Ethereum: Reloaded

Why I Still Pay Attention

I began following Ethereum in 2013, before the mainnet existed. The original idea was unusually ambitious: turn a blockchain from a ledger into a place where anyone could publish and run software. It brought together economic systems, incentives, game theory, and products that did not depend on one company’s permission. I was interested because Ethereum was proposing a new foundation for products, not simply a new form of money.

More than a decade later, my view is divided. Ethereum succeeded as open, programmable infrastructure and created markets and standards that now support a large part of the crypto economy. As a consumer product platform, however, it has fallen well short of its promise. A person still has to understand wallets, gas, signatures, bridges, Layer 2 networks, and several kinds of risk before completing actions that a conventional application handles quietly in the background. The technology has matured, but using it still feels too much like operating the technology itself.

Programmable Infrastructure

Ethereum gave developers a shared execution layer and left the applications open-ended. That decision led to token standards, decentralized exchanges, lending markets, digital collectibles, governance systems, and many experiments that would have been difficult to coordinate through a single company. ERC-20 and ERC-721 became important because unrelated teams could build around the same rules. The transition to proof of stake showed something equally significant: a large, live network could replace one of its foundations while continuing to operate. These are substantial achievements, and any honest assessment of Ethereum has to begin by recognizing them.

Product Friction

Ethereum’s scaling strategy makes technical sense. Layer 2 networks can process transactions more cheaply and settle their results back to the base layer. The problem is that this architecture has divided the experience across rollups, bridges, wallets, sequencers, and different security assumptions. Users encounter every seam. They can own the correct asset on the wrong network, sign an approval they do not understand, or find that moving a small balance costs more than the action is worth. These may look like edge cases from the protocol level. From the user’s perspective, they are the product.

Network Activity

I do not treat the following charts as a single score for Ethereum. They show different parts of the system at different moments: where transactions took place, where developers deployed contracts, and what users paid to participate. Read together, they reveal a network that continued to develop while much of its activity moved away from the base layer. They do not prove that the overall experience became easier or that more people found a lasting reason to use it.

Mainnet Activity

Daily mainnet transactions grew as Ethereum found real uses, but eventually levelled off. This does not mean the wider ecosystem stopped growing, because a meaningful share of activity moved to Layer 2 networks. It does mean that protocol capacity and user adoption should not be confused. More room for transactions is useful, but it does not give people a reason to make them.

Development Moved to Layer 2

In this 2024 dataset, Optimism accounts for 70% of the recorded EVM contract deployments. The figure belongs to this particular dataset and should not be treated as a complete measure of every chain. Its direction is still useful. Developers were already placing much of the application layer away from Ethereum mainnet, which is exactly what the rollup strategy was designed to encourage. The consequence is that Ethereum increasingly depends on products and interfaces it does not control from beginning to end.

Fees Are a Product Problem

Gas is necessary because computation and blockspace are limited, but it is a poor concept to place in front of a customer. EIP-1559 made fees more predictable, and EIP-4844 reduced the cost of publishing rollup data. Neither change removes the basic trade-off: when demand rises, using the network becomes more expensive. Applications therefore have to decide whether to expose that cost, subsidize it, or redesign the action. That is a product decision, even when the cause sits deep in the protocol.

Usage Outlasted the Hype

On-chain categories overlap, so this breakdown should be read as a direction rather than a precise division of activity. A token swap is also a contract call, and the same address may participate in several categories. The broader pattern matters more than an exact percentage. The NFT surge faded, but Ethereum continued to settle a wider mix of financial and application activity. The network proved more durable than any single story used to promote it.

Scaling Moved the Problem

During this period, Ethereum’s base layer processed roughly 10–15 transactions per second. That can support a settlement layer, but it cannot carry every action for a mass consumer product. Rollups provide the additional capacity. They also introduce a new responsibility: balances, identity, security, and movement between networks must still feel like one coherent experience. Scaling is not finished from a product perspective while users need to understand the architecture in order to avoid a mistake.

Why Adoption Still Feels Early

Ethereum has created durable financial tools, but very few consumer applications are used by people who have not already decided to become “crypto users.” That distinction is important. Adoption will not become mainstream because more people learn how bridges or gas estimates work. It will happen when an application is useful enough that those details fade into the background. Ethereum-based products can preserve openness and user ownership while still giving people clear actions, understandable consequences, and a realistic path to recovery when something goes wrong.

Protocol Roadmap

Ethereum continued to make substantial protocol changes after the Merge. Pectra went live in May 2025 with improvements to programmable accounts, validator operations, and blob capacity. Fusaka followed in December 2025, bringing PeerDAS to mainnet and creating room for much greater rollup data throughput. Glamsterdam is planned for the fourth quarter of 2026 and focuses on how Ethereum builds blocks, processes transactions, and manages long-term state growth. I find this progress encouraging because it addresses genuine limits in the network. It also makes the remaining question more urgent: when will these improvements become obvious to someone using an application rather than studying the roadmap?

Path to Adoption

The next meaningful test is not a single throughput number. It is whether a person can move through the ecosystem without first learning its topology. Wallets should be able to select the right network safely. Applications should be able to bundle or sponsor routine transactions without hiding meaningful risk. Assets and identity should move across rollups without turning bridges into the center of the experience. These changes are less dramatic than a major protocol milestone, but they are much closer to what ordinary adoption requires.

Conclusion

Ethereum has survived difficult market cycles, changed its consensus system, and become infrastructure that other networks and applications genuinely depend on. Those achievements are real. But as a consumer product platform, Ethereum has largely failed to make decentralized applications feel useful, ordinary, or meaningfully better for most people. The value of programmable blockchains no longer needs to be proved in theory. The challenge now is to turn a layered technical system into one coherent experience without giving up the openness that made Ethereum important. If the ecosystem succeeds, people will notice Ethereum less at the exact moment they are able to use it more.

Roadmap