Why is Cardano more decentralized than Ethereum?

The Ethereum community is sounding the alarm. The amount of staked ETH on Lido is approaching the first of three critical thresholds at which Ethereum is more easily manipulated by an attacker. Protocols that offer Liquid staking derivatives (LSD) are a stratum for cartelization and induce significant risks to the Ethereum protocol and to the associated pooled capital.

Why doesn't Cardano have the same problem? The simple answer is that it has liquid staking and users don't need to use LSD protocols. The IOG team managed to design Ouroboros Proof-of-Stake in such a way that its decentralization increases over time. Economic incentives motivate users to spread the total stake over multiple block producers. Ethereum may get into a situation where it becomes economically viable to use Lido for profit maximization reasons. This will further reinforce the monopoly. If the critical threshold is crossed, block production may be halted when Lido collapses.

TLDR

The quality of decentralization depends on small details in the PoS design. If stakeholders can delegate even a few ADA directly from the wallet and coins are liquid, decentralization flourishes. LSD tokens have a market value similar to ETH, but play no role in the Ethereum consensus. The centralization of ETH in Lido poses a serious risk.

Liquid staking and pool saturation

We see 2 main reasons why Ethereum has problems with the centralization of ETH. It's the need to have at least 32 ETH (~50K USD) to run the validator and locking of coins for a certain period of time (which is necessary due to slashing). Locking coins makes it impossible to delegate ETH in such a way as to make the coins liquid. If it is necessary to entrust the coins to a third party for any reason, the stakers lose control of the coins. From a usage perspective, users want ETH (or ADA) to be liquid. From a decentralization perspective, it is desirable that stakers retain control over the coins. It is necessary to reconcile these two principles well.

Staking on Cardano requires no minimum amount of ADA. During staking, staked ADAs are not locked (cannot be slashed) and users can spend them at any time. The key is that coins can be delegated to a pool. Delegating to a pool is essentially a decision to use someone else's infrastructure. This is because PoS coin holders usually don't want to run their own pool (node reliably running 24/7/365). They just want to make a profit. The IOG team was aware of this fact when designing PoS. They deliberately made staking in a way that it is possible to keep control of ADA and at the same time delegate coins' consensus power to whoever wants to produce the blocks (pool operators/SPOs).

It may not seem like such a fundamental difference to you, but in fact, these details determine the quality of decentralization.

ETH holders, like ADA holders, do not usually want to run their own block-producer node (pool/validator). If they want to make a profit, it is logical that they have to entrust the coins to someone who will run the nodes. While in the case of Cardano, coins can be delegated directly at the protocol level and stakers always have full control over the ADA (and the coins are also liquid), Ethereum needs a third party to achieve the same functionality. Lido, for example. ETH holders have to send ETH to the third party and in return, they get LSD tokens. LSD tokens have no meaning in the Ethereum consensus. However, LSDs have a market value close to ETH and users can sell them or use them in DeFi at any time.

PoS protocols can be attacked through native coins such as ADA and ETH, not through LSD. Centralizing native coins in one place (in the hands of one entity) is a critical problem and a single point of failure.

If there were more LSD protocols, maybe it wouldn't be such a big deal. Unfortunately, Lido may soon exceed the critical threshold of ⅓ of all staked ETH. How is it possible that people delegate more than 30% of all staked ETH to a single entity while Cardano's largest multi-pool operator (MPO) has less than 10%? This is due to the concept of pool saturation.

In the Cardano ecosystem, maximum rewards for pool operators and stakers can be achieved by having a preferred number of pools (defined by parameter k) that are saturated. The saturation point can be easily calculated by dividing the amount of all ADAs in circulation by the parameter k, currently 500. If you do the math, you get something like 70M ADA.

If the pool exceeds the saturation point, it will have a higher stake, but will still not receive higher rewards. Stakers have an economic incentive to delegate ADA to another pool that is not saturated.

Even this mechanism is not ideal. It does not prevent the emergence of entities that can operate multiple pools. In addition, blocks are produced by more than twice the preferred number of pools (approximately 1200). This means that instead of 500 saturated pools, there are a higher number of less saturated pools in the Cardano ecosystem. This is one of the reasons for the overall smaller staking rewards. The important thing is that there has not been the critical centralization of stakes that we see with Ethereum.

There is another reason that is worth noting. In order to make the decentralization of Cardano as good as possible from the beginning, i.e. from the transition to the PoS version enabling staking in 2020, pool operators were trained. There were hundreds of them right from the start. From my point of view, the Ethereum community underestimated the state of the ecosystem during the transition to PoS and basically allowed Lido to gain dominance. Maybe there should have been more education of stakers and more competing LSD projects should have been created. Ideally, however, the team should have delivered a PoS that did not require third parties to function.

Why is Lido a problem?

Lido (and similar protocols) is a stratum for cartelization and induce significant risks to the Ethereum protocol. If it exceeds critical consensus thresholds then the associated pooled capital is put in danger.

Lido is very close to exceeding the first critical consensus threshold which is 1/3 of staked ETH. The next critical thresholds are 1/2, and 2/3 of staked ETH. When this extreme situation happens, the staking derivative can achieve outsized profits compared to non-pooled capital due to coordinated MEV extraction, block-timing manipulation, or censorship of block space. In a scenario like this, staked capital becomes discouraged from staking elsewhere due to outsized cartel rewards, self-reinforcing the cartel’s hold on staking. In other words, it will be economically viable to stake ETH on Lido even though it will directly threaten Ethereum.

At the time of writing, 31.5% of all staked ETH is on Lido. Coinbase has 10% and Binance 5.4%. People predicted that after the Shapella upgrade (finally allowing unstake) people would migrate elsewhere and the amount of staked ETH on Lido would decrease. But that's not happening. The Lido team has no plans to limit the number of deposits (set the maximum below 1/3) and thus save decentralization.

Users behave rationally with respect to themselves first and with respect to others second. Stacking with Lido seems like a rational choice to people, but it is irrational with respect to Ethereum. Both protocol and capital are at risk.

Ethereum is a network in which there is an active vote on every block and it requires all nodes to be online. There is a risk that block production may be halted if more than 1/3 of validators go offline. This can happen due to a software bug or an attack (internal or external).

It's not nice to see that Lido has a chance to cross a critical threshold, and one entity will essentially be responsible for whether Ethereum will be able to produce blocks or stall. It's certainly a tempting spot for attackers. If Ethereum stops, there will likely be some sort of restart and probably a slashing of ETH.

Conclusion

The Ethereum team considers Lido to be an extreme risk. Influencers are sounding the alarm and asking people to delegate ETH wisely. If you stake ETH, listen to this smart advice. The Ethereum team, but also the community, should think about the quality of the PoS delivered. People were concerned about PoS because they thought centralized exchanges would stake large amounts of coins. It turns out that this is not happening to such a large extent. However, the Ethereum team delivered a PoS that gave rise to Lido, which appears to be an even bigger problem than centralized exchanges. This is a real shame. Decentralized finance cannot be built on a centralized protocol.

Cardano is years ahead of the PoS version currently used by Ethereum. This is mainly related to decentralization and staking. Ethereum has one advantage and that is the instant finality of transactions. Cardano has probabilistic finality of transactions, which means that it takes a relatively long time before a transaction can be considered forever stored in the ledger. These things are related. Ethereum must lock coins and have slashing, as it requires the constant and immediate active participation of nodes over each produced block. Cardano is able to produce a block even if only say 10% of the pools are online. The ultimate goal is fast finality, robustness, and decentralization. No project has yet achieved this ideal state.

IOG wants to keep the decentralization of Cardano (or even improve it) and improve the finality of transactions through the Ouroboros Leios upgrade. The Ethereum team has quietly given up on trying to improve scalability at the first layer. Unless this changes, Ethereum will not scale at the first layer, and moreover, it will be centralized. Cardano still has a chance to be the first protocol in the top 10 that will both be decentralized and scale very well (with the instant finality of transitions).

One more thought. The minimum for staking on Ethereum is 32 ETH. Currently, it is staked ~22,800,000 ETH. The maximum possible number of stakers (if each individual staked only 32 ETH) is 712,500. Of course, we are talking about the possibility to run a validator and we are not considering Lido (and others) where people can stake smaller amounts of ETH. Nearly 1.3M ADA holders stake on Cardano, which is almost double the number. Stakers trust Cardano more than Ethereum. Perhaps it is more accurate to say that significantly more people have direct control over Cardano consensus than in the case of Ethereum.