Mentioned in This Article
1. Starknet L1: what has been proposed
On October 8, Starknet's builders said they are weighing a move away from their current design as an Ethereum layer-2 network toward a standalone layer-1 blockchain, with StarkWare chief executive Eli Ben-Sasson describing the idea publicly the same day. The aim is a Starknet L1 that could harden itself against quantum computing and advanced AI faster than a rollup tied to Ethereum can. Nothing formal has begun, and the project has named 2027 as its target.
The announcement describes the destination as the first fully quantum-resistant network. That is a bold framing, and it explains why the discussion has drawn attention well beyond the Starknet community. The proposal arrives as a broader argument over how quickly major blockchains should prepare for cryptography-breaking advances, an argument sharpened by recent remarks from Ethereum researcher Justin Drake about adopting a so-called bunker mode.
It is worth being precise about the status. Starknet has said it is considering the change. There is no governance vote, no published specification and no migration schedule. What exists is a stated direction, a reasoning chain and a date to aim for, which makes the logic behind the idea more informative than any technical detail so far.
2. Why is Starknet weighing a move from L2 to L1?
The short answer is control over timing. A layer-2 rollup posts its data and proofs to Ethereum and relies on that base layer for part of its security, which means certain upgrades can only happen when Ethereum itself makes them. As a standalone chain, Starknet could set its own pace for security migrations instead of waiting.
Ben-Sasson has said the thinking grew out of concern that traditional cryptography could be broken sooner than many people expect. He pointed to both quantum computers and progress in AI-driven mathematics as risks, and he framed the response as a matter of urgency. The two threats differ in kind, since one attacks the hardness assumptions behind widely used signature schemes and the other could in principle find shortcuts in mathematics people have long treated as secure, but both argue for a network that can change its cryptography quickly.
The commentary around the proposal makes the same observation from the other direction: it is unclear whether Drake's timeline for such breakthroughs is too pessimistic or too optimistic. Because nobody can say with confidence, projects differ widely in how seriously they treat the risk. Starknet's builders are effectively betting that moving early is cheaper than being late.
The idea also did not appear in StarkWare's earlier planning. Its post-quantum roadmap, published on June 30, makes no mention of becoming an L1. The proposal therefore represents a shift in thinking over a few months rather than the next step of an existing plan, and it suggests that the roadmap's dependence on Ethereum looked more constraining the longer the team examined it.
3. What the June post-quantum roadmap already covers
StarkWare's roadmap divides the move to quantum-resistant cryptography into three phases. The first two are work Starknet can carry out by itself. They include replacing the Pedersen hash function with BLAKE2, introducing Falcon-512 signatures, and migrating legacy contracts to the new scheme. None of these steps requires anyone else's permission or timetable.
The third phase is where the difficulty lies, because it depends on two components borrowed from Ethereum. Those components are the messaging layer that connects Starknet to Ethereum through its bridge, and the way Starknet publishes its state data to Ethereum.
Understanding these two dependencies is the key to the whole proposal. They show that a rollup's security is partly inherited, and that inheritance cuts both ways: it spares a rollup from building everything itself, but it also ties the rollup's schedule to its parent chain.
4. Where Ethereum's timeline holds Starknet back
The bridge is the first problem. Messages passing between Starknet and Ethereum use elliptic-curve operations, which are vulnerable to a sufficiently capable quantum computer. StarkWare's position is that removing those calls before Ethereum has migrated would break the bridge, so Starknet cannot simply fix its side first.
The second problem is data publication. Starknet posts state data to Ethereum in blobs, and those blobs are anchored by KZG commitments, another elliptic-curve construction. If KZG were broken, the damage would fall on data availability rather than on the soundness of Starknet's state, because STARK proofs, which rely on hash functions instead of elliptic curves, protect state validity. Even so, losing data availability would be a serious failure for a network that users depend on to retrieve their balances. The June roadmap leaves this fix to Ethereum.
Ethereum's own post-quantum work may take a few years. If that estimate holds, key parts of Starknet's roadmap would be on hold for at least as long. This is the heart of the argument for an L1: a rollup effectively moves at the slower of its own pace and that of the chain beneath it.
The same point applies beyond Starknet. Any rollup inherits its base layer's cryptography and its security assumptions, so its quantum timeline cannot be shorter than the base layer's for the affected components. Other layer-2 projects may eventually confront the same constraint, and Starknet's decision could become a reference case.
5. Why Starknet can move faster than Ethereum
Starknet expects to hit this ceiling precisely because it can complete its own share of the work sooner than Ethereum can. Some of Ethereum's proposed directions, such as EIP-8141, would make the base layer behave more like Starknet does. That proposal would let accounts adopt quantum-safe signatures on their own schedules instead of requiring a single network-wide switch.
Starknet has offered native account abstraction for years, meaning that accounts are programmable and can define their own signature rules. A project called S2morrow could build a post-quantum account around Falcon-512 without any change to the protocol. In other words, the flexibility that Ethereum is considering adding is already part of Starknet's design.
Ethereum's plans on the validator side are separate. The network intends to replace BLS validator signatures with leanXMSS, a hash-based scheme, and to compress the results using zero-knowledge proofs through leanVM. STARK-based commitments are among the candidates to take the place of KZG. Those paths could eventually remove the dependencies that worry Starknet, but the timeline for them remains the open question.
6. What Starknet already has in place for a standalone chain
Over the past couple of years, Starknet built several pieces that a layer-1 would need. They include staking of the STRK token and support for multiple sequencers. These features were designed to decentralize Starknet as an L2, but they also make the step toward an independent network smaller than it would otherwise be.
Several parts are still missing. A standalone chain would need its own solution for data availability, a mechanism for final settlement, and a reworked official bridge. Each of these is a significant engineering and governance task, which is a reminder that the existence of staking and sequencer infrastructure does not make the transition simple.
Settlement deserves particular attention. Today, Starknet does not take the word of its operators for the state that bridged funds depend on. Withdrawals through the StarkGate bridge are released only after Ethereum verifies the proof for the relevant block and the matching state update has landed there. Ethereum effectively checks Starknet's mathematics, so operators cannot forge the state behind users' bridged assets.
7. How much trust would an L1 Starknet require?
This is the question that the proposal leaves open, and it may matter more than any other. An L1 version of Starknet could keep posting STARK proofs to Ethereum, which would preserve a trust-minimized bridge, or it could cut that connection entirely. The two approaches imply very different trust models for people holding assets on the network.
If proofs continue to flow to Ethereum, users would keep a strong guarantee that bridged funds are backed by verified state, while Starknet would control its own cryptographic upgrades. If the link were severed, bridging would depend on other mechanisms, and users would need to trust those instead. Which path the team selects is likely to determine how the community judges the plan.
Starknet has also presented itself as a layer that could one day settle on both Bitcoin and Ethereum. A proof-connected L1 that bridges to both would fit that vision, not contradict it. In that reading, independence from a single base layer is a continuation of the project's stated strategy and not a departure from it.
8. What happens next for the Starknet L1 plan
Nothing official has started. There is no vote scheduled, and no date has been given for a detailed proposal. The only timing signal is the 2027 target for a fully quantum-resistant network, which suggests the team wants to act soon, partly because some projects may not adapt in time.
For observers, the developments to watch are straightforward: whether Starknet publishes a formal proposal, how it intends to handle data availability and settlement, and whether it keeps a proof-based bridge to Ethereum. Ethereum's progress on its own post-quantum upgrades will also shape the argument, since a faster Ethereum timeline would weaken the case for leaving.
Whatever the outcome, the discussion highlights a practical lesson for the wider industry. Quantum and AI risks are no longer treated only as distant theory by major teams, and decisions about architecture are beginning to reflect them. An L1 architecture currently appears to be the path Starknet considers best suited to that challenge.