After L2 transaction fees fell below 1 cent, how does Ethereum, which no longer sells gas, make money?

ETH is no longer a radical dream, but a sober engineering project

Ethereum was once the narrative engine of the Web3 world. From the grand vision of "Merge" to the "ultrasonic currency" myth brought about by the EIP-1559 burning mechanism, every key node was accompanied by a consensus frenzy and soaring valuations. However, as we enter 2026, the sky for Ethereum has changed.

With the Ethereum Foundation recently updating its 2026 protocol priorities, a clear signal has been released: Scale, Improve UX, and Harden the L1 have become the three main themes.

As the Ethereum Foundation recently updated its 2026 protocol priorities, a clear signal has been released: Scale, Improve UX, and Harden the L1 have become the three main themes.

This shift, rather than a proactive strategic adjustment, is more accurately described as a choice of "engineering-driven survival" under competitive and realistic pressures. Industry competition is forcing this behemoth to shift from "storytelling" to "engineering," and from "narrative-driven growth" to "engineering-driven survival." Looking back at Ethereum's history, from smart contracts in the ICO era to DeFi Summer, and then to the transition to PoS and deflationary narratives, each leap has been accompanied by a strong market narrative ability. However, entering 2026, the marginal utility of narratives is diminishing, replaced by cold data metrics and underlying architecture restructuring. The most significant engineering leap in the roadmap is the upcoming Glamsterdam hard fork in the middle of the year. This upgrade directly addresses the long-standing pain points of Ethereum's mainnet performance, with two core metrics being particularly crucial: first, significantly increasing the mainnet's gas limit from 60 million to 200 million; and second, officially introducing a parallel execution architecture to the mainnet. For a long time, Ethereum's EVM has used a single-threaded serial processing mode. This mode has advantages in ensuring state consistency, but it becomes a fatal bottleneck in high-concurrency scenarios. Introducing parallel execution means that Ethereum is expanding from a "one-way street" to a "multi-lane highway." By using a block-level access list, nodes can predict which transactions will not involve state conflicts, thus processing multiple transactions simultaneously. With the gas limit increasing to 200 million, the computational and transaction volume that each block can accommodate will grow exponentially. However, this is not without cost. The increase in the gas limit directly challenges Ethereum's long-held principle of "democratizing full nodes." State expansion will accelerate, drastically increasing the storage and network bandwidth requirements for node hardware. To mitigate this risk, the Ethereum engineering team plans to push approximately 10% of validators to shift from "re-executing all transactions" to "verifying zero-knowledge proofs" within the year. This is known as "SNARKing the L1," which not only significantly lowers the hardware barrier for full nodes but also marks a watershed moment in Ethereum's evolution from "repetitive labor" to "intelligent verification." This means that Ethereum's underlying computational model is undergoing a qualitative change, outsourcing or fronting up heavy computations, and gradually shedding the complex execution layer burden from L1. This is purely an engineering compromise and progress. The performance anxiety stems largely from the disruptive impact of competitors. In 2026, the performance war in the public blockchain sector had reached a fever pitch. Solana, through its Alpenglow upgrade, completely abandoned its previous Proof-of-History (PoH) and Tower BFT consensus mechanisms, adopting a brand-new Votor and Rotor architecture. The direct result of this underlying restructuring was that Solana's transaction finality was reduced from 12.8 seconds to less than 150 milliseconds. This is a highly disruptive metric. A latency of 150 milliseconds falls within the response range of traditional Web2 internet infrastructure (such as Google Search or the Visa payment network). For applications extremely sensitive to latency, such as high-frequency trading (HFT), full-chain derivatives exchanges, and real-time payments, this constitutes a fatal attraction. In contrast, while Ethereum's Glamsterdam upgrade and subsequent Heze-Bogota fork aimed to improve TPS and censorship resistance, its modular and complex architecture is inherently disadvantaged in terms of cross-chain composability and latency. Although Ethereum's current block cycle is 12 seconds, true finality takes several minutes. While this architecture is robust for settling high-value, low-frequency assets, it is too cumbersome for consumer-grade applications targeting a massive number of retail users. Ethereum's performance anxiety is essentially a debate between monolithic and modular architectures during the technological explosion expected in 2026. If Solana's relentless pressure is an external threat, then Ethereum also faces an internal paradox stemming from its own strategy—the "L2 paradox." With the implementation of Pectra and Fusaka upgrades, and the maturity of PeerDAS technology, Ethereum's Rollup-centric scaling strategy has achieved a tremendous engineering victory. L2 data availability throughput has increased several times over, and the capacity of data blobs continues to expand. The direct result is that L2 transaction fees have plummeted to $0.001 or even lower. From a user experience perspective, this is a huge success, perfectly aligning with the "Improve UX" theme of the 2026 roadmap. The widespread adoption of native account abstraction and intent frameworks has completely hidden complex on-chain interactions within seamless wallet operations. However, this raises a pointed question: when users enjoy a smooth transaction experience of $0.001 on L2, do they really care what consensus mechanism the underlying Ethereum mainnet uses? The "decentralized legitimacy" that the Ethereum community is so proud of, and the censorship-resistant network composed of tens of thousands of independent validators, are becoming, in the eyes of most end users, an invisible, abstracted backend database. When application execution is completely migrated to Arbitrum, Base, or ZKsync, and the mainnet merely serves as a verification layer for data availability and state roots, Ethereum not only loses direct access to end-users but also faces the risks of liquidity fragmentation and hollowing out of the application layer. This is not only a decoupling of technical architecture but also a decoupling of brand perception and user mindset. From "selling Gas" to "selling secure settlement services," ETH's value capture method has changed. This evolution of the technical roadmap will ultimately be reflected in the asset's pricing model. Ethereum's current transformations are triggering a fundamental reshaping of ETH's value capture logic. For most of 2021 to 2024, ETH's value was primarily supported by the "world computer" narrative and the gas fee burning mechanism introduced by EIP-1559. Higher on-chain activity meant more ETH was burned, strengthening the deflationary expectation of "Ultra Sound Money." This model is essentially a consumer-facing retail logic—Ethereum is "selling gas." However, by 2026, the situation had changed dramatically. As execution layer activity irreversibly migrated to L2, mainnet gas consumption decreased significantly. Although L2 required payment for data availability (DA) to L1, with the continuous expansion of the Blob space, this revenue was far from sufficient to fill the gap left by L1 execution layer fee losses. ETH's burn rate dropped significantly, even returning to microinflation during the trough, severely testing traditional deflationary expectations. From a quantitative finance valuation perspective, Ethereum's DCF (Discounted Cash Flow) model is undergoing a rewriting. Ethereum is transforming from a high-margin computing platform geared towards retail into a low-margin, high-certainty "secure settlement layer" geared towards B2B (L2 and even L3) clients. Its new business model is no longer "selling gas," but rather "selling economic security" and "censorship-resistant finality." Under this paradigm, the return structure of ETH as a monetary asset is changing. The implementation of ePBS (proposer-builder separation at the protocol level) will restructure the MEV supply chain, making the distribution of MEV rewards within the validator network smoother and more predictable. The benchmark returns from staking and restaking will replace gas burning as the core supporting ETH's valuation. This makes ETH's asset attributes more similar to traditional government bonds or institutional-grade clearing and settlement assets. It no longer needs fancy Meme coin transactions to contribute fees, but instead relies on its massive staked capital to provide an immutable trust guarantee for the entire decentralized finance empire. In 2026, Ethereum will no longer try to convince the world with narratives, but will prove itself with engineering capabilities. This transformation is not only Ethereum's "engineering survival" under competition and real-world pressure, but also a redefinition of "what ETH is." When users no longer care about the underlying L1, and when ETH's value capture model shifts from gas sales to security and settlement, ETH must find a new narrative to establish its position in the digital world. Whether Ethereum can successfully transform, and whether ETH can capture the value of its thriving ecosystem, will be a key issue that quantitative finance practitioners and all finance enthusiasts must closely monitor in the coming years.