# Ethereum Thesis: Risks By [recizk.eth](https://paragraph.com/@recizk) · 2022-12-19 --- When coming up with an investment thesis, an investor needs to think about the risks involved to make a rational choice. Cryptocurrencies investors are already familiar, often tragically so, with the many risks they face when investing in cryptocurrencies. This post isn’t anywhere near exhaustive about those risks. In fact, it will just list the ones I believe are underexplored when talking about Ethereum, its future as money, and as smart contracts layer for financial and non-financial applications. Today, Ethereum is made by 2 separate layers: the consensus layer, the execution layer. Both layers have more than one production-level implementation, but the consensus layer has more client diversity. The vast majority of both validators and users run Geth as the execution client. Splitting Ethereum into 2 separate layers was a difficult and brilliant feat of engineering, enabling developers to ship upgrades to the consensus and execution layer in parallel. ### Execution Layer Risks The execution layer is a production environment that executes billions of dollars worth of transactions, touching tens of thousands of deployed contracts. Those contracts execute their bytecode on the EVM. But as domain expertise has accumulated and evolved, protocol developers now know that’s a very limited environment that could benefit from a lot of upgrades. Moreover, the EVM interactions with Ethereum’s state have also consequences on the execution layer scalability and the possibility of having trustless light clients. On one hand, Ethereum’s rollup-centric roadmap will reduce the pressure of having to upgrade the EVM as rollups can have somewhat more flexibility on the execution environment they ship (see FuelVM for example). On the other, most high-value applications will remain on the execution layer for the foreseeable future and they wouldn’t be able to take advantage of those features, which could include security upgrades such as improved cryptography and user experience. Moreover, improving the execution layer would also allow rollups to have fewer constraints for both optimistic and validity-proof types. Unfortunately, not all upgrades are backward compatible and, in fact, there is significant tension between guaranteeing backward compatibility and enabling new features and use cases. Smart contract code is immutable. Although some high-value applications have adopted update proxy patterns, it cannot be relieved by protocol developers, as a way to ship backward-incompatible code without affecting users. This is true for both intentional and unintentional backward-incompatible updates. I believe both intentional and unintentional backward-incompatible updates to be equally dangerous as they hurt Ethereum's claim of neutrality, security, censorship resistance. In particular, even if only one smart contract is affected in the name of the collective benefit, it would possibly unrecoverably hurt Ethereum’s shot at being self-sovereign money. Therefore, every time an execution layer update is shipped, there are significant understated risks for Ethereum’s investors. .. to be continued.. --- *Originally published on [recizk.eth](https://paragraph.com/@recizk/ethereum-thesis-risks)*