Chainlink v. API3: A Case for Protocol Owned Liquidations

Ideas
, ,
1

Chainlink v. API3: A Case for Protocol Owned Liquidations

TLDR:

@ugurmersin from API3 has assembled a Dune Dashboard which shows that Compound has leaked over $5M in Liquidation value in 2025 – and we’re only into February.

Value accrual for Lending Protocols has changed. Both API3 OEV, and Chainlink SVR challenge the status quo by enabling more liquidation value to be returned to the DAO, rather than captured by Validators in exchange for preferential Order Flow.

These changes take place by introducing bidding at the Oracle Feed, rather than the Validator level – and then direct profits from those bids back to Compound.

Compound DAO has a responsibility to re-examine our business model and optimize our liquidation processes to close these leaks on every market, so that when volatility hits the broader markets, we can strengthen our treasury.

Context:

Our current system concentrates wealth amongst early players, with preferential access to infrastructure, but it doesn’t have to be that way – and those same players can still provide the bids necessary to handle liquidations for Compound, allowing that value to circulate back into our markets, rather than be lost to MEV.

As a DAO, we need to ask ourselves, “Where does the value for Compound Protocol accrue?”

Compound currently earns around $9M a year through borrows, and yet, 5 weeks into 2025, Compound has paid out $5.3M as liquidation incentives. That means that currently, the greatest benefactors of Compound are external actors, rather than the protocol itself.

API3 and Chainlink have solved this problem, and they have data to prove it.

How Does API3 OEV Function?

In the current system, a bidding war takes place to allow one Liquidator to out-compete all others in transaction priority. The bidding occurs after the Oracle Price Feed has given its information, and before the transaction has been entered at the Validator level. The highest bid is included in the block, and up to 99.9% of the profit of the liquidation is spent on bidding for that block placement.

API3 is a Price Feed Oracle which offers a bidding market for liquidators on which they can acquire oracle updates, and directs captured value back to Compound DAO.

Compound currently uses API3 as its oracle on the Mantle Deployment, and on February 4th, this bidding system was used on Mantle to return $150K to Compound that would otherwise have been bribed to validators. If we used this same tech on all of our markets, we would have been able to recapture most of the $5.3M mentioned in Ugur’s dashboard.

Chainlink SVR:

Chainlink has their own OEV solution called Smart Value Recapture (SVR). Chainlink’s articles and recent Aave case study are freshly published, as of the end of January.

Given that Compound already uses Chainlink on existing markets, it makes sense to inquire about the costs for upgrading our current Liquidation processes to SVR.

Choosing Both: Chainlink SVR & API3 OEV

As stated above, Chainlink’s SVR was recently announced and AAVE passed a proposal for a trial run. The proposal as well as the Chainlink Blog post outline technical and commercial aspects of utilizing Chainlink SVR.

Important factors to consider for Compound DAO are:

  • Currently Chainlink SVR is limited to Ethereum Mainnet due to the reliance on Flashbots.

  • Chainlink outlines a 60/40 split between integrating protocols and themselves

  • AAVE gets a preferred 65/35 split due to being a launch partner

  • The solution hasn’t been tested in production on scale yet as its development is fairly recent

API3’s OEV solution was announced in July 2024 and has been in use from multiple protocols across multiple chains including Compound’s USDe Comet on Mantle. Important factors for Compound DAO to consider are:

  • API3s solution works across all chains and can be used across all Compound deployments immediately

  • API3 outlines a 80/20 split between integrating protocols and themselves

  • the solution has been live for multiple months and has proven to be working for several dApps including Compound itself

  • the solution proved itself in the biggest liquidation event in DeFi history without any issues

Compound can choose to use both Chainlink SVR and API3, however, we cannot afford to continue without OEV.

Applying OEV Savings to Growth:

To drive the point home, $5.3M is the equivalent of 5 Arbitrum LTIPP Ecosystem Grants.

If AlphaGrowth deployed the $5.3M in a similar manner to LTIPP it would bring around $700M in new TVL to Compound, which would net Compound reserves an additional $4M/year in Fees and $XXM/year in OEV profits. Making Compound more sustainable.

OEV Savings might also be used to provide seed liquidity for chains like:

  • Celo
  • Mode

These chains have offered Compound $1 or $2M in incentives for deployment, but due to liquidity constraints we were unable to come to terms. With OEV Savings, we could form a seed liquidity fund to be matched by these growing ecosystems – then re-absorb that matched capital with interest once the markets have matured.

There are so many possibilities, but they are only available to us if we take the steps to adapt to what DeFi looks like in 2025.

Our Recommendations:

AlphaGrowth recommends Chainlink SVR and API3 OEV type solutions across all markets.

OEV represents DeFi pivoting in 2025, to empower Lending Protocols. Chainlink and API3 are leading this innovation, and Compound DAO must be responsive, and seize this opportunity for Growth.

Additional Resources:

7 Likes
2

Thank you for the writeup, @Kyle! Really appreciate it.

If there are any questions, feel free to reach out—I’d be more than happy to go over everything in detail.

When it comes to OEV and the critical need to recapture it for protocols, we’ve been at the forefront of this discussion for well over two years—ever since we introduced how to best recapture this value for protocols.

Early on, there was skepticism from the industry, especially from other oracle and risk providers. However, even the most reluctant players—including Chainlink—have now recognized its importance and are actively working on their own OEV solutions.

From day one, our approach was designed around three key principles:

  1. Independence: No reliance on third parties (unlike Chainlink’s or UMA’s dependency on Flashbots, which limits their solutions to Ethereum Mainnet).
  2. Scalability & Simplicity: OEV recapture should be a seamless, standardized, and easily replicable “copy-paste” experience across all chains for dApps.
  3. Decentralization with Safety Nets: While some centralization may be acceptable for OEV recapture, underlying core functions (like data feeds) must have decentralized fallback mechanisms baked in.

Our solution has been live and operational since July 2024, with multiple dApps using it successfully. These include:

  • Lendle (a smaller lending protocol)
  • Yei & Init Capital (medium-sized protocols that surpass some of Compound’s Comets in TVL)
  • Compound on Mantle, following a successful OZ audit

Most importantly, our solution just underwent the ultimate stress test:
On Monday, the industry saw the largest liquidation event in crypto history. A common concern has been that OEV solutions remain untested in extreme conditions. Even risk providers like Gauntlet have questioned whether they would hold up.

The results? Every liquidation was executed without failure across all protocols using API3 oracles, including Compound on Mantle. Significant value was recaptured across all platforms—Compound, Yei, and Lendle.

I’ve long advocated for Compound to adopt OEV recapture, but recent events have underlined that free money is being left on the table. Money that could be used to further grow Compound. Your biggest competitor is already moving forward with this, potentially expanding their advantage.

However, Compound has a unique edge right now:

  • API3 is already audited and proven in production for Compound.
  • Unlike others still piloting their solutions, we are ready to immediately replicate what we’re already doing on Mantle, wherever else you want. — All it takes is reading a new price feed!
5 Likes
3

It’s a convincing proposal. As someone that was involved with pioneering MEV for eth I am a little familiar with the process, how do you avoid being frontrun this way? It doesn’t seem like you are doing a private tx by skipping the public mempool.

2 Likes
4

Say Compound uses the ETH/USD feed of API3. API3 runs two feeds with identical configurations and security guarantees:

  • One global ETH/USD feed that anyone can update (and API3 updates it based on deviation)
  • One Compound-specific ETH/USD feed that the current OEV auction winner can update for the next 30 seconds

These two feeds are put behind a proxy that Compound uses to read them. The proxy simply returns the value of the feed that is updated with more recent data.

To ensure that OEV updates can frontrun global updates reliably, API providers delay the data they have signed for global updates for an additional 30 seconds before publishing them. See the link for implications Security considerations | Documentation

1 Like
5

Thanks @Kyle for opening up conversation for such critical topic as value leaking with liquidations!

As long time supporters of Compound, we’ve presented an alternative with no latency compromises that can scale to all networks supported by Compound currently and in the future.

Glad we can be part of the discussion.

3 Likes
6

Hey Matt - Grateful for your post. Thank you for joining the conversation, and sharing what Redstone can do.

It probably makes sense to organize some kind of comparative case study, to evaluate performance from different providers.

1 Like
7

Hey @Kyle, we’d be happy to chime in for a comparative case study. We’ve just witnessed the biggest liquidation event in history, so everybody should have sufficient data to showcase how their solution performed under the biggest stress test possible. When it comes to OEV, the worth of a solution can only be truly tested during such extreme market conditions. :slight_smile:

8

Thank you @Kyle for kicking off this discussion. We’d like to provide some additional context on Chainlink SVR, Compound’s historical usage of Chainlink oracles, and how Compound can capture liquidation MEV (OEV) in a risk-adjusted manner.

First, it’s important to note that this discussion isn’t just about which OEV recapture solution the Compound protocol should integrate, but also which price oracle provider Compound deployments will use, as the two are intrinsically tied together (i.e., using API3 for OEV also means using API3 for price oracle reporting). Modifying the core price oracle system used by Compound markets has serious security implications that could increase the risk of bad debt and insolvent markets, resulting in a loss of user funds if not considered carefully.

While recapturing OEV presents a real and tangible revenue opportunity for the Compound community, it is important to analyze this situation from a holistic perspective, which informs our recommendation of Compound to adopt Chainlink SVR for OEV recapture.

Background

Compound upgraded to Chainlink Price Feeds in June 2021 for the Compound V2 market on Ethereum, and integrated Chainlink at launch for the Compound V3 market on Ethereum in August 2022. Chainlink Price Feeds were adopted after $89M in user positions were incorrectly liquidated when the Coinbase Oracle (Compound’s former oracle provider) misreported the price of DAI, leading to value being siphoned from users. Chainlink Price Feeds have since become the default price oracle solution for Compound market deployments across chains, and continues to secure the primary and largest Ethereum V2 and V3 markets. Since integrating Chainlink nearly four years ago, there have been zero oracle-related exploits across Compound markets using Chainlink feeds, with Chainlink securing $12B in Compound TVL at its peak.

Through our long-standing collaboration with the Compound community, we have always taken a security-first approach, ensuring Compound markets have reliable access to accurate market data via Chainlink’s Decentralized Oracle Networks (DONs). Chainlink DONs leverage multiple layers of decentralization data aggregation, including the use the use of multiple independent oracle node operators, such as global telecommunication firms, who fetch and aggregate data from multiple professional data aggregation firms via paid data API subscriptions, to create tamper-proof oracle reports that are reliably published onchain even during the most extreme market volatility and blockchain network congestion conditions. This track record of securing billions of dollars in TVL for Compound over a period of nearly four years without any loss of user funds speaks to the integral nature of Chainlink’s usage by Compound.

It’s important to emphasize that the integration of any particular OEV recapture solution is inextricably linked to also integrating that OEV solution’s price oracle provider for market data. Therefore, it’s critical for protocols to not compromise on security by integrating an OEV solution linked to an unproven price oracle that can increase risks for the sake of rushing to capture a bit of extra revenue, as the risks to the protocol greatly outweigh the potential of small revenue recapture.

Expanding to using other, less proven price oracles in an attempt to recapture OEV would significantly alter the risk profile of the Compound protocol, potentially increasing risk for the end-users and billions of dollars in TVL.

Chainlink SVR

As the largest oracle in the industry by a wide margin, as measured by Total Value Secured (TVS) and number of integrations, we cannot afford to put DeFi at risk by taking shortcuts when creating and improving Chainlink services. There is no margin for error. When developing Chainlink Smart Value Recapture (SVR)—the OEV recapture solution native to Chainlink Price Feeds running on the same exact proven DON infrastructure—we took a systematic approach to ensuring security and reliability. Chainlink Labs and the broader Chainlink community have been actively researching MEV mitigation solutions for a number of years, including Fair Sequencing Services (FSS) and Protected Order Flow (PROF). As a subset of MEV, we’ve also engaged in research around OEV and how DeFi protocols can recapture this value and to help support the economic sustainability of oracles.

Rather than rush a product to market, we engaged in multiple cycles of R&D to arrive at an initial design that maximizes security, reliability, and long-term economic viability. Chainlink SVR was built in collaboration with industry-leading organisations such as Flashbots and other contributors to the Aave DAO, in order to meet the specific security and reliability requirements of the largest DeFi lending markets. SVR is undergoing testing on mainnet and is currently being integrated into the Aave V3 deployment on Ethereum pending Aave’s governance processes.

SVR also includes critical security functionality, such as in the case of transmission failure during an SVR auction, there is a fail-safe mechanism to ensure that the feed can still report an accurate price to DeFi protocols. This failsafe mechanism can be set at any threshold to manage risk according to an integrating protocol’s needs, highlighted in our recently published research report on SVR. Furthermore, SVR is not dependent on any third party bridges or external blockchain networks to process OEV auctions, reducing both reliance on third party dependencies and minimizing downtime risks. Chainlink SVR feeds are deployed natively on each supported blockchain and operate offchain OEV auctions using widely adopted MEV infrastructure. This maximizes searcher diversity and creates more competitive OEV auctions, leading to higher recapture rates for the protocol. More information about SVR can be read in the original SVR announcement and the follow up SVR research analysis.

Chainlink SVR is currently live on Ethereum and we are actively working on making SVR available across additional blockchains. Base and other OP Stack chains are expected to be supported by SVR in the near term, while chains such as Arbitrum, Scroll, Sonic, and other chains are being evaluated to be rapidly supported next. We are happy to collaborate with the Compound community to align our chain expansion plans for SVR. Additionally, future iterations of SVR are planned to introduce further improvements, including increased decentralization, enhanced gas efficiency, continuous price updates, and chain-agnostic functionality.

Recommendation

Based on the above considerations, we strongly recommend that the Compound community does not introduce unnecessary risk and continues to benefit from the proven security track record of using Chainlink Price Feeds. To build on this secure foundation, we recommend that Compound integrate Chainlink SVR as the OEV recapture solution of choice starting with Ethereum mainnet before expanding to other chains as SVR continues to become available across additional blockchain networks.

Beginning with an integration of Chainlink SVR for Compound markets on Ethereum mainnet would not only maximize the OEV recapture revenue opportunity compared to longer tail chains, but would also maintain the usage of Chainlink’s proven decentralized Price Feed oracle networks for core pricing data, protecting the protocol and its users.

2 Likes
9

I hate that these things always turn into a dick measuring contest, but here we go i guess. Thanks for opening up this discussion. To address some of the things above - Chainlink wants you to believe they’re the only project capable of building infrastructure in this space, but that simply isn’t the case. Leading projects are already proving this by going with solutions outside the “Chainlink Suite”— USDT <> LayerZero or Securitize <> Wormhole.

OEV isn’t the first time Chainlink has been caught off guard. When Pyth introduced pull-oracles, it was a game-changer for Synthetix. Chainlink spent months reacting, and eventually rolled out Data Streams—mostly because GMX forced their hand.

Now we’re watching the same thing unfold with Oracle Extractable Value (OEV). For years, Chainlink took a hard stance against MEV, saying Fair Sequencing Services (FSS) was “the way”.

Then in January 2024, UMA introduced Oval, their OEV product, and started an attempt to get Aave. Chainlink fought hard against it, but Aave wasn’t going to leave millions of dollars on the table just because Chainlink didn’t like the idea of recapturing MEV.

The result?

Chainlink, forced into action again, rushed to build SVR together with and for AAVE (which is essentially just a carbon-copy of oval with UMA cut out of the picture). The solution is everything that Chainlink critizied Oval for when they released it: They still delay data and they still utilize centralized players (a.k.a Flashbots Servers) in their solution while claiming “superiority”.

Let’s be clear: Chainlink didn’t want to build SVR. They were forced to by the market, because Aave wasn’t going to ignore fuckloads of money because “Fair Sequencing”.

In turn this means one thing for Compound: You’ll always play second fiddle to your biggest competitor. No matter if its in terms of innovation, revenue splits or expansion aspirations.

OEV solutions show their worth during heavily market volatility. The recent nuke—the biggest in terms of size—proved that we’re able to handle real stress without issues, while other solutions are mere theoreticals at this stage.

To summerize: Compound can continue to wait and choose to be treated as a secondary priority while being dependant on the goodwill of a project already having a favourite child, or Compound can choose to be made a priority. We can help Compound today, at a faster rate, with a better split, with a more proven solution, on all chains that comets are deployed on currently and that are in the pipeline.
Cheers.

PS: One thing to note. Compound is already utilizing API3 on Mantle. It is audited and contrary to what the above post makes you want to believe, other oracles are already making Compound a priority helping you expand and seize opportunities instead of waiting around.

3 Likes