Hook
Delta's CEO just told CNBC that oil prices are “sticky for longer” because travel demand is too strong for airlines to cut fares. The market yawned. But if you swap “oil” for “Ethereum gas” and “travel demand” for “L2 activity,” you get the exact same macro tension playing out on-chain right now — and almost nobody is talking about it.
I sat through the Uniswap v2 launch party in 2020 where the constant product formula was still news to 90% of the room. I learned then that the most dangerous trade is the one everyone agrees on. Today, the consensus is that L2s will solve Ethereum’s congestion permanently. The code didn’t say that. And the on-chain data is screaming something else.
Context
Since the Dencun upgrade in March 2024, blob space has slashed L2 costs by 90%+ on Optimism and Arbitrum. Transaction volumes on Base and Arbitrum hit all-time highs in Q2. And yet — check Etherscan right now — median gas on Ethereum L1 has been hovering between 8 and 15 gwei for weeks. Not the sub-5 gwei we saw during the ‘crypto winter’ of 2022. Not the 100+ gwei during NFT mania. It’s sticky. Exactly like Delta’s oil.
The reason? L2s don’t eliminate L1 demand; they aggregate it. Every L2 batch submission, every state root update, every proof verification — they all consume L1 gas. When Base processes 100 million transactions in a month, that’s thousands of batches landing on L1, each paying a baseline fee. The link between L2 activity and L1 gas price is not linear, but it’s real. And it’s tightening.
Core: The On-Chain Data Doesn’t Lie
Let me show you the numbers I’ve been tracking since I built my own gas monitor during the Fomo3D era (yes, that wallet-dormancy trap I broke 4 hours before CoinDesk).
Over the past 30 days:
- Total L2 transaction count: 1.2 billion (source: L2beat). Up 40% from the same period pre-Dencun.
- Ethereum L1 median gas: 11 gwei. It hasn’t dropped below 7 gwei in any 24-hour window since March.
- Blob utilization: consistently above 80%. The memory pool for blobs is routinely full.
This is the equivalent of Delta saying “load factor is 95% and fuel costs are not coming down.” The demand side (L2 users) is structurally high. The supply side (L1 block space) is fixed at ~30 million gas per block with an adaptive base fee mechanism. When both forces are strong, the equilibrium price — gas — becomes sticky.
But here’s the part that market cheerleaders are missing. Most analysts treat L2s as a “scaling solution” that will eventually reduce L1 congestion. They see L2 volume and assume L1 gas will fall. The contrarian truth is that L2 activity itself becomes a new source of L1 demand. Ethereum is becoming a settlement layer that also brokers data availability. More L2 usage = more blobs = more L1 base fee pressure.
We didn’t account for that in the early DeFi Summer models. I remember sitting in that San Francisco loft, sketching out fee curves on a napkin with a developer who later built one of the first L2 rollups. We assumed that once L2 fees dropped, users would flee L1 and gas would collapse. We were half right. Users fled L1 for execution. But they brought back demand as proofs.
Let me harden this with a specific on-chain signal. On June 12, when Arbitrum processed 2.1 million transactions (its third-highest day ever), the daily gas consumed by L1 just from Arbitrum batch submissions was over 1.2 billion gas — about 4% of all L1 gas for the day. Multiply that by 40 rollups and validiums, and you’re looking at 30-40% of total L1 gas now being consumed by L2 infrastructure. That’s not going away.
Contrarian Angle: The “Sticky Floor” Is a Feature, Not a Bug
If you’ve read my work before, you know I hate the phrase “this time is different.” But this time, the macro of on-chain economics really is different. The traditional narrative says that high gas fees are a sign of failure — that Ethereum is too expensive and will lose users to Solana or other L1s. The contrarian take, reinforced by the Delta analogy, is that a sticky gas floor is the natural equilibrium of a mature network that serves both as a settlement layer and a data availability layer.
This has profound implications for portfolio positioning:
- ETH becomes a revenue-generating asset even in a sideways market. If median gas stays at 10 gwei forever, ETH’s annualized burn rate at current block times is about 1.5% of supply. That’s a yield no one is pricing in.
- L2 tokens (ARB, OP) benefit from volume but face margin compression if L1 gas stays high. Their own cost base (blob fees) rises, but they can’t easily pass that to users because competition is fierce.
- The real alpha is in infrastructure that sits between L1 and L2 — bridge protocols, sequencing providers, and data availability layers like Celestia that can offer cheaper alternative revenue for rollups.
I saw this same structural tension play out in real time during the Bored Ape floor dip in early 2021. Everyone thought the floor would recover only when “paper hands” washed out. I organized a private dinner with Toronto whales and learned they were buying the dip for branding — the floor was sticky because demand was sticky even as price fell. The parallel here is that L1 gas is sticky because L2 demand is structurally high, not because of speculative bots or NFT mints.
Takeaway
The market is too busy chasing the next L2 airdrop to notice that the base layer’s fee market has fundamentally repriced. If Delta’s CEO is right about oil, then the on-chain analog is clear: don’t bet on cheap Ethereum gas ever returning to sub-5 gwei levels. The next time you see a headline calling Ethereum “dead” because gas is “only” 12 gwei, remember: sticky is not the same as broken. It’s a signal of structural demand. The question is whether you’re positioned to benefit from the stickiness — or to panic sell into it.