Curve Logs 1.84 Million Gas in a Single Ethereum (ETH) Leveraged Borrow

Curve's Mainnet Gas Checks

The Ethereum (ETH) price slipped 3.8% in the past 24 hours, but the number drawing developer attention this week comes from Curve Finance's own mainnet checks: a single leveraged borrowing transaction that consumed roughly 1.84 million gas, against 125,000 for a direct stablecoin swap. On those figures the borrowing operation cost about 14.7 times as much to execute as the swap, a gap the DeFi protocol presents as evidence that complex transactions need the most help reaching blocks. Leveraged borrowing sits at the core of margin trading on lending platforms, and it is where the measured gap shows. Curve, the team behind the crvUSD stablecoin, argues that Glamsterdam's value should be measured by how transactions perform during congestion rather than by the block gas limit alone. It named transaction inclusion times and the fees users actually pay during demand spikes as the practical tests of whether added capacity improves access, and made easier validation a condition of successful scaling: capacity only counts if the network can still be verified independently. During sharp market moves, liquidators, oracle updates and borrowers compete for blockspace, and Curve expects extra capacity to help their transactions clear, though larger blocks would still face demand. The checks carry weight because operations that combine borrowing with a collateral purchase can be delayed at several points inside the same transaction, so congestion hits them repeatedly. Cheaper execution could also make smaller arbitrage trades worth completing, with longer swap routes becoming economical and lifting the prices available through Curve's pools. The team stopped short of claiming every action would become cheaper, pointing to Glamsterdam's revised charges for state operations as a source of higher gas requirements for some transactions. Its standard for judging the fork is blunt: “What we think should be watched most closely is execution under real load.”

EIP-8037 Repricing and the Sepolia Test

The protocol work behind those expectations is already running on testnets. Sepolia is being prepared for a block gas limit of up to 200 million, up from roughly 60 million in current settings, and developers have stated plainly that the test setting will not automatically become mainnet's limit. Prysm's Oct. 5 release notes show version 7.2.1 introducing the Sepolia schedule, under which each validator defaults to 200 million gas once the fork arrives; the prior release, though upgrade-capable, kept 60 million until an operator adjusted its configuration. The repricing Curve flagged lives in two specifications: EIP-8037 raises what it costs to create state, accounts and contract storage included, while EIP-8038 overhauls the charges applied when existing state is accessed. The Ethereum (ETH) Foundation set out the compatibility risks in an Aug. 26 report, finding most contracts unaffected in transaction replays while noting that many flagged failures could be fixed by supplying higher gas limits. Hardcoded call limits, fixed gas allowances and logic sensitive to gas were listed among the risk factors, and the warning extended to wallets, infrastructure providers and gas-estimation tools that rely on cached assumptions. Developers were urged to test applications before mainnet activation. Curve ties the design back to proof of stake's core promise: Ethereum should preserve the ability to verify its ledger without asking permission, and scaling would compromise that principle if added capacity made independent verification impractical. On the risk side, Curve expects cheaper and faster arbitrage to help LLAMMA, its liquidation mechanism, follow market prices more tightly as collateral moves through the liquidation range, limiting losses when prices shift before transactions complete and supporting crvUSD stability. The upgrade has rehearsed on Devnet-11 with 84,000 validators across multiple clients, though that controlled run excluded deliberate attacks, and developers have warned that free testnet ether could let a hostile builder win block auctions and withhold transaction payloads. Hoodi and mainnet activation dates remain unset, with separate announcements to follow agreement among client teams.

Capacity Versus Cost

Our reading of the 14.7x spread is that it marks where Glamsterdam's benefit and its cost meet. The gas gap between a borrow and a swap is widest during volatility, exactly when liquidations are most urgent, so capacity that shortens inclusion times is worth more to lending systems than to simple swaps. If the 200 million gas configuration holds on Sepolia and the EIP-8037 and EIP-8038 schedules land with contracts patched, complex operations gain room; if repricing raises costs faster than capacity arrives, the net effect on borrowing-heavy positions could be neutral or negative. The fork is ultimately a bet on scaling Ethereum (ETH) without breaking independent verification, and the unset mainnet date is the binding constraint. Readers tracking levels while the schedule waits can follow our Ethereum technical analysis, since a recent range breakdown shows the market is not waiting for the fork.