DSS 2026 — Quantum-Safe Is Not Just Post-Quantum Signatures Abhinav Sharma Source video: https://www.youtube.com/watch?v=ovoTmF01nzQ Automatic captions — not human-reviewed. Complete spoken-audio coverage is unverified. Source language tag: en (English). Source export SHA256: 0cd7f453dc15418cf0e8c9928003efb65cb4fa8c588118d2574a9bb5e59d1288 [0:00] Perfect. [0:01] So, [0:02] hello everyone. My name is Abhinav. [0:04] I'll be talking something other than AI. [0:08] Of course, why not? [0:10] I'll be talking about how quantum safe [0:12] is not just about post quantum [0:14] signatures [0:15] and we have five more layers including [0:19] that that we need to take care of. [0:21] So, [0:23] the general reflex around [0:25] securing ourself from [0:28] post quantum world. [0:30] We only take care [0:33] of one failure mode which is just that [0:35] will swap the signatures with quantum [0:37] proof signatures. [0:38] But [0:39] I want to make sure that there are four [0:41] more layers that we could get our [0:43] awareness on and what are they and what [0:45] we need to take care of. [0:48] The modern cryptography relies on one [0:51] assumption which is that certain [0:53] mathematical problems [0:55] are very hard to solve [0:58] so that even if we have enormous [1:00] computing power [1:01] the goal to get the right answer is [1:04] practically infeasible. [1:05] And [1:06] quantum computing challenges this [1:08] assumption by saying hey [1:10] what if we change the architecture of [1:12] the computation itself? So, instead of [1:14] representing our information in the bits [1:16] of zeros and one what we'll do is we'll [1:19] use something called as qubits and this [1:21] qubit represents [1:22] a simultaneous state [1:25] by using a principle called as [1:27] superposition. I won't go into the [1:28] details due to the lack of time, but [1:31] current estimate says that we need [1:32] around 1200 logical qubits and around [1:35] less than 9 minutes to crack [1:38] one private key [1:39] and less than 9 days to crack top 1000 [1:43] Ethereum accounts. [1:46] Basically decimation. [1:49] There are five layers that we need to [1:50] take care of. [1:52] The first one is ownership that we all [1:53] know which is what? [1:55] Cracking your private key. So, each one [1:57] of us must have a wallet address that [1:59] wallet address is nothing but a public [2:02] key. [2:03] And quantum computers can break your [2:05] private key by using that public key. So [2:07] when you're transacting, you're exposing [2:09] your public key. [2:11] When I'm talking about ownership, I'm [2:12] not just talking about these signatures, [2:14] but also the admin accounts that are [2:17] being used to transact multi-billion [2:20] cryptos from one point to another. [2:23] And so the fix here will be to focus [2:25] more on account abstraction that makes [2:27] ownership much more flexible and key [2:30] rotation followed by quantum proof [2:32] signatures. [2:33] The second one will be the code base. So [2:35] in order to stay under the gas limit, [2:38] smart contracts [2:40] give their heavy lifting to something [2:42] called as Nates pre-compiles, which is [2:44] EC recover, we have BLS 12 pairing and [2:47] many more. [2:48] These are also dependent on the elliptic [2:51] curve cryptography, which means that if [2:53] I can break them [2:55] or an attacker who has a quantum [2:56] computer can break them, [2:58] the impact can become that I can just [3:00] forge a proof on a rollup, an invalid [3:03] proof. [3:04] So the fix here will be to use [3:05] post-quantum pre-compiles and slowly [3:08] move forward to the lattice-based [3:09] designs. [3:11] Third one, which is the consensus. So [3:13] validators are using the BLS key [3:16] which are [3:18] curve friendly [3:19] and [3:20] if an attacker can sort of get their [3:23] hands on cracking the validator key, [3:26] they can forge attestations at a large [3:29] scale. We need [3:30] two-thirds of the majority and believe [3:32] me, this is not about stealing funds. It [3:35] is about the network's ability to agree [3:37] on what is true. [3:39] The fix here will be to harden [3:41] the validator key rotation and [3:43] post-quantum signature aggregation as it [3:45] matures. [3:46] Data availability. So [3:48] rollups [3:51] they upload, for lack of better words, [3:53] their data [3:55] and these validators do not just extract [3:58] it and verify it. They are dependent on [4:00] something called as KGZ commitments. [4:02] These commitments are essentially the [4:04] proof that what is being transported is [4:07] actually what we want. [4:10] The [4:11] quantum attacker can basically forge [4:14] these since KZG commitment also uses [4:17] elliptic curve cryptography. [4:19] If this goes [4:21] of the if this breaks, that means I can [4:24] forge the proofs that are invalid in [4:25] nature [4:26] and that means the entire roll up [4:28] infrastructure needs to [4:30] be quantum safe. [4:33] The fifth and I think the [4:36] most underestimated one is the privacy. [4:39] So all of these protocols that [4:41] essentially say that they are [4:42] anonymizing the user [4:44] it breaks into two parts. First is the [4:46] key exchange that is responsible for [4:49] sharing the secret between two users. If [4:51] this goes down, that means we can [4:54] deanonymize a user followed by checking [4:57] all of this his old transactions as [4:59] well. [5:00] The second is the same called as [5:02] pedersen commitments which is [5:04] essentially [5:06] hiding the amount of value that is being [5:08] transported. [5:10] If that goes down, which means it is not [5:12] just about the value that I can get from [5:14] this, but I can also force the invalid [5:17] proofs which is essentially [5:18] counterfeiting [5:20] minting money out of thin air. [5:22] So [5:24] in order to be [5:25] in order for us to say that we are [5:27] quantum safe [5:29] there were five layers that we need to [5:30] take care of and believe these are not [5:33] just [5:34] race against a particular transaction as [5:36] some of the things like the KZG [5:38] commitments need a trusted setup and one [5:41] offline compute can crack that [5:43] which means it is not just about race [5:44] against time, but we need to have a five [5:47] layer checklist as a security researcher [5:50] that could focus on auditing the stack [5:53] and focus more on the [5:56] thinking statement that this is not not [5:58] just about [5:59] a checkbox but a checklist. [6:02] This marks the end of my talk. I hope [6:04] you have a great time. [6:05] Thank you.