Bring ETH Back: Based revolution towards One Ethereum

The Spaces brought together builders and researchers to examine base rollups: what they are, why they matter, and how they can surpass today’s run-ahead, centralized-sequencer L2s. Tom Lehman (Facet), Gustavo (Taiko), and Yunga (Rice) converged on a definition rooted in L1-aligned sequencing where users see “real” chain state (not sequencer-only state), yielding stronger liveness, censorship resistance, and a path to synchronous composability. The panel weighed Ethereum alignment and the risk that successful, centralized L2s could “eject” due to settlement costs—concluding base rollups must win on UX and cost, not just values. They explored product–market fit, arguing apps (especially institutions and high-value DeFi) will prioritize custody-minimized, resilient infrastructure. ZK proving is advancing rapidly; hybrid designs (optimistic by default with ZK on contention) and prover markets are key to cost and permissionless operation. Synchronous composability—especially L1↔L2—is viewed as the endgame, with assured sequencing and real-time proving as prerequisites; some cautioned against reliance on third-party shared sequencers, preferring Ethereum-native paths. Scaling blob space and enabling blob sharing/aggregation were highlighted as near-term bottlenecks and opportunities. Roadmaps: Facet is focused on disaster-mode UX and blob sharing; Taiko on permissionless preconfs, Shasta to reach Stage 1 and 10x cheaper proposing; Rice on ultra-low latency mainnet while decentralizing components.

Base Rollups: Definitions, Benefits, Risks, ZK, Synchronous Composability, Interop, and the Next 6 Months

Participants and Roles

• Host/Moderator: Name not explicitly given in the recording (addressed as “Ryan” in closing thanks; uncertain).
• Tom Lehman (Facet cofounder): Building a base rollup; describes Facet as the only general-purpose Stage 2 Ethereum base rollup at present.
• Gustavo (Engineering Lead, Taiko): Taiko positions itself as the first base rollup; focused on permissionless precomps (pre-confirmations), cost reductions, and a roadmap to Stage 1.
• Yunga (Research/Lead on base design, Rice): Building a base rollup with a strong emphasis on ultra-low latency finality (tens of milliseconds), composability with Ethereum L1, and decentralized architecture.

Executive Summary

• Definition and properties: A base rollup keeps the user-facing chain closely aligned with what is posted to the L1 data availability layer (as opposed to “run-ahead” rollups where users interact with blocks that exist only in the sequencer’s view). This architecture emphasizes liveness and censorship resistance, and is a prerequisite for same-block L1↔L2 and L2↔L2 interactions (synchronous composability).
• User experience vs. resilience: Today’s large non-base rollups provide great “happy-path” UX, but panelists argue their reliance on centralized sequencers creates significant tail risks (censorship, prolonged downtime, mass exit disasters). Base rollups aim to deliver comparable cost/latency with materially better worst-case guarantees.
• Alignment with Ethereum: Base rollups can better align with Ethereum’s security/validator set and support deeper L1↔L2 composability. But alignment must come with competitive costs—otherwise rollups may be economically tempted to “eject.”
• ZK trajectory: ZK proving is maturing rapidly. Panelists expect meaningful ZK adoption for base rollups within ~6–12 months, enabling tighter latency guarantees and ultimately same-block cross-domain actions (e.g., L1 flash loans used on L2 and repaid in the same block). Incentive design for prover markets and proof aggregation remains crucial to keep costs low.
• Synchronous composability: Viewed as a natural fit and long-term differentiator for base rollups. It requires assured (L1-based) sequencing and real-time proving, plus specialized infrastructure (block builders, blob aggregators). Opinions diverge on reliance on shared external services (e.g., shared sequencers/publishers) vs. using Ethereum-native pathways.
• Interop and scaling: Interoperability (interop) is a major bottleneck for new rollups. The near-term push is improving asynchronous interop (open intents, standardized frameworks) while investing in synchronous composability. Scaling blob space (and sharing it) is critical to ensure small/new rollups can compete.
• Next 6 months: Facet will prioritize resilience and blob sharing infra; Taiko targets shipping permissionless precomps with costs at or below centralized alternatives by EOY and early next year; Rice aims to launch with ultra-low-latency UX competitive with top CLOBs, then decentralize while maintaining strong performance.

What Is a Base Rollup and Why It Matters

• Tom Lehman (Facet):
  • Origin: The early base-rollup idea (referencing the “giving superpowers to the L1” thread) was that L1 block proposers could serve as sequencers, permissionlessly including the next rollup block.
  • Evolution: Core idea expanded beyond only L1 proposers, but the essence remains: user-visible blocks are tightly coupled to what’s posted to the DA layer.
  • Contrast with “run-ahead” rollups: In run-ahead designs (e.g., many op-stack deployments), users interact with blocks that live only in the sequencer’s head before they are posted to L1. Base rollups keep user-visible state much closer to what’s on L1, enabling stronger censorship resistance and synchronous composability.
• Gustavo (Taiko):
  • Key properties: Liveness and censorship resistance. With base rollups, liveness hiccups from a faulty proposer are limited to seconds/minutes, vs. potentially hours under a single centralized sequencer. Censorship resistance is stronger because inclusion isn’t gated by a single operator; rotating sequencers and assured inclusion paths are more natural in base designs.

Alignment With Ethereum and Centralization Risks

• Host question: As rollups scale, could they “eject” from Ethereum due to L1 costs, especially if sequencers are centralized and costs rise with success? Can base rollups mitigate this and keep L2s aligned with Ethereum?
• Gustavo (Taiko):
  • Alignment must be paired with competitive economics. Teams won’t stick with Ethereum settlement if costs are 10x–100x higher than alternatives. The job is to keep improving L1 scale and base-rollup tech so that cost is comparable while delivering superior properties (bridging UX, potential synchronous composability, censorship resistance).
• Tom (Facet):
  • Challenge the premise that legacy large L2s already offer the “best” UX. They shine on the happy path, but base rollups offer better UX in adverse scenarios (sequencer failure, censoring, mass exits). Many existing L2s cannot operate without their centralized sequencers; force-inclusion paths are insufficient to avoid catastrophic mass-exit risk.
• Bankruptcy/custody risk (Host): If a large L2 faces bankruptcy/legal action, could user assets and state be imperiled?
  • Tom: The risk is broader than bridge assets; the entire interdependent application state is at stake, and migrating it could cost billions. This is far worse than a simple exchange custody scenario. Base rollups reduce this risk bucket by tying execution more tightly to L1 and minimizing trusted operator assumptions.
• CEX vs. DEX analogy applied to rollups:
  • Gustavo: Not a perfect analogy. Even a Stage 2 rollup still allows users to recover funds back to L1. But we must plan for pessimistic cases while still optimizing happy-path UX to attract users.
  • Tom: The CEX analogy understates rollup complexity—unwinding a Turing-complete blockchain state is much harder than liquidating an exchange’s positions.

Current State and Product-Market Fit (PMF)

• Yunga (Rice): After ~3 years working on base rollups, PMF is still elusive. “Eventual inclusion” is a clear value, and payments may be the most obvious early use case, but few projects are focused here.
• Tom (Facet): Target apps, not end users. Users chase happy-path UX; app developers are the decision-makers who should care about resilience and portability of state. Example: an app like Polymarket on a centralized-sequencer L2 could face existential risk if that L2 fails; exporting intertwined state is prohibitively costly.
• Gustavo (Taiko): PMF candidates include institutions (banks, hedge funds) and high-value DeFi—users least tolerant of counterparty risk. Internal shift: move from “shiny research” to practical engineering—reduce latency now, and drive proposing/settlement costs to parity (or better) versus centralized models.

ZK Proofs: Costs, Incentives, and Roadmaps

• Yunga (Rice):
  • ZK proving costs and times have dropped dramatically; ZK is the only way to capture the full base-rollup benefits while accessing Ethereum liquidity/network effects.
  • Near-term vision (6–12 months): broad rollup adoption of ZK proofs enabling same-block L1↔L2 actions (e.g., L1 flash loans used on L2 and repaid within the same L1 block).
• Gustavo (Taiko):
  • Incentive design is critical. Even if proving is cheap, L1 verification can remain costly. A permissionless prover market that aggregates proofs across rotating proposers can amortize costs and avoid duplication.
  • Taiko aims for extensive ZK coverage by year-end and is designing incentives to enable efficient aggregation and permissionless proving.
• Tom (Facet):
  • Pragmatic approach: Pure per-block ZK proving is too expensive for small/new chains. Blend optimistic proposals with (uninteractive) fault proofs so only contentious blocks must be ZK proven; as usage grows, more blocks can be proven.
  • Core challenge: Permissionless proofs risk redundant proving races. Aligning incentives so one proof “wins” without waste is a hard but necessary design problem for base rollups.

Synchronous Composability (Sync-Comp): Why Base Rollups Are a Natural Fit

• Tom (Facet):
  • Concept: Pack L1 and L2 transactions into a single L1 block so outputs from one domain can be inputs to another in the same block—effectively making domains feel like one chain.
  • Today’s benefits: Faster, more reliable L1↔L2 contract calls (e.g., deposits/bridges included same-block versus waiting for deferred inclusion on run-ahead rollups).
  • Risk on non-base designs: Current op-stack interop approaches magnify centralized-sequencer dependencies—composing across multiple sequencers compounds failure/censorship risk.
  • Infrastructure needs: Specialized, compensated operators (block builders, blob aggregators) will be needed to implement same-block packing at scale. Blob aggregation services (e.g., Spire’s DA builder) are already essential due to the scarcity/competition for blobs per slot.
• Host: Synchronous composability between Ethereum and L2s is the end goal. Using L1 validators to coordinate L1↔L2 sync-compat gives base rollups an inherent advantage over legacy designs.
• Gustavo (Taiko):
  • Only base rollups are realistically sync-composable with Ethereum L1 because they require assured L1 sequencing plus real-time proving. Base rollups also make shared sequencing fairer: sharing L1’s neutral sequencers is easier than trusting a third-party shared sequencer.
• Yunga (Rice):
  • Skeptical of shared external infrastructure (shared sequencers/publishers, etc.). Every added component is another trust and complexity vector. Prefers deriving all benefits directly from Ethereum (assured L1 sequencing) with ZK as the path to sync-compat.

Industry Direction: Rollups vs. New L1s, Scaling Limits, and Blob Space

• Host question: With TradFi and brands entering, should they build L2s or new L1s? Is Ethereum’s L2 stack the “enterprise superpower”?
• Gustavo (Taiko):
  • Most apps should be rollups for 90%+ of use cases: you inherit L1 security, DA, and consensus rather than building your own validator set.
  • Concern: L1 must scale quickly enough (e.g., for hyper-throughput apps like Hyperliquid). If blob space grows too slowly, some categories will be forced elsewhere and network effects could fragment.
  • On blob growth (6 → 9 next fork, etc.): Not fast enough. Community should be more aggressive while demand is growing; underutilization today shouldn’t deter scaling targets.
• Yunga (Rice):
  • Rollups are the most efficient way to run onchain businesses: operators run minimal infra while inheriting Ethereum’s security. With data availability sampling (DAS), thousands of blobs per slot are feasible; Ethereum can host massive scale if it continues on this trajectory.

Interoperability: Today’s Bottleneck and Near-Term Pragmatism

• Host: Launching a rollup is easy; making it successful is hard. Will interop be the bottleneck?
• Gustavo (Taiko):
  • Partly yes, but we may be over-indexing on perfect synchronous interop today. Many use cases can be unlocked with improved asynchronous interop while we build toward sync-compat.
  • Taiko is funding the “Opening Time Framework” (led by “AF”) to standardize/open intents so users can reach new rollups from anywhere (L1 or any L2) without painful multi-bridge detours. Asynchronous interop improvements are a necessary bridge to the synchronous future.
• Host: Teases upcoming news in the next few weeks on progress toward synchronous composability.

The Next 6 Months: Roadmaps and Priorities

• Tom (Facet):
  • Focus on non-happy-path resilience and realistic operations (sequencer failure, governance constraints). Today’s leading rollups still have centralization choke points (e.g., centralized sequencers, security councils), which an external observer might deem unacceptable.
  • Major innovation: Blob sharing. Without blob aggregators, blob scarcity advantages the largest players with the most traffic and best compression. Sharing blob space is necessary to avoid pricing out new/small rollups and to enable the innovation base rollups promise.
• Gustavo (Taiko):
  • Goal: Prove in production that a base rollup can run with permissionless precomps and be as cheap or cheaper than centralized alternatives. Ship visible milestones by EOY; target fully permissionless precomps and very cheap transactions by Q1 next year.
• Yunga (Rice):
  • Rice emphasizes engineering for ultra-low latency UX (claims down to ~5ms per transaction). Near-term aim: host performance-critical venues (e.g., CLOBs) and payments with base sequencing.
  • After decentralizing base components, latency may increase modestly, but Rice expects to maintain category-leading performance—targeting to be faster than Hyperliquid within ~6 months while adding decentralized components.

Risks, Open Questions, and Points of Emphasis

• Cost competitiveness: If base rollups can’t match or beat centralized-sequencer cost/latency, projects may be incentivized to leave Ethereum or avoid base designs. Teams are prioritizing engineering to close that gap now.
• Prover markets and incentives: Avoiding wasted proofs, aggregating across rotating proposers, and minimizing L1 verification costs are open design and market-structure challenges.
• Blob space growth and sharing: Scaling the number of blobs per slot and enabling fair access via aggregation services is critical to prevent concentration of advantage and to support many rollups concurrently.
• Interop pragmatism: Improve asynchronous interop now (intents, standards, user flows) while building infrastructure and proofs for synchronous composability.
• Governance and failure modes: Reducing reliance on centralized sequencers and security councils remains a priority for true rollup robustness.

Notable Concepts and Terms (as used by speakers)

• Base rollup: A rollup where user-visible blocks closely track what’s posted to L1 DA, enabling assured inclusion by L1 proposers and stronger censorship resistance; contrasts with “run-ahead” designs.
• Run-ahead rollup: A design where users interact with sequencer-only state that hasn’t been posted to L1 yet, improving responsiveness on the happy path but increasing censorship/liveness risk.
• Synchronous composability: Same-block cross-domain calls (L1↔L2 and L2↔L2) that require assured L1 sequencing and real-time proving.
• Precomps (as referenced by Taiko): Pre-confirmations delivered permissionlessly to improve UX while retaining base properties.
• Stage 1 / Stage 2 rollup: Maturity classifications indicating the degree to which training wheels (centralization/trust assumptions) have been removed.
• Blob space: Data-availability bandwidth introduced by EIP-4844; scarcity necessitates aggregation/sharing for fair access.
• Prover market: A permissionless market of provers that compete/coordinate to generate and aggregate proofs at minimal cost.

Closing Sentiment

The panel views base rollups as an essential evolution for Ethereum-scale systems: more censorship-resistant, L1-aligned, and the only realistic path to robust synchronous composability. The near-term work is intensely pragmatic—cut costs, reduce latency, improve asynchronous interop, scale and share blob space, and ship permissionless components—so that the market can directly experience these advantages without sacrificing UX.