TechFlow Logo
Login/ Sign up
ETH Gas
Gwei
Fear
gas
Vitalik's Latest Article: Why EIP-8037 Is the Only Way Forward for Ethereum's Secure Scaling?

Vitalik's Latest Article: Why EIP-8037 Is the Only Way Forward for Ethereum's Secure Scaling?

2026.11.16
Share
Ethereum's short-term scalability must decouple execution from state growth. EIP-8037 re-prices state creation through a state Gas pool mechanism, paving the way for a substantial increase in the Gas limit while ensuring decentralization and node security.

Author: Vitalik Buterin

Translated by: TechFlow

TechFlow Summary: The core contradiction in Ethereum's scaling lies in resource heterogeneity: computational execution and network transmission can be efficiently scaled through hardware iteration and parallelization, while the uncontrolled expansion of the global state size is strictly constrained by bottlenecks in hardware storage and synchronization.

As the core proposal of the Glamsterdam upgrade, EIP-8037 introduces a state Gas reservoir mechanism, significantly repricing the cost of creating new state, and decouples execution scaling from state expansion at the protocol level for the first time. Although this breaks some legacy application assumptions based on hardcoded Gas limits, it remains the only viable engineering path for safely scaling Ethereum.

Special thanks to Binji, Lukasz Rozmej, and other colleagues for their feedback and review of this article.

The core objective of EIP-8037 is to decouple Ethereum's scaling boundaries: allowing a significant increase in easily scalable execution computation while strictly limiting the growth of hard-to-scale state volume. This is the most pragmatic solution to balance scalability and system security in the short term.

This article will focus on the following five topics:

· Explaining the core rationale behind advancing EIP-8037;

· Arguing why mechanisms like EIP-8037 are essential paths for the secure scaling of Ethereum;

· Exploring countermeasures to mitigate potential compatibility issues with EIP-8037 (including current progress, proposed solutions, and the underlying constraints behind them);

· Analyzing the communication gap between the core layer and the application layer, and reflecting on why expected mismatches arise in future EVM rule evolution;

· Providing a survival guide for developers writing immutable contracts and applications, explaining how to maintain contract security during subsequent roadmap evolutions.

Background: Capacity Limits of Different Resource Types in Ethereum

Over the past decade, the two most widely recognized core objectives of the Ethereum community can be summarized as: maintaining decentralization and security, and achieving scalable expansion.

To achieve scaling without sacrificing decentralization, we must first clarify the capacity limits of Ethereum's underlying resources: if we simply bluntly raise the Gas limit without making any other improvements, where exactly will the system break down?

The following outlines the main resource load currently borne by the Ethereum network:

Figure: Baseline operational load of Ethereum's primary resource dimensions

Under the most basic linear extrapolation, directly increasing the Gas limit by 10 times would cause all the above values to increase synchronously by 10 times.

An increase of 10 times for the first item (average verification time) remains within an acceptable range (though it will put some pressure on staking nodes running on personal laptops), but magnifying the remaining three items by 10 times would be an absolutely intolerable disaster for the network. Except for Blobs, any scaling on the Ethereum execution layer will directly hit these three red lines:

· Scaling execution computation: directly increases worst-case block verification latency;

· Scaling Calldata: directly increases worst-case network-wide broadcast transmission time;

· Scaling state (especially creating new state): directly leads to an explosive expansion of state volume. A larger state not only raises disk capacity and performance thresholds, drastically prolonging initial sync times for full nodes, but also weakens the network's deterministic guarantee for successful full-state synchronization, slowing down all state I/O reads and writes during daily execution.

Changes Brought by Other Proposals in Glamsterdam

Before discussing EIP-8037, let us first examine the improvements brought by several other key proposals in the Glamsterdam upgrade:

· EIP-7976: Increases the baseline cost of Calldata from the previous (non-zero byte 40, zero byte 10) to a uniform 64 per byte, effectively compressing the maximum Calldata size per block;

· EIP-7928: Introduces Block-Level Access Lists (BAL), supporting parallelized verification to significantly reduce verification time and improve sync speed;

· EIP-8038: Increases the Gas cost for state I/O operations;

· EIP-7732 (ePBS): Expands the secure verification window for a single block from approximately 2 seconds to about half a Slot (i.e., 6 seconds) through pipelined processing mechanisms.

Let us incorporate these proposals into consideration and reassess the system load performance when the Gas limit is increased by 10 times:

Figure: Stress test projection for 10x Gas limit after implementing Glamsterdam optimizations

From the data above, it is evident that there are no longer major bottlenecks at the computational execution level. Calldata transmission is also approaching compliance, and subsequent EIP-8279 and EIP-8131 in the Hegota upgrade will further reduce the worst-case volume by approximately 1.6 times.

Combined with planned P2P transmission optimizations, the network can easily gain more bandwidth redundancy; if further expansion is needed, EIP-8288 will also support stripping signatures and proofs out of Calldata.

Therefore, bolstered by the Glamsterdam and Hegota upgrades, Ethereum is fully capable of sustaining a 10x or even higher Gas limit increase at the computational and network levels.

The only remaining fatal bottleneck in the entire system is state volume.

When Resource Constraints Are Asymmetric: The Logic of Repricing

For nearly a decade, Ethereum has followed a clear and unwavering principle: the Gas pricing of a certain resource must accurately reflect its actual physical burden on the underlying network. If resource A imposes a burden on the network that is 10 times that of resource B, then the Gas costs for A or B must be dynamically adjusted so that the cost of A reaches 10 times that of B.

A classic historical example is the SLOAD instruction: initially costing only 50 Gas, it was gradually repriced to 200 due to genuine hardware bottlenecks in state read/write operations, later raised to 800, and finally adjusted to the current 2100.

Faced with the severe asymmetry in resource capacity described above, what EIP-8037 does is completely consistent: significantly increasing the cost of creating new state, whether adding new storage slots, writing new code, or creating new accounts.

The following shows the recommended cost adjustment comparison for EIP-8037 (already converted to final total overhead):

Figure: Repricing scheme for state creation operations under EIP-8037

Note that state Gas is not subject to the single-transaction 16.7M execution Gas limit, so deploying a full-size contract of 24 KiB can still execute normally.

Furthermore, EIP-8037 introduces entirely new logic to eliminate loopholes for cost avoidance. For instance, under current rules, creating a new account via a standard ETH transfer does not even require paying the 25,000 NEWACCOUNT overhead; after EIP-8037 goes live, regardless of the path taken to create a new account, the full 183,600 state Gas must be paid without exception.

More profoundly, EIP-8037's design allows these state overheads to be further increased over time. The core parameter determining this cost, CPSB, is initially set at 1530 and will be progressively raised in tandem with Gas limit increases.

This design reflects the consensus among core developers: computation and data transmission will become increasingly easier to scale, but the physical limits of state storage are difficult to break through. Before introducing an entirely new state paradigm, every byte of storage occupation is a genuine burden, and state cannot be easily sharded because any block producer or replay node must possess the complete state data to function properly.

Trade-offs of Alternative Solutions

Facing the dilemma of state bloat, aside from repricing, there are actually only four alternative choices in engineering practice:

· Never scale Ethereum;

· Calmly accept the centralization of nodes brought about by massive state;

· Do not increase the cost of state creation, but instead drastically reduce the cost of all operations other than state creation;

· Introduce a dedicated mechanism to decouple this additional Gas overhead from the execution path, allowing users to complete payment in an independent dimension.

The first option clearly severely contradicts the wishes of the vast majority of Ethereum community members; the second option directly undermines Ethereum's foundational security promises.

The third option was discussed extensively but ultimately rejected, as its disruptive impact would be far more severe than a price hike: drastically reducing Gas across all dimensions except state creation would trigger unpredictable chain reactions and unknown risks.

Moreover, to achieve more than 3x scaling, the costs of fundamental opcodes such as ADD and PUSH would be forced below 1 Gas, compelling Ethereum to introduce an extremely complex sub-Gas mechanism.

Thus, the only remaining way forward is the fourth option: this is precisely the breakthrough strategy of EIP-8037, although it can currently only resolve certain derivative issues.

Design and Operation of the Reservoir Mechanism

EIP-8037 represents a significant step for Ethereum toward "multi-dimensional Gas." Under this mechanism, a single transaction can simultaneously specify Gas limits across multiple dimensions: one portion allocated to execution Gas, and another portion designated as a state Gas Reservoir.

By default, the portion of a single transaction within the 16.7M execution limit is entirely classified as execution Gas, with any excess automatically injected into the reservoir (subsequent EIP-8141 frame transactions will allow users to explicitly define these two values, eliminating the need to artificially inflate the transaction Gas limit for funding).

Any operation involving state creation must deduct fees first from the state Gas reservoir; only after the reservoir is depleted will it fall back and deduct standard execution Gas. Crucially, the state Gas reservoir is completely transparent and accessible to sub-calls: a parent call can strictly limit the execution Gas allocated to a child call, but cannot intercept the child call's access to the state Gas reservoir.

This design philosophy closely mirrors the access list mechanism introduced by EIP-2930 years ago: when EIP-2929 increased storage read costs and broke some legacy contracts, EIP-2930 allowed users to pre-declare and pay read overhead in the transaction header, thereby keeping reads within the execution path cheap again.

The reservoir mechanism successfully resolves the following core challenges:

· Allows a single large contract deployment to consume Gas far exceeding the 16.7M execution limit;

· Allows certain external calls containing hardcoded fixed Gas limits to operate normally, provided the entire transaction is wrapped in a higher Gas transaction equipped with a reservoir;

· Allows parent calls to invoke child contracts using constructs similar to `gasleft() - N`, ensuring that even if a child call exhausts its Gas and reverts, the parent call retains sufficient N Gas to complete necessary post-execution state cleanup.

In the long term, this mechanism enables Ethereum to independently calculate block-level caps, target capacities, and dynamic Gas pricing for different resource dimensions (state creation, standard execution, state I/O), providing a minimum-cost engineering solution to the architectural challenges of multi-dimensional Gas.

Edge Cases Not Yet Resolved by the Reservoir

However, the reservoir is not a panacea, and the following categories of compatibility challenges remain:

· Constructing ultra-large Gas transactions is cumbersome: before EIP-8141 officially launches, requiring frontends to construct transactions exceeding 16.7M Gas appears bloated, demanding substantial frontend redevelopment efforts from application developers;

· Hardcoded configurations cannot be upgraded: core applications like cross-chain bridges have their transaction construction or call parameters hardcoded into modules that are difficult to upgrade;

· Reservoir injection is purely optional: taking the OP Stack cross-chain bridge mechanism as an example, its underlying architecture contains a hardcoded 40,000 Gas limit. Malicious attackers could deliberately send a withdrawal transaction without a reservoir, intentionally exhausting the call Gas to cause a fund withdrawal failure, while external logic might mistakenly mark it as successfully processed;

· Instant refunds disrupt Gas accounting: the proposal adopts an instant refund mechanism rather than unified settlement at the end of the transaction, causing abnormal, sharp jumps in the remaining Gas value returned by the GAS opcode;

· EIP-8038 lacks reservoir protection: EIP-8038 increases the SSTORE cost from 5,000 to 12,100 without configuring a reservoir, directly breaking a very small number of legacy applications;

· Hinders account abstraction: once the reservoir is used, the GAS opcode can no longer accurately track the total energy consumed by sub-calls, making it difficult for ERC-4337 abstract accounts to smoothly utilize reservoir Gas.

To address these pain points, the community has explored various protocol-layer remedial measures: such as reverting to end-of-transaction refund settlement, canceling EIP-8038 at the cost of narrowing scalability by 2.4 times, and adding a reservoir for state I/O.

Other concepts include prematurely introducing the EIP-8141 base framework, distinguishing empty storage slots from zero-value slots, forcing a full-frame rollback upon sub-call overruns, or granting exemptions for legacy contracts (EIP-8488), among others.

However, objectively assessed, each of the aforementioned remedial measures constitutes a major structural modification, highly prone to introducing numerous unforeseen new vulnerabilities. The only relatively mild instant refund adjustment can only salvage a minuscule fraction of affected applications; with the Glamsterdam upgrade now in late-stage testing, any protocol-layer alteration would significantly escalate operational and deployment risks.

How Should Application Developers Respond Today?

Ethereum's scaling is fundamentally an asymmetric resource game. Pure computational execution is the easiest to scale, state I/O and Calldata come next, while the expansion of state volume is the most restrictive.

To implement precise, differentiated scaling for different resources, the system must dynamically reset their relative Gas costs; and within the current realistic engineering boundaries, EIP-8037 and EIP-8038 represent compromise solutions remarkably close to the optimal answer.

The core developer team assesses that forcefully applying various fragile patches at the protocol layer not only incurs extremely high development costs but also introduces unknown systemic risks, severely delaying the overall pace of quantum-resistant upgrades and mainnet scaling.

Therefore, a more pragmatic approach is to focus primarily on application-layer coordination: fully assist the broad community of application developers to ensure that all applications and smart contracts affected by EIP-8037 and EIP-8038 complete adaptation and fixes before the Glamsterdam upgrade officially launches on the mainnet.

This includes:

· Actively delaying the Glamsterdam launch window (already executed);

· Fully and transparently clarifying these potential technical impacts (multiple public workshops and conference calls have been held);

· Dedicating specialized ecosystem resources to assist in mapping affected projects and driving patches (work is currently underway).

Survival Rules for Immutable Contracts Facing the Future

In response to this controversy, many developers have raised a completely reasonable and profound question: Ethereum's core value lies precisely in empowering users with heavily secured decentralized applications; writing immutable contracts without administrator upgrade private keys is meant to fulfill this promise. If the next hard fork could potentially break existing rules at any time, how exactly should developers craft safe and enduring smart contracts in the future?

Regarding this event, the answer boils down to a single sentence: never assume that Gas cost rules are eternally constant.

In Ethereum's early core developer and security auditing circles, this has always been an implicit foundational consensus. However, this understanding has rarely been formally documented and repeatedly emphasized since 2021, leading many newly joined developers in subsequent cycles to remain unaware that it is the default foundational assumption assumed by the entire core developer community when advancing EIP-8037.

Looking ahead, we must formalize these implicit rules: reach a strong consensus across the entire ecosystem regarding which protocol-layer characteristics belong to mutable zones, and promote them widely until they become common knowledge industry-wide. With Ethereum's development roadmap (Strawmap) now highly clear and mature, we can make relatively precise predictions about characteristics likely to undergo disruptive adjustments in the future.

Below is a personally compiled list of Ethereum characteristics expected to change in the future:

Gas Mechanisms of Any Form

When designing smart contracts, any specific Gas cap should be dynamically specified by the party bearing the loss due to an incorrect cap setting. In the vast majority of scenarios, applications are already naturally executing this principle: if transaction Gas is insufficient, the user sending the transaction suffers the loss; therefore, the transaction Gas cap is entirely voluntarily chosen and configured by the user's wallet.

Moving forward, dynamic specification must be extended to all stages, completely abandoning the practice of hardcoding any specific Gas values within contract code.

Developers are also advised not to assume that sub-calls can always operate well below the total Gas limit: after EIP-8141 lands, this mechanism may be completely removed to simplify the protocol, as its use cases can be implemented more elegantly within frames.

State Tree and Block Hash Structure

The current state tree is highly likely to be restructured into a binary partition tree based on Keccak, and block structures and transaction receipts will also fully shift to a Keccak-based SSZ serialization format.

Mathematical Finite Field Used for Blob Point Evaluation

The current system employs an internal domain of approximately 253-bit BLS-12-381, which is highly likely to migrate to a 128-bit binary finite field in the future.

Subsequent Evolution of Blob Data Structures

With the changes to the finite field, the maximum size of a single Blob is expected to be standardized to exactly 131,072 bytes, potentially triggering even more extensive structural restructuring subsequently.

ECRECOVER Precompiled Contract

To ensure that accounts upgrading to quantum-resistant cryptography in the future do not suffer from replay attacks due to historical off-chain pre-authorization vulnerabilities, ECRECOVER is expected to be transformed into a stateful operation.

Almost All Characteristics of the Consensus Layer

Within the next two years, Ethereum's consensus layer algorithms will undergo sweeping upgrades, including significant changes to validator staking participation thresholds, punishment and reward mechanisms, distributed verification technology (DVT), and light client verification protocols.

Slot Block Production Time

As the slot production duration is further compressed and shortened, schemes that directly rely on `block.slot` as a timestamp proxy will become completely obsolete.

Distributed Block Construction Mechanisms

With the advancement of mechanisms such as ePBS and FOCIL, the network-wide block construction model will undergo deep restructuring.

Transaction Serialization Format

Historical transaction formats may be gradually phased out in the future; several years from now, EIP-8141 transactions based on SSZ serialization will likely become the only supported standard transaction format on the Ethereum mainnet.

It is worth specifically noting that the EVM as a whole is not included in the mutable list. Even if Ethereum's underlying infrastructure introduces entirely new native instruction sets (such as RISC-V) in the future, compatibility layers can completely guarantee that existing bytecode logic executes identically.

Among all these, the only aspect that will truly undergo dynamic adjustment remains the Gas billing overhead for various opcodes.

This list is merely a preliminary summary compiled after a few hours of research. The core developer team should subsequently collaborate to formulate a more detailed evolutionary whitelist/blacklist, clearly delineating mutable zones and backward-compatibility foundations.

The vast majority of standard applications (such as Uniswap-style DEXs) are naturally immune to most upgrade shocks. However, when facing immutable contracts and complex financial applications, the entire industry must still establish defensive regulatory frameworks as early as possible.

Join TechFlow official community to stay tuned

Add to Favorites
Share to Social Media