ZK rollups hold approximately 20% of Ethereum Layer 2 total value locked despite offering faster finality and stronger cryptographic guarantees than their optimistic counterparts. As of late July 2026, L2BEAT recorded $33.8 billion in total value secured across Ethereum scaling solutions. Base ($...
"The team acknowledged the project never gained meaningful adoption due to technological limitations." — Loopring Foundation, June 2026 Shutdown Announcement
ZK rollups hold approximately 20% of Ethereum Layer 2 total value locked despite offering faster finality and stronger cryptographic guarantees than their optimistic counterparts. As of late July 2026, L2BEAT recorded $33.8 billion in total value secured across Ethereum scaling solutions. Base ($11.5B) and Arbitrum One ($10.1B) alone command roughly 64% of that figure. zkSync Era, the largest ZK rollup by TVS on L2BEAT, registers $202 million — less than 1% of the total.
The first half of 2026 saw three ZK rollup shutdowns: Loopring wound down its DEX on June 28, Polygon sunsetted its standalone zkEVM chain on July 1, and ZKsync Lite entered formal deprecation. Zero Network, a gasless Layer 2, ceased operations after 18 months. These closures follow a pattern identified by 21Shares: smaller rollups' usage has dropped 61%, and the top three networks now process approximately 90% of all L2 transactions. The data presents a paradox: ZK technology is maturing, proving costs are falling, and new hardware startups have raised over $100 million — yet the market continues to consolidate around optimistic rollups.
L2BEAT's late-July 2026 snapshot recorded approximately $33.8 billion in total value secured across Ethereum scaling projects. The distribution is heavily skewed:
Optimistic Rollups (~80% of L2 TVL):
ZK Rollups (~20% of L2 TVL):
The methodological gap between L2BEAT's TVS metric (which measures bridged token value) and DeFiLlama's TVL metric (which counts redeemable deposits) explains some discrepancies in zkSync Era's figures. Regardless of which data source is used, ZK rollups collectively command a fraction of the capital that optimistic chains attract.
Post-EIP-4844, transaction fees have converged across all major L2s to the $0.02–$0.09 range. Base averages approximately $0.05 per transaction, Arbitrum and Optimism $0.08–$0.09, zkSync Era $0.07, and Starknet $0.05–$0.19. Fee parity has removed one of the few user-facing differentiators.
Three ZK-based networks ceased operations in H1 2026, each illustrating a different failure mode.
Loopring (Shutdown: June 28, 2026): Ethereum's earliest production ZK rollup launched in 2020. TVL collapsed from approximately $760 million in November 2021 to roughly $8 million at closure. The LRC token fell from $3.75 to approximately $0.01 — a decline exceeding 99%. The team cited weak user adoption, limited business development capabilities, and competition from newer zkEVM-based networks. Loopring confirmed all historical L2 data remains accessible on-chain via The Graph, and remaining user funds were distributed to Ethereum wallets in batches.
Polygon zkEVM (Shutdown: July 1, 2026): Polygon's standalone zkEVM Mainnet Beta sequencer stopped producing blocks. Wallet-held assets not bridged before the deadline were auto-migrated to Ethereum L1. Funds locked in DeFi protocols, however, could not be automatically migrated and may become inaccessible. A claim window remains open until December 31, 2027. Polygon stated that the zkEVM shutdown does not affect CDK-based chains.
ZKsync Lite (Deprecation: 2026): Ethereum's first dedicated ZK rollup, launched by Matter Labs in June 2020, entered formal deprecation. The network, which supported only simple transfers and NFT minting without smart contract capability, was rendered obsolete by zkSync Era's March 2023 launch. Withdrawals to Ethereum L1 remain functional during the transition period.
Zero Network (Shutdown: July 31, 2026): Though not strictly a ZK rollup, this gasless Layer 2 built by Zerion lasted only 18 months before the company pivoted resources back to its wallet and API products.
According to CoinDesk, over 100 crypto projects shut down in H1 2026 — a pattern described as a "dot-com style shakeout." For L2s specifically, the consolidation is stark: per L2BEAT, over 73 rollups exist, but the top three process 90% of transactions.
ZK rollups carry a structural cost that optimistic rollups do not: generating cryptographic proofs for each batch of transactions. This "ZK tax" manifests in sequencer margins.
According to industry analyses, specialized ZK-ASICs and optimized GPU clusters have reduced proving costs by over 90% since early 2025. Proof generation for batch rollups that once took hours now completes in seconds to minutes on specialized hardware. Despite this improvement, zkSync Era operates at thinner sequencer margins than comparable optimistic rollups. The proof generation cost remains a meaningful component of the operating cost base.
For context, Base was the only L2 that turned a profit in 2025, earning approximately $55 million according to Yellow Research. ZK rollups, which must amortize proving costs across their transaction volume, face a circular problem: lower volume means higher per-transaction proving cost, which makes fees less competitive, which suppresses volume.
Post-EIP-4844, data availability costs dropped for all rollups, but this benefited optimistic rollups disproportionately — they had no proving overhead to begin with.
ZK rollup tokens have underperformed significantly in 2026.
ZK (zkSync): Trading at approximately $0.010–$0.015 as of August 2026. Market cap: ~$99 million. Fully diluted valuation: ~$206 million. The token has declined 74.66% over the past year. A token unlock of 173 million ZK ($1.35 million) is scheduled for August 17, representing 0.82% of total supply.
STRK (Starknet): Trading at approximately $0.024 with a 24-hour volume of $8.8 million. Starknet's TVL (~$179 million) lags mid-tier rollups, though a $43 million net bridge inflow in the week of July 22 indicated selective capital rotation.
LINEA: Generated by Consensys in September 2025, the token spiked to $0.047 on launch day, crashed 85% within two months to $0.006, and declined further to $0.003 by April 2026.
The market is pricing ZK rollup tokens at a steep discount to their optimistic counterparts. This reflects both the TVL gap and uncertainty around when — or whether — proving cost economics will converge to parity.
One data point cuts against the narrative of ZK decline. According to multiple 2026 analyses, the dominant ZK rollups collectively settle more stablecoin volume than every optimistic rollup combined. This is attributed to ZK rollups' faster finality characteristics — no 7-day challenge period — which aligns with the requirements of payment and treasury operations.
This suggests a bifurcation in L2 use cases. Optimistic rollups dominate general DeFi activity and hold most stablecoin TVL. ZK rollups may be capturing a disproportionate share of high-velocity stablecoin transfers, where finality speed matters more than deposited value. The economic value, measured by settlement throughput rather than locked capital, may be less skewed than TVL metrics suggest.
However, this claim is difficult to independently verify with consistent cross-chain data. If accurate, it implies ZK rollups may serve institutional payment rails even as they lose the consumer DeFi market to optimistic chains.
The thesis behind continued ZK investment rests on proving cost trajectories. Multiple hardware startups have raised significant capital:
Collectively, these companies represent over $100 million in venture funding. According to a June 2026 IEEE paper, the NoCap accelerator achieves proof generation 586x faster than a 32-core CPU and 41x faster than PipeZK, a prior state-of-the-art ZK accelerator.
The economic question is whether these hardware improvements can close the proving cost gap before the L2 market fully consolidates around optimistic incumbents. The timeline remains the primary uncertainty.
With its standalone zkEVM chain shuttered, Polygon has repositioned CDK (Chain Development Kit) as a managed, enterprise-grade service — "chain-as-a-service, not a self-serve kit," according to official documentation. Live CDK deployments include OKX's X Layer, Immutable's gaming chain, and Astar Network.
This represents a strategic pivot from competing with Arbitrum and Base for general-purpose L2 market share toward providing ZK infrastructure to enterprises building application-specific chains. Arbitrum Orbit and the OP Stack dominate the framework market for deploying app-specific chains, but CDK targets a different buyer — enterprises that value the ZK security model for specific compliance or settlement use cases.
Linea has made a parallel pivot since March 2026, shifting from retail engagement farming to positioning itself as a settlement rail for SWIFT and global custodians, according to Consensys communications. Whether these institutional pivots generate sufficient transaction volume to sustain ZK proving economics remains unproven.
The data presents a ZK rollup sector caught between technical progress and market reality. Proving costs are falling, hardware is improving, and the cryptographic guarantees remain objectively stronger than optimistic fraud proof models. None of this has translated into TVL dominance, fee revenue competitiveness, or token value retention.
The stablecoin settlement data — if confirmed across independent sources — suggests ZK rollups may find durable demand in institutional payment corridors where finality speed matters more than DeFi composability. The enterprise pivots by Polygon CDK and Linea align with this thesis.
The counter-argument is network effects. Base has Coinbase's 100+ million user base. Arbitrum has the deepest DeFi liquidity. Switching costs are high and rising. By the time ZK proving costs reach parity with optimistic rollups, the market may have already consolidated beyond recovery.
The L2 landscape is repricing ZK technology from "inevitable future" to "niche infrastructure." The next 12 months of proving cost data, enterprise adoption metrics, and stablecoin settlement volumes will determine which assessment is correct.