The turning point of the 10-year debate: Can Ethereum truly solve the "trilemma"?

Article by: imToken

The term “trilemma” has probably made your ears numb by now, right?
In the first decade since the birth of Ethereum, the “trilemma” has been like a physical law hanging over every developer— you can choose any two among decentralization, security, and scalability, but never all three at once.
However, looking back from the beginning of 2026, we find that it seems to be gradually transforming into a “design threshold” that can be crossed through technological evolution, as Vitalik Buterin pointed out in a disruptive perspective on January 8:

Compared to reducing latency, increasing bandwidth is safer and more reliable. With PeerDAS and ZKP, Ethereum scalability can be increased by thousands of conflicts, and decentralization does not conflict with it.

So, can the “trilemma,” once considered insurmountable, really dissipate in 2026 with the maturity of PeerDAS, ZK technology, and account abstraction?
1. Why has the “trilemma” remained unsolvable for so long?
We need to revisit the concept of the “blockchain trilemma” proposed by Vitalik Buterin, which was specifically used to describe the dilemma faced by public chains in balancing security, scalability, and decentralization:

• Decentralization means low node barriers, broad participation, and no trust in a single entity;
• Security means the system can maintain consistency against malicious acts, censorship, and attacks;
• Scalability means high throughput, low latency, and a good user experience;

The problem is that these three often hinder each other under traditional architectures. For example, increasing throughput usually means raising hardware requirements or introducing centralized coordination; reducing node burden might weaken security assumptions; insisting on extreme decentralization can inevitably sacrifice performance and user experience.
It can be said that over the past 5-10 years, different public chains have given different answers— from early EOS to later Polkadot, Cosmos, and the ultimate performance seekers like Solana, Sui, Aptos, etc.— some sacrificed decentralization for performance, some improved efficiency through permissioned nodes or committee mechanisms, and others prioritized performance with free verification and validation.
But the common point is that almost all scaling solutions can only satisfy two of the three at the same time, inevitably sacrificing the third.
Or in other words, almost all solutions are stuck in a tug-of-war under the “monolithic blockchain” logic— if you want to run fast, you need strong nodes; if you want many nodes, you have to run slow. This seems to be a dead-end proposition.
If we temporarily put aside the debate over the pros and cons of monolithic vs. modular blockchains and carefully review Ethereum’s transition from a “monolithic chain” to a multi-layer architecture centered on “rollups” since 2020, as well as the recent maturity of ZK (zero-knowledge proof) and other supporting technologies, we will find:
The underlying logic of the “trilemma” has been gradually reconstructed over the past 5 years through Ethereum’s modularization process.
Objectively, Ethereum has, through a series of engineering practices, decoupled the original constraints one by one. At least in terms of engineering pathways, this problem is no longer just a philosophical discussion.
2. The engineering solution of “divide and conquer”
Next, we will break down these engineering details and see how Ethereum has advanced multiple technical lines in parallel during the five-year empirical period from 2020 to 2025 to dissolve this triangular constraint.
First, by decoupling data availability through PeerDAS, freeing the inherent limit of scalability.
As is well known, in the trilemma, data availability is often the first bottleneck for scalability because traditional blockchains require each full node to download and verify all data, which guarantees security but limits scalability. This is also why DA solutions like Celestia, which emerged in the last cycle (or even the cycle before that), have seen explosive growth.
Ethereum’s approach is not to make nodes stronger but to change how nodes verify data, with PeerDAS (Peer Data Availability Sampling) as the core idea:

It no longer requires each node to download all block data but instead verifies data availability through probabilistic sampling—block data is split and encoded, and nodes randomly sample parts of the data. If data is hidden, the probability of sampling failure rapidly increases, allowing significant improvements in data throughput while still enabling ordinary nodes to participate in verification. Does this mean it’s not a step toward optimizing decentralization through structural improvements?

Vitalik emphasizes that PeerDAS is no longer just a conceptual plan but a real deployed system component, which means Ethereum has taken a substantial step forward in balancing “scalability × decentralization.”
Next is zkEVM, which attempts to solve the problem of “whether each node must repeat all computations” through zero-knowledge proof-driven verification layers.
The core idea is to enable the Ethereum mainnet to generate and verify ZK proofs. In other words, after executing each block, a verifiable mathematical proof can be produced, allowing other nodes to confirm correctness without re-executing all transactions. Specifically, zkEVM’s advantages focus on three aspects:

• Faster verification: nodes do not need to re-execute transactions, only verify zkProofs to confirm block validity;
• Lighter burden: significantly reduces full node computation and storage pressure, making light nodes and cross-chain verifiers easier to participate;
• Stronger security: compared to the OP approach, ZK state proofs are confirmed on-chain in real-time, with higher anti-tampering capabilities and clearer security boundaries;

Recently, the Ethereum Foundation (EF) officially released the L1 zkEVM instant proof standard, marking the first time ZK technology has been formally incorporated into the mainnet-level technical plan. Over the next year, Ethereum mainnet will gradually transition to a zkEVM-supported execution environment, shifting from “re-execution” to “verification proof” structure.
Vitalik judges that zkEVM has preliminarily reached a stage suitable for production in terms of performance and functionality, with the main challenges being long-term security and implementation complexity. According to EF’s published roadmap, block proof latency is targeted to be within 10 seconds, with individual zk proofs less than 300 KB, using 128-bit security, avoiding trusted setup, and planning to enable household devices to participate in proof generation to lower decentralization barriers.
Finally, besides these two, there are ongoing efforts based on Ethereum’s 2030 roadmap (such as The Surge, The Verge, etc.) to improve throughput, reconstruct state models, raise Gas limits, and enhance execution layers.
These are all trial-and-error and cumulative paths in crossing the traditional trilemma. They resemble a long-term mainline, aiming for higher blob throughput, clearer Rollup division of labor, and more stable execution and settlement rhythm, laying the foundation for future multi-chain collaboration and interoperability.
Importantly, these are not isolated upgrades but are explicitly designed to be layered and mutually reinforcing modules, which also reflects Ethereum’s “engineering attitude” toward the trilemma: not seeking a magic solution like a single monolithic chain, but redistributing costs and risks through multi-layer architecture adjustments.
3. The 2030 vision: Ethereum’s ultimate form
Even so, we must remain cautious. Because elements like “decentralization” are not static technical indicators but long-term evolutionary results.
Ethereum is gradually exploring the constraints of the trilemma through engineering practices—as verification methods (from re-computation to sampling), data structures (from state bloat to state expiry), and execution models (from monolithic to modular) change, the original trade-offs are shifting, and we are infinitely approaching that “want, need, and must have” endpoint.
In recent discussions, Vitalik also provided a relatively clear timeline:

• 2026: With some improvements in execution layers/construction mechanisms, after introducing directions like ePBS, the Gas limit without relying on zkEVM can be increased first, creating conditions for “more widespread zkEVM node operation”;
• 2026–2028: Adjustments around Gas pricing, state structure, and execution load organization to maintain security under higher loads;
• 2027–2030: As zkEVM gradually becomes a key method for verifying blocks, Gas limits may further increase, with the long-term goal of more decentralized block construction;

Combining recent roadmap updates, we can glimpse three key features of Ethereum before 2030, which together constitute the final answer to the trilemma:

• Minimalist L1: L1 becomes a solid, neutral bottom layer responsible only for data availability and settlement proofs, no longer handling complex application logic, thus maintaining extremely high security;
• Thriving L2 and interoperability: Through EIL (interoperability layer) and fast confirmation rules, fragmented L2s are stitched into a whole, users do not perceive the chain’s existence, only experiencing hundreds of thousands of TPS;
• Extremely low verification barriers: Due to mature state processing and lightweight client technology, even mobile phones can participate in verification, ensuring the cornerstone of decentralization remains unshaken;

Interestingly, as I write this article, Vitalik again emphasizes an important testing standard—the “Walkaway Test”—reaffirming that Ethereum must have the ability to operate autonomously, even if all service providers (Server Providers) disappear or are attacked, so that DApps can still run and user assets remain safe.
This statement essentially brings the evaluation criteria of this “ultimate form” back from speed/experience to the most critical aspect Ethereum cares about—the system’s trustworthiness in the worst-case scenario, and whether it remains independent of single points.
In closing
People always need to look at problems with a developmental perspective, especially in the rapidly changing Web3/Crypto industry.
I also believe that many years later, when people look back at the intense debates about the trilemma from 2020–2025, they might think it’s like the serious discussion about “how to make horse-drawn carriages simultaneously fast, safe, and capable of carrying loads” before the invention of cars.
Ethereum’s answer is not to make painful choices among the three vertices but to build a digital infrastructure that belongs to everyone, is extremely secure, and can carry all human financial activities through PeerDAS, ZK proofs, and clever economic game design.
Objectively speaking, every step forward in this direction is stepping on the end of the “trilemma” chapter.