The Most Ambitious Software Roadmap in History
Ethereum’s development roadmap is one of the most ambitious in the history of computing: transform a public blockchain that currently processes 15 transactions per second into infrastructure capable of handling 100,000 transactions per second, while maintaining and even improving decentralization and security, with no downtime and no ability to pause the system for maintenance. This must be done through coordination among hundreds of independent developers, clients, and stakeholders who have no central authority requiring their cooperation.
Ethereum co-founder Vitalik Buterin has outlined this roadmap across five major phases he calls “The Surge, The Scourge, The Verge, The Purge, and The Splurge” — a conceptual framework for understanding what the protocol needs to achieve and roughly in what order. Understanding this roadmap is essential for understanding Ethereum’s long-term value proposition and the technical foundation supporting the ecosystem built upon it.
The Merge: The Foundation (September 2022)
Before discussing the future roadmap, it’s essential to understand what The Merge accomplished. In September 2022, Ethereum transitioned from Proof of Work (PoW) to Proof of Stake (PoS) in a seamless upgrade that had no downtime and did not interrupt any running applications. The Merge reduced Ethereum’s energy consumption by approximately 99.95% overnight — one of the most significant environmental improvements in the history of technology.
Beyond energy, The Merge established the validator-based consensus system that enables all subsequent roadmap items. Proof of Stake validators can be economically punished (slashed) for misbehavior in ways that PoW miners cannot, enabling stronger security guarantees. PoS also enables faster finality — blocks are considered economically finalized after approximately 12.8 minutes, a much stronger guarantee than PoW’s probabilistic finality.
The Shanghai/Capella Upgrade: Enabling Staking Withdrawals (April 2023)
The Shanghai upgrade (execution layer) paired with Capella (consensus layer) enabled validator withdrawals for the first time since the Beacon Chain launched in December 2020. Validators who had staked 32 ETH could finally exit their positions and withdraw accumulated staking rewards. This completion of the staking lifecycle — validators could now both enter and exit — was essential for the economic health of Ethereum staking and significantly expanded participation.
Dencun: The Blob Revolution (March 2024)
The Dencun upgrade (March 2024) was the most significant Ethereum improvement since The Merge. Its key component, EIP-4844 (Proto-Danksharding), introduced “blobs” — a new type of data specifically designed for Layer 2 rollups. Before Dencun, L2s posted their transaction data to Ethereum through standard calldata, which is expensive and permanently stored by all Ethereum nodes. Blobs are a separate data space that is cheaper, has a separate gas market (blob gas), and is only stored by Ethereum nodes for approximately 18 days before being pruned.
The impact was immediate and dramatic. L2 transaction fees dropped 10 to 100 times across all major rollups within days of Dencun’s activation. Transactions on Arbitrum, Optimism, and Base that previously cost $0.01-$0.50 dropped to $0.001-$0.01. For users, this made Ethereum’s L2 ecosystem genuinely competitive with dedicated high-throughput blockchains on fee grounds for the first time.
Dencun also initiated the “blob fee market” — blobs have their own target (3 per block) and maximum (6 per block), with fees rising when demand exceeds the target and falling when below. This separate market ensures blob demand doesn’t compete with regular Ethereum transaction fees, and vice versa.
The Surge: Full Danksharding and Maximum L2 Scalability
“The Surge” refers to the set of upgrades aimed at maximizing Ethereum’s throughput through rollup-centric scaling. The centerpiece is Full Danksharding — the completion of what EIP-4844 began. While Proto-Danksharding introduced blobs with a limit of 6 per block, Full Danksharding aims for 64 or more blobs per block, dramatically increasing data availability for L2s. With Full Danksharding, the aggregate L2 throughput could reach 100,000+ transactions per second across the ecosystem.
Full Danksharding requires Data Availability Sampling (DAS) — a technique where nodes sample random small pieces of blob data to probabilistically verify the entire blob is available, without any node needing to download all blob data. DAS is what enables the blob count to scale dramatically without requiring nodes to download proportionally more data, preserving decentralization as capacity increases.
The timeline for Full Danksharding is measured in years — it is a complex upgrade requiring significant cryptographic engineering and coordination. Interim solutions like PeerDAS (a simpler version of DAS implementable sooner) are being developed to increase blob capacity more quickly than Full Danksharding allows.
The Scourge: MEV and Validator Economics
The Scourge addresses Maximal Extractable Value (MEV) — the profit that can be extracted by controlling transaction ordering in blocks. MEV creates centralization pressure (validators who can extract more MEV earn more, incentivizing sophisticated, well-connected operations rather than independent validators) and negative externalities for users (sandwich attacks, front-running).
Attester-Proposer Separation (APS) is the primary Scourge upgrade: separating the role of the validator who proposes a block (who currently controls transaction ordering and thus MEV extraction) from the validators who attest to its validity. This separation prevents proposers from maximally extracting MEV while maintaining security, and can be combined with inclusion lists that prevent censorship of specific transactions.
The Verge: Verkle Trees and Stateless Clients
The Verge’s most important component is replacing Ethereum’s current Merkle Patricia Trie state structure with Verkle Trees — a more efficient cryptographic data structure that enables much smaller “witnesses” (proofs that specific data is in the state). Currently, fully verifying Ethereum blocks requires downloading and maintaining the full Ethereum state (hundreds of gigabytes). With Verkle Trees, nodes can verify block validity using only the small witnesses included in the block itself, without needing to store the full state — enabling “stateless clients.”
Stateless clients represent a major advance for decentralization: currently, running a full Ethereum node requires significant storage (a pruned node still needs hundreds of gigabytes of SSD storage). Stateless clients would reduce hardware requirements dramatically, enabling many more people to run nodes — including on consumer hardware, mobile devices, or resource-constrained environments.
The Verkle Tree transition is technically complex, requiring migration of all existing Ethereum state from one data structure to another in a live system — an enormously challenging engineering task being carefully planned by Ethereum core developers.
The Purge: Simplification and History Expiry
The Purge removes historical data and protocol complexity that has accumulated over Ethereum’s history. EIP-4444 (History Expiry) is the key component: nodes would no longer be required to store block history older than approximately one year. This reduces full node storage requirements from the current hundreds of gigabytes to manageable tens of gigabytes, improving node accessibility.
History older than the expiry window would be stored by a distributed network of “history archivists” rather than requiring every full node to maintain it. The Portal Network — a lightweight peer-to-peer network for serving historical data on demand — is being developed to fulfill this archival role. This means historical data remains accessible (via Portal Network) without burdening every node with its storage.
The Purge also includes deprecating rarely-used EVM features, simplifying the execution environment, and removing technical debt that has accumulated since Ethereum’s 2015 launch.
The Splurge: Miscellaneous Improvements
The Splurge encompasses various improvements that don’t fit neatly into the other categories. Key items include EVM Object Format (EOF) — a restructuring of the EVM’s contract bytecode format for better tooling and gas metering; Account Abstraction improvements (building on ERC-4337) to make wallets programmable smart contracts by default; and various cryptographic upgrades improving efficiency and preparing for post-quantum security.
Single Slot Finality: The Ultimate Upgrade
Single Slot Finality (SSF) is perhaps the most ambitious long-term goal: reducing Ethereum’s economic finality from approximately 12.8 minutes (two epochs) to a single slot (12 seconds). This would dramatically improve user experience and security for applications that need to wait for finality, but it requires significant consensus algorithm changes given Ethereum’s scale.
The Timeline and Investment Implications
Executing the full roadmap is a decade-long project. Dencun (done), Pectra (upcoming, 2024-2025, with EIP-3074 for account abstraction improvements and more), Fusaka (2025-2026, potentially including Verkle Trees and PeerDAS), and subsequent upgrades follow a cadence of approximately two major upgrades per year. Each upgrade is thoroughly tested on testnets for months before mainnet deployment.
For ETH investors, the roadmap represents a long-term fundamental driver of value: as Ethereum’s capacity grows, its utility expands, fee revenue increases, and the ETH burned through EIP-1559 becomes more deflationary — a direct value capture mechanism that makes ETH more scarce as network usage grows.
Conclusion
Ethereum’s roadmap is a coherent, technically rigorous plan for building the most capable and decentralized computing platform ever created. Each component addresses a specific limitation — scalability through blobs and danksharding, MEV through APS, node requirements through Verkle Trees and history expiry, client complexity through the Purge. The most remarkable aspect is that this development is happening on a live system handling hundreds of billions of dollars in value, with no ability to pause or revert, through coordination among hundreds of independent developers globally. If executed successfully, the result will be a blockchain capable of supporting truly mass-market global adoption while remaining decentralized — a combination that no existing computing platform has achieved.