Ethereum is aiming to move beyond the throughput and programmability limits that characterize Bitcoin’s base layer, with a long-term ecosystem target of up to 100,000 transactions per second (TPS), according to Vitalik Buterin. The core thesis is that Ethereum can scale without sacrificing its programmability by pushing most execution off the base layer while preserving Layer 1 security.
Buterin framed the roadmap as more than incremental tuning. He is positioning these upgrades as a shift that turns Ethereum from a simple ledger into a higher-throughput, highly programmable platform, narrowing the bandwidth and cost trade-offs that constrain Bitcoin’s roughly 7 TPS base-layer throughput.
Now that ZKEVMs are at alpha stage (production-quality performance, remaining work is safety) and PeerDAS is live on mainnet, it's time to talk more about what this combination means for Ethereum.
These are not minor improvements; they are shifting Ethereum into being a…
— vitalik.eth (@VitalikButerin) January 3, 2026
Why Ethereum’s model differs from Bitcoin’s base layer
Bitcoin’s architecture prioritizes broad decentralization and a conservative scripting model, which in practice caps base-layer transaction capacity and limits on-chain programmability. This design choice favors restraint and simplicity, even if it leaves less room for complex on-chain computation.
Ethereum’s approach layers scalability mechanisms over a Turing-complete execution environment, aiming to keep base-layer security intact while moving most transaction processing and complex computation to secondary layers. The model is built around the idea that the base layer should secure the system while rollups and Layer 2 networks carry the bulk of activity.
Buterin has argued that enshrined protocol upgrades plus mature rollups can reconcile decentralization, security, and scalability, which he framed as solving the “blockchain trilemma.” That claim rests on a mix of live code already running in production and a staged pipeline of future upgrades that still need smooth rollout and adoption.
The upgrade stack behind the throughput target
A major pillar is data availability and verification efficiency, including PeerDAS and data-availability sampling, with prototypes described as going live on mainnet in 2025. The practical goal is to help light clients verify that shard or blob data has been published without downloading entire datasets, which increases usable data throughput for rollups.
Another foundational change is proto-danksharding (EIP-4844), implemented in 2025, which introduced cheaper dedicated data blobs for rollups and has already reduced Layer 2 transaction costs. This upgrade is presented as a near-term cost lever that makes rollups cheaper today while setting the stage for larger scaling later.
Looking into 2026, proposer-builder separation (ePBS) is described as a Layer 1 change intended to decouple block proposers from block builders to improve fairness, reduce censorship risk, and enable higher safe gas limits. In roadmap terms, ePBS is positioned as a key 2026 milestone that improves how blocks are built and may expand effective throughput headroom.
On Layer 2, the roadmap highlights ZK-EVMs and rollups as the path to high-assurance off-chain execution, with limited mainnet adoption expected to begin in 2026 and a broader shift toward ZK validation projected across 2027 to 2030. Optimistic rollups are already handling a large share of activity and delivering materially lower fees today, while ZK systems are positioned as the next step in verification maturity.
Buterin and the broader roadmap tie these pieces to measurable scaling outcomes: Ethereum’s base-layer TPS is typically in the mid-teens, roughly 15 to 30 TPS, while Layer 2 adoption plus full data sharding is intended to multiply effective capacity many times over. The 100,000 TPS ambition is therefore not framed as a Layer 1 number, but as an ecosystem-wide outcome driven by rollups plus deeper data scaling.
Near-term proof points already exist, especially where proto-danksharding reduced rollup costs in 2025 and Layer 2 networks account for a substantial portion of transaction volume. At the same time, the “trilemma solved” framing sits alongside real implementation and adoption risk, because the largest throughput gains still depend on ePBS, full danksharding, and broad uptake of ZK validation.
Investors and institutions are now focused on 2026 delivery, particularly ePBS and the initial ZK-EVM rollouts, as practical tests of the scaling thesis. The market impact will hinge on whether developers and users migrate workloads to rollups quickly and whether remaining Layer 1 changes land without friction, which will shape fee trajectories, liquidity distribution across L1 and L2, and execution risk for smart-contract-based products.
