Thesis: Post-Dencun, the rollup wars stopped being about fees and TPS. Those are commoditized. The contest now is pure capital gravity, and the leaderboard is consolidating into an oligopoly faster than the long tail will admit.

The Data

Collective rollup TVL exceeds $48 billion across more than 73 networks, but the distribution is brutally top-heavy.

  • Arbitrum One leads at $13.8 billion, roughly 40% market share. Base sits at $11.2 billion.
  • Five chains (Arbitrum, Base, OP Mainnet, zkSync Era, Linea) command about 75% of total L2 TVL.
  • Fees are a rounding error post-EIP-4844: $0.02 to $0.06 per transaction. Throughput clusters in a tight band, Arbitrum near 62 TPS, Base up to 89 at peak, zkSync near 28, Starknet near 19.

When fees and speed converge, they stop being a differentiator. What is left is liquidity, app gravity, and distribution.

The Risk

The bull case for the long tail is app-specific chains and the rollup-as-a-service thesis fragmenting demand. That is real, but fragmentation without liquidity is just dilution. The counter-risk to the incumbents is governance and token-incentive dependence: some of that top-tier TVL is mercenary, parked for emissions, and rotates the moment yields normalize. Base also carries single-sequencer centralization that an institutional allocator will eventually price.

The Take

The era where a new rollup could win on cheaper, faster is finished. With throughput solved and fees near zero, capital concentrates where capital already is. And with the Crypto Fear and Greed Index pinned at 13 (Extreme Fear), mercenary liquidity is exactly the kind that flees first, leaving the deepest books even more dominant. The top five own the liquidity, and in market structure, liquidity is the only moat that does not erode. Treat sub-scale L2s as venture bets, not infrastructure positions.

Sources: Spoted Crypto, Eco, Coin Bureau, L2 TVL trackers, June 2026. Editorial research. No financial advice.