Keynote lecture
Why blockchain scalability requires new and better hardware

Blockchains are rather slow computers, but they are very trustworthy and hence carry immense financial and social value. One of the most interesting ways to increase their scale, and also improve the privacy of transactions over them, uses advanced mathematical and cryptographic proofs known as ZK-STARKs. This technology has already been used to settle upwards of $800B on top of Ethereum, and forms the basis of Starknet - a new extension of Ethereum that offers exponentially greater scale with no reduction in trust and/or security.

Generating ZK-STARK proofs requires immense computational resources, needed to compute algebraic computations over finite fields. These computations are modular arithmetic and finite precision versions of the Fast Fourier Transform (FFT), executed over huge and ever growing domains.

This talk will survey the fascinating world of blockchain, their scalability challenges and the types of computations and hardware needed to scale blockchains using ZK-STARKs.