Ethereum 2029 Roadmap: Rebuild itself from the ground up, but the ship cannot stop

Author | James/Snapcrackle

Translation | Deep Tide TechFlow

Introduction: Ethereum researcher Justin Drake released “Strawmap” — the first structured upgrade roadmap for Ethereum with clear timelines and performance goals. Vitalik called it “very important” and described the overall effect as a “Ship of Theseus” style reconstruction. This article is the clearest popular science explanation of Strawmap so far, covering its principles, five major goals, and seven upgrades—all understandable even if you’re not technical.

The full text follows:

Ethereum has just released its most detailed upgrade plan ever. Seven upgrades, five goals, one large-scale rebuild.

If you’re wondering who this guide is for… it’s for me.

Ethereum researcher Justin Drake released what he calls “Strawmap,” a seven-major-upgrade timeline extending to 2029. Ethereum co-founder Vitalik Buterin called it “very important,” describing the cumulative effect as a “Ship of Theseus” style reconstruction of Ethereum’s core.

This metaphor is worth understanding.

The Ship of Theseus is an ancient Greek thought experiment: if you replace each plank of a ship one by one, is it still the same ship?

That’s what Strawmap proposes for Ethereum.

By 2029, every major component of the system will be replaced. But there’s no planned “big shutdown” or overhaul. The goal is backward-compatible upgrades—replacing parts while keeping the chain running—though each upgrade still requires node operators to update their software, and edge cases may occur. This is a full rebuild disguised as a gradual upgrade. Strictly speaking, while both consensus and execution layers are being reconstructed, the state (user balances, contract storage, history) is preserved across all forks. “The ship is being rebuilt while carrying its cargo.” Everyone, get on board!

“Why not just restart from scratch?” Because you can’t reboot without losing what makes Ethereum valuable: the applications running on it, the flowing funds, the trust built over time. You must replace the planks while the ship is still sailing.

“Strawmap” is a portmanteau of “strawman” and “roadmap.” A strawman is a preliminary proposal known to be imperfect, meant for critique and debate. So this isn’t a promise but a starting point for discussion. But it’s the first time Ethereum builders have laid out a structured, time-bound upgrade path with clear performance targets.

The work involves the world’s top cryptographers and computer scientists, and it’s all open source. No licensing fees, no vendor contracts, no corporate sales teams. Any company, developer, or country can build on it. JPMorgan will benefit from these upgrades just as much as a three-person startup in São Paulo.

Imagine a coalition of top engineers rebuilding the internet’s financial infrastructure from scratch, and you can directly connect to it.

How Ethereum Works (60-Second Version)

Before discussing where it’s headed, let’s clarify what it is today.

Ethereum is essentially a shared global computer. Not operated by a single company on one server, but by thousands of independent operators running copies of the same software.

These operators independently verify transactions. Some are called validators, who also stake their ETH as collateral. If a validator tries to cheat, their staked ETH is confiscated. Every 12 seconds, validators reach consensus on which transactions occurred and in what order. This 12-second window is called a “slot.” Every 32 slots (about 6.4 minutes) makes an “epoch.”

Finality—the point when transactions become irreversible—takes about 13 to 15 minutes, depending on where your transaction falls in the cycle.

Ethereum’s processing speed is roughly 15 to 30 transactions per second, depending on transaction complexity. In comparison, Visa processes over 65,000 transactions per second. That gap is why most Ethereum applications today run on “Layer 2” networks—independent systems that bundle many transactions and then submit summaries back to the main chain for security.

This system of reaching consensus among operators is called the “consensus mechanism.” Ethereum’s current mechanism works well and has been battle-tested, but it was designed for an earlier era, limiting the network’s capacity.

Strawmap aims to fix all these issues, one upgrade at a time.

Strawmap’s Five Core Goals

The roadmap is organized around five goals. Ethereum is already operational, with billions of dollars flowing daily. But it has real limitations on what can be built. These five goals aim to eliminate those limits.

1. Fast L1: Finality in seconds

Today, confirming a transaction on Ethereum takes about 13 to 15 minutes—until it’s truly irreversible.

Solution: Replace the engine that all operators agree on. The goal is to achieve finality within a single slot through a single round of voting. A leading candidate is Minimmit, a protocol designed for ultra-fast consensus, still under development. The key goal: finality within one slot. Slot times themselves will also be compressed: proposals are 12 seconds → 8 → 6 → 4 → 3 → 2.

Finality isn’t just about speed; it’s about certainty. Think of wire transfers—the time between “sent” and “settled” is the window where errors can occur. If you’re settling a million-dollar payment, bond transaction, or real estate deal, those 13 minutes of uncertainty matter. Compressing to seconds fundamentally changes what the network can do—beyond native crypto apps, anything involving value transfer.

2. Gigagas L1: 300x faster

Ethereum mainnet currently processes about 15 to 30 transactions per second, which is the bottleneck.

Solution: Strawmap aims for 1 gigagas of execution capacity, roughly translating to 10,000 transactions per second (depending on transaction complexity). The core technology is “zero-knowledge proofs” (ZK proofs).

The simplest way to understand: now, each operator must recompute every transaction to verify correctness—like every employee redoing their colleague’s work. Secure but highly inefficient. ZK proofs let you check a compact mathematical proof that the computation was correct, with minimal work and trust.

Generating these proofs is currently slow—minutes to hours for complex tasks. Compressing this to seconds (a 1000x speedup) is an active research challenge, not just engineering. Teams like RISC Zero and Succinct are making rapid progress, but it’s still cutting-edge.

A mainnet with 10,000 TPS and quick finality means simpler systems, fewer moving parts, and less room for errors.

3. Teragas L2: Cross-fast channels at 10 million transactions per second

For truly large-scale transactions (and custom needs), Layer 2 is still necessary. Today, L2’s capacity is limited by what the mainnet can process.

Solution: “Data availability sampling” (DAS). Instead of every operator downloading all data, they check random samples and verify the full dataset mathematically. Like checking if a 500-page book is on the shelf by flipping through 20 random pages—if all are present, you can statistically confirm the rest.

PeerDAS is live on Fusaka, laying the groundwork for Strawmap. Extending this to full capacity involves iterative upgrades: each fork adds more data capacity and stress-tests network stability.

A 10 million TPS L2 opens the door to applications impossible on current blockchains—global supply chains with digital tokens for each product, millions of IoT devices generating verifiable data, micro-payments at fractions of a cent. These workloads are too large for existing networks, but at 10 million TPS, they’re easily handled.

4. Post-quantum L1: Preparing for quantum computers

Ethereum’s security relies on math problems that are hard for today’s computers—used in signatures and consensus. Once quantum computers are powerful enough, they could break these, allowing forgery or theft.

Solution: Transition to new cryptography (hash-based schemes) believed to resist quantum attacks. This is a late-stage upgrade because it affects nearly everything—larger data sizes (kilobytes instead of bytes), changing network economics, block sizes, bandwidth, and storage.

Quantum threats may still be years or decades away, but building infrastructure that might hold trillions of dollars in value must consider this now.

5. Privacy L1: Confidential transactions

By default, everything on Ethereum is public. Unless you use privacy tools like Railgun, ZKsync, or Aztec, every transaction, amount, and counterparty is visible to anyone.

Solution: Build privacy directly into Ethereum’s core. The goal is to verify transaction validity (sender has enough funds, math checks out) without revealing details. You could prove “this is a legitimate $50,000 payment” without revealing who paid whom or for what.

Current workarounds exist. EY and StarkWare announced Nightfall on Starknet in February 2026, bringing privacy to L2. But these add complexity and cost. Embedding privacy into the base layer eliminates the need for middleware.

This also intersects with post-quantum work: any privacy solution must be quantum-resistant. Both must be solved simultaneously. Solving this major obstacle could unlock widespread adoption.

Seven Forks (Upgrades)

Strawmap proposes seven upgrades, roughly every six months, starting with Glamsterdam. Each upgrade is deliberately limited to one or two major changes, so if issues arise, the cause is clear.

After Fusaka (already live, laying groundwork with PeerDAS and data tuning), the first upgrade is Glamsterdam, restructuring how transaction blocks are assembled.

Hegotá follows with further structural improvements. The remaining forks (I to M) extend to 2029, gradually introducing faster consensus, ZK proofs, expanded data availability, quantum resistance, and privacy features.

Why until 2029?

Because some problems are genuinely unsolved.

Replacing consensus mechanisms is the hardest. Imagine replacing an airplane’s engine mid-flight with thousands of co-pilots needing to agree on every change. Each change requires months of testing and formal verification. Eventually, compressing cycle times below 4 seconds hits physical limits: signals take about 200 milliseconds to circle the Earth, and at some point, you’re racing the speed of light.

Making ZK proof systems fast enough is another frontier. Current speeds (minutes) versus target speeds (seconds) differ by about 1000x, requiring mathematical breakthroughs and specialized hardware.

Expanding data availability is less difficult but still complex—mathematics is open, but operating a real-time network holding trillions in value requires caution.

Post-quantum migration is a nightmare at the operational level, as new signatures are larger and change the entire economic model.

Native privacy adds political sensitivity: regulators worry privacy tools enable money laundering. Engineers must build solutions that are private enough to be useful but transparent enough to meet compliance, and quantum-resistant.

All these upgrades depend on each other. You can’t scale to 10,000 TPS without mature ZK proofs, nor expand L2 without data availability work. These dependencies shape the timeline.

Considering the scope, three and a half years is already ambitious.

2029?

First, there’s a variable. Strawmap explicitly states: “The current draft assumes human-led development. AI-driven development and formal verification could significantly shorten the timeline.”

In February 2026, a developer named YQ bet Vitalik that a single AI agent could program the entire Ethereum system for a 2030+ roadmap. Weeks later, he released ETH2030: an experimental Go client claiming 713,000 lines of code implementing all 65 items of Strawmap, tested on testnet and mainnet.

Is it production-ready? No. As Vitalik pointed out, it almost certainly has critical bugs, stub implementations, and AI has not attempted full versions. But his response is worth reading: “Six months ago, even such a thing was far beyond possibility; what matters is the trend… People should remain open to this possibility (not certainty! possibility): Ethereum’s roadmap will be completed much faster than expected, and with higher safety standards than anticipated.”

Vitalik’s core insight is that using AI isn’t just about speed; it’s about splitting the benefits—half for faster development, half for safety: more testing, more mathematical validation, more independent implementations of the same system.

The Lean Ethereum project is formalizing some cryptography and proof stacks through machine-verified code. Flawless code—long considered an idealistic fantasy—may truly become a basic expectation.

Strawmap is a coordination document, not a promise. Its goals are ambitious, its timeline visionary, and execution depends on hundreds of independent contributors.

But the real issue isn’t whether each goal is achieved on time. It’s whether you want to build on this platform or compete with it.

And all of this—research, breakthroughs, cryptography migrations—is happening openly, freely, and accessible to anyone… That’s the part of this story that should be getting far more attention than it currently does.