Sealed-Bid Auctions

The Problem

Sealed-bid auctions are a much older and broader pattern than crypto. They're the default mechanism for government procurement contracts, real estate sales, spectrum license allocation, M&A bidding, and public sale / token launch auctions alike. The whole point of the format is that bidders submit one bid each, blind to what everyone else is offering, so the price reflects genuine valuation rather than reactive jockeying. The problem is that "sealed" is often more of a procedural promise than a technical guarantee. In paper-based or centralized processes, the auctioneer (or whoever handles the envelopes) still sees every bid before announcing a winner, leaving room for leaks, favoritism, or a trusted party simply making a mistake. Move the same auction on-chain and the problem flips but doesn't go away: bids sitting in a public mempool are visible to everyone before settlement, and even commit-reveal schemes eventually force every bidder to reveal their number in plaintext. Either way, something leaks — how much you were willing to pay, how aggressive your strategy was, where your true valuation sat — and that invites bid sniping, collusion based on visible patterns, and front-running by anyone watching the process.

Why FHE Fits This Problem Well

Sealed-bid auctions are a good match for FHE precisely because the underlying computation is simple and bounded: find the maximum (or second-highest, for Vickrey-style auctions) among a fixed, known number of encrypted values. Each bidder submits their bid encrypted under the protocol's public key, and a smart contract runs the comparison logic directly on ciphertexts. No bid is ever decrypted individually. Only the final result (the winning price, and which bidder won) goes through threshold decryption, handled by a distributed set of key-holders so no single party can unilaterally decrypt anything.

A Proven Cryptographic Pattern

The comparison-only computation pattern behind this, encrypt every bid, find the max (or second-highest) without decrypting any individual value, then reveal only the result, has a long research lineage in cryptography, well before FHE became practical enough for production use. FHE-based sealed-bid auction protocols, including second-price (Vickrey) constructions designed specifically to keep losing bids permanently secret while still proving the winning price correct, have been studied in the cryptography literature for over two decades. The same approach has also been proposed for spectrum allocation: bidders encrypt their bids under the auctioneer's public key before submission, so that channel assignment can be computed without the auctioneer or any third party ever seeing a bid in the clear. What's changed recently isn't the idea — it's that the underlying encryption schemes have gotten fast enough that "comparison among a bounded, known set of values" is now something you can actually run in a smart contract or a production auction system, rather than only in an academic prototype.