Why Contract Upgradability is a Gas Optimization Problem

The dominant upgrade pattern adds a persistent overhead to every single function call. This is the baseline cost of choice.

• Delegatecall Overhead: Adds ~2k-5k gas per transaction for the proxy jump.
• Storage Slot Clash Risk: Mismanagement leads to catastrophic bugs, as seen in early OpenZeppelin implementations.
• Immutable Trade-off: The gas tax is permanent, paid by users for the lifetime of the protocol.

Uniswap v3 consolidated all pools into a single contract to maximize capital efficiency, but this created a massive, monolithic upgrade surface.

• Upgrade Complexity: A single upgrade touches $3B+ TVL across thousands of pools, making governance and testing a high-stakes bottleneck.
• Gas Savings vs. Risk: The singleton architecture saves ~20-30% on pool creation gas but makes systemic upgrades prohibitively risky and slow.
• Contrast with v4: The upcoming hooks-based architecture is a direct response to this, decentralizing upgrade risk.

Lending giants use time-locked, governance-controlled proxies. Security is paramount, but agility suffers.

• Time Lock Drag: Every upgrade requires a 2-7 day delay, preventing rapid response to exploits or market opportunities.
• Cascading Upgrades: Adding a new asset often requires upgrading multiple contracts (oracle, comptroller, market), multiplying gas and coordination costs.
• The Fork Escape Hatch: Competitors like Euler Finance (pre-hack) used this rigidity as a growth vector by moving faster.

dYdX abandoned Ethereum and its upgrade constraints entirely, building v4 as a Cosmos app-chain.

• Eliminated Proxy Tax: Native control over the state machine removes all EVM-based upgrade overhead.
• New Costs: Introduces validator coordination, cross-chain bridging complexity, and fragmented liquidity.
• The Ultimate Trade-off: Shows the extreme length protocols will go to escape the gas and governance costs of Ethereum-native upgradability.

EIP-2535 Diamonds allow unlimited functions via facet contracts, pushing modularity to its limit.

• Per-Facet Overhead: Each external call to a different facet incurs a new proxy jump cost.
• Tooling Fragmentation: Debugging and verification become exponentially harder, increasing audit costs and security risk.
• Adoption Curve: Used by ambitious projects like Aave Gotchi, but the complexity cost limits mainstream adoption. It's flexibility for power users only.

Some protocols reject on-chain upgrades entirely, forcing all changes into social consensus or layered systems.

• Zero Upgrade Gas: Core contracts are immutable; no proxy tax exists.
• The MakerDAO Model: Changes happen via voter-approved new contracts and migration. Users bear the one-time migration gas cost.
• Social Scalability: The true cost shifts from blockchain gas to governance coordination and user education overhead. This is the ultimate decentralization tax.

Treats a contract as a collection of independent logic facets, enabling surgical upgrades without monolithic redeployment. This is the dominant pattern for modular, gas-efficient systems like Aave V3.

• Key Benefit: ~90% gas savings for targeted function upgrades vs. full proxy redeployment.
• Key Benefit: Eliminates storage collisions, enabling independent team development on different facets.

Anchors core logic (e.g., Uniswap V3's AMM math) in an immutable contract, while pushing upgradeable logic (e.g., fee management, governance) to periphery contracts. This is the security-first optimization.

• Key Benefit: Zero-risk core with verifiable, permanent state guarantees.
• Key Benefit: Peripheral upgrades are cheap and can be permissionlessly deployed, as seen in the Uniswap V3 to V4 migration path.

The classic transparent proxy pattern (e.g., OpenZeppelin) incurs a persistent ~2.7k gas overhead per call due to storage slot management and delegatecall context switching. This is a tax on every user transaction.

• Key Benefit: Simplicity for early-stage projects with low transaction volume.
• Key Benefit: High long-term cost that scales linearly with protocol adoption, creating a multi-million dollar inefficiency at scale.

Moves upgrade logic from the proxy to the implementation contract itself (EIP-1822). This reduces proxy deployment cost by ~50% and minimizes proxy attack surface, but places upgrade responsibility on implementation logic.

• Key Benefit: Lighter proxy footprint and reduced initial deployment gas.
• Key Benefit: Self-destruct risk if upgrade function is improperly implemented, trading some security for upfront efficiency.

A single, massive contract where all functions are internal and an external facade routes calls. This minimizes external call overhead and is used by gas-obsessive protocols like Solmate and early MakerDAO.

• Key Benefit: Minimizes JUMPI and external call opcodes, the single largest gas cost in function execution.
• Key Benefit: Nightmare to upgrade—requires complex migration or eternal locking, making it a high-efficiency, low-flexibility trade.

L2s like Arbitrum and Optimism use a centralized sequencer with upgrade keys, making on-chain upgrade logic irrelevant. The gas optimization is shifting cost from L1 security to off-chain governance trust.

• Key Benefit: Near-zero on-chain gas for upgrades, handled at the protocol level.
• Key Benefit: Re-introduces a trust assumption that the L2 operator will follow governance, a regression in decentralization for efficiency gains.