a16z Crypto Unveils Aerie to Accelerate Post-Quantum Signature Aggregation

Venture firm a16z crypto released Aerie on Oct. 16, a lattice-based SNARK designed to aggregate post-quantum signatures at speeds exceeding 50,000 per second on a single laptop ¹. The framework presents an alternative to the hash-based post-quantum cryptography designs currently pursued across major proof-of-stake networks ¹.

Proof-of-stake blockchains rely on signature aggregation to manage large validator sets ¹. While networks like $ETH use BLS signatures to compress tens of thousands of validator approvals into single group elements per slot, BLS lacks quantum resistance ¹. Post-quantum signatures are between 10 and 100 times larger than legacy elliptic-curve signatures, creating potential throughput bottlenecks and raising node operating expenses without effective aggregation ¹. In benchmark tests conducted on an Apple M4 Max laptop without GPU acceleration, Aerie aggregated 32,768 Falcon-512 signatures in 0.65 seconds, compressing the data down to roughly 42 bytes per signature with a single-core verification latency of 54 milliseconds ¹.

Challenging Hash-Based Approaches

The prevailing approach in the Ethereum post-quantum cryptography roadmap pairs hash-based signatures with hash-based SNARKs, relying on the leanVM zero-knowledge virtual machine to recursively merge proofs in batches of approximately 900 signatures ¹. That design targets 1,000 signatures per second for 2026 ¹. According to a16z crypto, recursive proving adds systemic bug risk and leaves proof security unestablished under the random oracle model ¹. By utilizing Akita—a polynomial commitment scheme developed alongside LayerZero Labs—and sum-check techniques, Aerie enables a single node to generate an entire slot's proof directly without intermediate recursion ¹.

Aerie targets Falcon, a lattice-based signature scheme selected by NIST for standardization under FN-DSA ¹. While Falcon previously faced implementation hurdles on hardware wallets due to floating-point math and memory footprints, recent refinements have reduced Falcon-512 signing requirements to 11 kilobytes of RAM ¹. The research team indicated ongoing development to support ML-DSA and hybrid post-quantum signatures, as networks like Solana evaluate Falcon adoption and Algorand maintains it in production ¹. At the time of writing, $BTC traded at $81,424 and $ETH stood at $2,435.