethereum community, scarred by ICO-era vulnerabilities, has long sought a protocol-level fix.
EIP-8184 (LUCID) proposes a commit-reveal scheme: builders submit sealed transactions, then reveal decryption keys after inclusion. EIP-8105 suggests a directed trust graph where providers whitelist each other. FOCIL (EIP-7805) adds a committee of attesters to constrain block builders. The three are intertwined, forming a roadmap for the Hegotá upgrade, targeted for 2027.
But the cryptographic glitch at the heart of LUCID is not a minor bug. It is a structural flaw.
Core: Code-Level Analysis of LUCID
LUCID's design is elegant in theory. A transaction is encrypted, sent to the mempool, and included in a sealed block. The block is revealed only after the builder commits. The key is then released by the sender or a key publisher. The encrypted block top segment is limited to one-eighth of the block gas limit, using a reservation fee. Successfully revealed transactions get the fee back; failures lose it.
Based on my audit experience in 2018 with 0x Protocol v2, I learned that elegant theory often hides expensive edge cases. The LUCID model assumes the key publisher is honest. If the publisher fails to reveal—due to malicious intent, network failure, or economic incentive—the builder loses the reservation fee, and the block may be invalid. There is no way to distinguish between an attack and a technical glitch. The protocol penalizes all failures equally.
This is not a trustless system. It is a trust-shifted system. The trust location moves from the relay operator to the key publisher. The publisher is a new central point of failure. The EIP-8184 authors acknowledge that no known construction meets all requirements: small public key, non-interactive decryption, no trusted setup, realistic ciphertext size, strong chosen-ciphertext security, and a quantum-safe path. That is a list of six constraints. Zero satisfied simultaneously.
Quantitative Impact of the 1/8 Gas Limit
LUCID restricts encrypted transactions to one-eighth of the block gas target. At current Ethereum block gas limits (~30 million), that caps encrypted throughput at ~3.75 million gas per block. For context, a single Uniswap V3 swap can consume 200,000 gas. That means fewer than 20 complex swaps per block can be encrypted. The rest remain exposed. This is a performance bottleneck that will force users to compete for scarce encrypted space, likely driving up reservation fees.
Compare this to private relays, which have no such gas limit but rely on a trusted operator. The trade-off is stark: performance vs. trust. LUCID chooses trust reduction at the cost of throughput. But the trust reduction is incomplete—the key publisher remains.
Contrarian: The Real Goal Is Not Privacy, It's Cost Elevation
The community frames these proposals as a war on MEV. The contrarian view: they are designed to raise the cost of MEV attacks, not eliminate them. The 1/8 gas limit and reservation fee make it expensive for bots to front-run at scale. The key publisher introduces a bottleneck that can be regulated by a whitelist (EIP-8105's trust graph). The real innovation is not cryptographic—it is economic. The protocol is building a toll booth for MEV.
But this creates a new blind spot. The trust graph in EIP-8105 outsources penalty enforcement to off-chain reputation. No on-chain slashing. No proof of misbehavior. This is a regression to the pre-smart-contract era of trust-based networks. Based on my stress testing of Curve Finance pools in 2020, I found that economic incentives alone cannot prevent insolvency during high volatility. The same applies here: off-chain enforcement fails when the stakes are high.
Furthermore, the Hegotá upgrade timeline (2027) places this squarely in the shadow of quantum computing. The Q-Day is estimated by some at 2029. If Ethereum waits until 2027 to deploy a non-quantum-safe encryption scheme, it will have two years before a potential existential threat. The authors explicitly require a 'credible path to quantum security,' but that path does not exist yet. The ledger remembers what the code forgot—and the code is forgetting to plan for post-quantum.
Takeaway: A Necessary Direction, Not a Solution
The encrypted mempool is a necessary direction. The private relay model is fragile. But LUCID, EIP-8105, and FOCIL are not ready for production. They are research proposals that trade one trust problem for another. The key publisher is a new central party. The gas limit is a performance cap. The quantum threat is a looming deadline.
Silence in the logs speaks loudest. The lack of independent peer review, the absence of a testnet, the admission of no known cryptographic construction—these are red flags that the market is ignoring. The real impact of these proposals will be on the MEV supply chain: private relays, builders, and searchers will need to adapt. But the adaptation will not happen overnight. It will happen in 2027, at the earliest.
Until then, the mempool remains a glass house. And the bots are still throwing stones.