status

Can DeFi threaten Cardano's security?

Published 26.4.2023

Cardano's security is based on ADA coins' distribution and market value. This is based on the assumption that one entity cannot quickly and cheaply acquire a supermajority of all coins. The demand for ADA coins increases the market value, making a potential attack more expensive. There is one clever attack vector. It is possible to create a DeFi service offering a higher yield than delegators can get for staking ADA. For example, it could be a lending platform that offers a 10% yield for the ADA coins you provide to the service. Such a DeFi service would directly compete with staking. The service would siphon ADA from the staking and thus theoretically weaken the security of the PoS protocol. Can an attacker get the required number of ADA coins to attack in this way?

TLDR

Trying to obtain a supermajority of PoS coins through the DeFi service is a theoretically possible attack, but very likely unfeasible in practice. The attack would be very costly and it would be necessary to create the most attractive service on the market.

The nature of the attack

Stakers are economically motivated to delegate ADA to the chosen pool because they will be rewarded for doing so. This is say something like 5% per year. If a staker delegates say, 1000 ADA, he can earn something between 46 and 59 ADA per year depending on many factors.

Currently, 65% of the ADA coins in circulation are staked, which is approximately 24,000,000,000 ADA. There are almost 35,000,000,000 ADA coins in circulation. An attacker would need to acquire enough coins to have more than half of the coins that are staked. The attacker needs, say, 12,000,000,000 ADA.

He try to create a DeFi application that will offer a higher yield for ADA than what is offered by the Cardano protocol for staking. He will assume that people will be greedy and want to earn more. So they will stop staking ADA and use the attacker's DeFi service instead. By doing so, people will essentially hand him ADA coins.

This is how the attack could theoretically be carried out. Is this feasible in practice?

In practice, the attack is almost impossible

This attack on PoS networks was described back in 2019 and no one has been able to carry it out since. I dare say it is almost impossible to execute it successfully. There are many arguments why it cannot be done.

First and foremost, it is important to note that an attack would be very expensive, as in order to get a huge number of people to stop staking ADA and instead lock them into the DeFi service, the service would have to pay out the promised yield on a regular basis. That's a huge number of ADA coins. The attacker would have to give greater rewards than what the Cardano protocol offers for staking.

Every epoch (5 days) the Cardano protocol will reward stakers and SPOs with 10M ADA coins (4M USD). To be attractive, the DeFi service would have to offer at least double the reward, i.e. 20M ADA every 5 days. The attacker could run the staking pools himself or delegate the ADA. He would still have to give 10M ADA every 5 days out of his own pocket. The attack would probably be very long, as it is almost unrealistic for people to start using a single DeFi service in say a month and put in 50% of the ADA coins they were previously staking.

If the attack lasted only one year, the attacker would have to invest 730,000,000M ADA in the attack (at the current market value of ADA it would be about 300M USD).

The problem with the attack is that it is not at all certain to succeed. If the attacker only got 40% staked ADA, it would not be enough to attack. It would be possible to increase the yield to make the service attractive to the last people hesitating. But this would make the attack even more expensive. It is very likely that the attacker would be financially exhausted and the attack would end in failure. The attacker would have a pile of ADA coins that he would be forced to sell incrementally to compensate for the loss.

Assume that an attacker would be willing to invest hundreds of millions of dollars in an attack, perhaps even more if the attack took place in future years when the market value of ADA coins could potentially increase. What else makes an attack unlikely?

Our view is that some stakers will never use ADA coins in DeFi services, no matter how attractive the yield, because they consider it risky. A large portion of stakers will only delegate ADA coins to pools and do nothing else with them. If they do, it will only be with a portion of the portfolio. We imagine many people will stake say 60% of ADA coins and only use 40% in DeFi services. If all stakers behaved similarly, they would essentially prevent the attack.

Furthermore, it is very unlikely that a single service could be so attractive that almost all users would use it. The Cardano ecosystem is a very competitive environment. The more DeFi services there are, the more diversified the risk will be.

It is very difficult to create a service that would siphon off ADA from existing ones. Of course, the high yield may be a reason for change, but it certainly won't be appealing to all people. If a new successful service emerged, developers would try to copy the business model. If the new service was not transparent and had closed source code, many people would start warning against using it. This is a relatively big hurdle for an attacker.

If everything is as users expect from the DeFi service, it is not possible for a smart contract application to allow a team (attacker) to freely manipulate ADA coins. ADA coins are locked in the smart contract and depositors receive tokens through which it is possible to get ADA back at any time. The attacker would have to create a service that allows him to dispose of the ADA at will. If someone discovered this (back door), it would be a red flag.

Another big obstacle for an attacker is the fact that many DeFi services offer staking. This basically means that, for example, liquidity providers who provide ADA to a DEX are still stakers. The liquidity provider receives staking rewards plus a reward for providing liquidity. Thus, there is no economic incentive to change his behavior and start using the attacker's DeFi service. Unless the attacker offers a much more attractive yield than what DEX offers. This makes it much more difficult for an attacker to get more than half the amount of staked ADA coins.

The perception of ADA coins can change over time and this can affect the attacker's options. If people started to see the ADA as a store of value, it is very likely that they would only stake coins and not use them in DeFi. If people ever paid with volatile cryptocurrencies, they would keep them in their wallets (while staking). This would essentially make an attack impossible, since if a portion of the population used ADA instead of money, it could be a significant percentage of the total number of coins in circulation.

But even if it ever happened that only say 40% of ADA coins were staked and 60% were used in other ways, it's still very unlikely that all holders would be using a single non-transparent application with a shady business model, and with a suspiciously high yield.

Conclusion

Staking on the first layer of the Cardano protocol will always be the most secure yield in the ecosystem. This is certainly not to discourage you from using DeFi services but to be aware of certain risks. If you want to actively prevent the attack described above, stake a portion of your portfolio from your own wallet and use only a portion of the ADA in DeFi services. Try to use multiple DeFi services simultaneously if possible.

There is another active way to defend against the attack and that is decentralized governance. If there was a service that was suspicious and had a large number of ADA coins locked in it, the community could decide to increase the staking rewards. This would actively combat a competing DeFi service but at the cost of reducing the reserve.

This type of attack has not yet appeared on any existing PoS platform and it is very likely to remain theoretical forever. However, it is good to know that such a thing exists and to be vigilant if you encounter it. High yield without a working business model is always a red flag.

Featured:

Related articles

Did you enjoy this article? Other great articles by the same author