Ethereum's Real Problems and What Is Being Done About Them
Ethereum settled the smart-contract question and inherited four harder ones: throughput, security, fees and governance. A sober look at each problem, what causes it, and which fixes have shipped versus which are still promises.
Ethereum won the smart-contract platform race. It is the second-largest cryptocurrency by market cap and the settlement layer for most of DeFi and whatever remains of the NFT market. Winning bought it a harder set of problems, the ones that come from actually being used. Four stand out, and it is worth being precise about which fixes have shipped and which are still promises.
1. Throughput
The base chain processes roughly 15 transactions per second. That number is not an oversight. It is the price of letting anyone with modest hardware verify the chain, and it is hopeless for global usage, so when demand spikes, the shortage shows up as congestion and fee auctions.
The strategy Ethereum settled on is to stop scaling the base layer directly and scale through rollups instead: Layer-2 networks that execute transactions elsewhere and post compressed results back to Ethereum for security. The base chain's job shifted to serving those rollups cheap data space. EIP-4844 (proto-danksharding, shipped March 2024) added blob storage for exactly that, and full danksharding is the long-term extension of the same idea. The old plan of sharding execution across the base chain itself was abandoned in favor of this.
2. Security
The protocol itself has held up well. The losses, and they run to billions cumulatively, happen one layer up:
- Smart-contract bugs: an exploitable contract is an open vault; code is law until it is a hack.
- Staking concentration: large pools controlling a big share of Proof-of-Stake validation is a standing centralization worry.
- Plain fraud: phishing and fake sites take more user money than any protocol flaw.
The countermeasures are unglamorous and incremental: better audit culture and formal verification of contracts, incentives for solo staking, wallet-level protections against approval scams. None is a silver bullet. Together they have made each ETH locked measurably harder to steal than in 2021.
3. Fees
Gas fees are a pure auction for scarce block space, which is why they swing from cents to hundreds of dollars with demand. The realistic answer was never "make L1 cheap." It is "move activity to L2s," where EIP-4844 cut typical rollup transaction costs to cents. On the base chain itself, fees remain what they are: the cost of settling on the most secured smart-contract layer available. For most users the practical fix is defaulting to an L2 and touching L1 rarely.
4. Governance
Ethereum upgrades happen through off-chain rough consensus among core developers, client teams, stakers and users. It is slow and argumentative by design. That is what not having a CEO looks like. But it concentrates informal influence in ways that are hard to audit, and regulators in most jurisdictions are still deciding what ETH legally is. Proposals for more formal on-chain voting exist, with their own capture risks, since token voting hands power to the largest holders. This problem is less solved than the technical ones, mostly because it is not technical.
Why a trader should care
These are not abstract concerns. Congestion and fee spikes are direct signals of on-chain demand. L2 adoption reshapes which tokens have a reason to exist. Every major upgrade is a volatility event for ETH and its ecosystem. Ethereum's bet, a deliberately slow and maximally secure base chain with speed provided one layer up, is coherent, and since 4844 it is visibly working. Whether it holds off the faster monolithic chains is one of the live questions of this cycle. The two companion posts on Layer 1 upgrades and Layer 2 solutions cover the mechanics behind it.