Decentralization is at the heart of the crypto space. When it comes down to it, everything we do and accomplish as an industry is meaningless without a clear path toward decentralization. If AAVE were centralized, it would have been closer to a Celsius; if Rocketpool were centralized, it would have been just another Coinbase staking service.

Decentralization as a buzzword is dear to our hearts, but people have only now started to really pay attention to it in extreme scenarios. The recent Tornado cash and OFAC debacle immediately comes to mind. Enforcing agencies were literally able to apply stringent measures against a protocol and its developers. The extreme measures taken by government and regulators have led to imprisonment and over-compliance by DeFi protocols and services (Uniswap, USDC…)

One of the most topical conversations to be had at the moment is, ‘how antifragile is the Ethereum ecosystem in the face of government scrutiny?’

Ethereum is about to complete its transition to a POS blockchain. This multi-year transition was initially conceived as a way to make Ethereum more decentralized and antifragile. What everyone is trying to wrap their heads around is if this transition will actually result in Ethereum being more antifragile and decentralized - or - if we are providing new tools for enforcement agencies to step in and regulate an open source public good.

The problem with validator concentration

Let’s skip over the PoS inner workings and basic staking concepts for now. Today, there is around 13 mil ETH at stake, and this number is projected to reach 30 mil ETH in the year following The Merge. Taking a closer look at staked ETH distribution, it becomes apparent that some have not achieved the decentralization goals that were set out following the Beacon Chain’s genesis.

At the network’s current state, 30% of the total staked ETH sits with Lido. Adding Binance, Kraken, and Coinbase to the pool will probably get us over 60%. The obvious conclusion is that Ethereum is dominated by 4-6 primary staking services that are able to wreak havoc and control the blockchain at their own will (if they wanted/were forced to). This scenario brings to light the systemic risk that centralization poses to Ethereum.

The question thus far was, what will happen if Lido (for example) gets a bug or if a good actor becomes malicious? The entire protocol will be at stake (no pun intended). With recent developments, the question has evolved to ‘how antifragile is the current system against involvement by nation-states?’ The answer is grim, to say the least.

As it stands, nation-states can approach the “big 5” staking companies fairly easily and get them to comply at gunpoint (threats of incarceration). If we’re not careful and thoughtful about how we approach this challenge, it will be the end of blockchain as we know it.

Radical decentralization

If it was not clear until now, the US government has made it more evident than ever before. The only way to achieve antifragility against nation-states is through radical decentralization. Nowadays, when you stake with Coinbase (one of many possible examples), they will be the ones running validators in the Ethereum protocol on your behalf. Running an at-home validator or staking through Coinbase is almost similar from the Ethereum protocol’s perspective. Every batch of 32 ETH staked today requires running a validator client connected to the Consensus and Execution chains (ETH1+2).

The current industry standard is fragile by design:

• Validator keys are being managed by a single validator
• Validators have a “single setup”, single cloud account, single ETH2 client, etc.
• Active to active redundancy is impossible due to the risk of slashing (i.e., we cannot deploy 2 or more validators to run in parallel)
• Active to passive redundancy opens stakers up to penalties for double attesting if not correctly set up

In the case where the US government approaches Coinbase or Kraken with a cease and desist order on specific addresses or transactions. They have nothing to do but comply. Shutting down their services will risk the protocol’s liveliness (imagine 20%+ percent of all validators suddenly going offline).

Decentralize Coinbase

One of the ways to stand against the inevitability of government intervention is through distributed systems and by introducing new technologies that will add robustness to the network and its participants. A relevant technology taking shape right now is DVT (Distributed Validator Technology). DVT allows the creation of ‘multi-node operators’, essentially creating a new type of validator which operates under consensus and is run by multiple nodes.

Instead of having a single node, with one setup managing a validation key. The same validation key is deconstructed into Shared keys, which are distributed between multiple node operators. The end result is a distributed validator that no one can censor effectively.

This means that Coinbase and the like can effectively take their validators and split them between 10,13, or more different service providers, with different setups, geo-locations, etc. The result will be an antifragile Ethereum validator. Done, everyone can continue buying their favorite NFTs uninterrupted.

Imagine the following scenarios:

Scenario A:

Uncle Sam -  “Hi Coinbase, you are holding 20% of the validation power on the Ethereum network. Please stop the such and such transaction from being included in your blocks.”

Coinbase: “Yes sir, would you like some coffee while you wait?”

Scenario B:

Uncle Sam - “Same question…”

Coinbase - “Sorry, sir, not my validators, not under my control. There is a group of 13 companies working under consensus to operate these validators. Some are in the US, some in India, Japan, and the list goes on. Why don’t you fly over to 13 different countries and try to convince all of them? Good luck… Oh, and do you want some coffee to go”?

DVT is not the end-all-be-all solution to our decentralization challenges, but it is a worthwhile technology to involve in the conversation. There are a few implementations out there that will probably reach mainnet sometime next year. Check out ssv.network and obel for a deep dive into the tech.