A&T Capital

What Does the Ideal Cross-Chain Bridge Look Like?

Written by Todd Zhu, Vice President of Investment at A&T Capital.

(Thanks to our interns Fuming and Steven for their support on the research)

Illustration: Saidler & Co.

Every time we talk about cross-chain, we always appreciate Polkadot for the thrilling picture it painted for the crypto industry in 2016. However, we still need to take these points into our consideration:

(1) The realization of heterogeneous chain information on the cross-chain basis is extremely difficult.

(2) Current asset cross-chain bridging have already met the needs from DeFi in the short-term.

The article will not cover information on the cross-chain basis, while only discuss asset cross-chain bridging. (Unless otherwise noted, all following “cross-chain bridges” refer to “asset cross-chain bridges”)

Before we dive into the cross-chain bridge industry, let us recall what is the best cross-chain bridge you ever used? I am not sure whether there are any fellows who share the same answer with me — if you don’t take the issue of decentralization into account, the experience using CEX for on-chain transactions is fantastic!

Now, please keep the answer in your mind, let’s move on.

The time of cross-chain bridge is now.

The past one year witnessed the booming of heterogeneous chain ecosystems (i.e., Matic Network, BSC and Solana). Regarding the congestion on the Ethereum, a situation is gradually formed in which the Ethereum users want to go out of Ethereum and try out other chains while the heterogeneous chains are in need of more liquidity. Correspondingly, a set of easy-to-use cross-chain bridges becomes increasingly important.

(P.S. DeFi genesis block mining on heterogeneous chains is charming)

Now, let’s have a glance at the cross-chain asset flow network composed by a series of bridges pursuant to the graph below:

The optimization of cross-chain bridges:

① The chart below is drawn from following dimensions to evaluate some of the mainstream cross-chain bridges currently on the market.

② Degree of Decentralization (to some extent, the degree of decentralization is equivalent to the degree of cross-chain security);

③ The efficiency of the cross-chain process (covering the time-consumed by the cross-chain process, user’s experience, and the capital efficiency)

*The bridge part of Aurora has an information cross-chain embryonic form.

Identify cross-chain bridges by their technical implementations:

To understand the whole picture of cross-chain bridge, we will start from the simplest and most mature cross-chain bridge solution-wBTC:

Taking the following scenario as an example:

Alice plans to transfer one BTC to her Ethereum account via wBTC. She needs to fulfill the steps below:

① Transfer: Alice pledges one BTC to a pledge address managed by a multi-sig address controlled by wBTC bridge on the Bitcoin network.

② Confirmation: wBTC detects Alice’s pledge from the Bitcoin network and informs this to its smart contracts in the Ethereum network.

③ Minting: The smart contracts of wBTC on Ethereum mint a wBTC in the Ethereum network based on the information obtained from the step above and transfer it to Alice’s Ethereum address to complete Alice’s BTC cross-chain request.

Let’s have a summary of the three steps above. In the transfer and confirmation process, Alice needs wBTC bridge to complete two core steps for her:

a. Inform the target chain (Ethereum) the specific information and progress of Alice’s cross-chain requirements, or say an “oracle”.

b. Give Alice the assets that Alice should obtain in the target chain (Ethereum), or say asset minting.

“Oracle” (focusing on the degree of decentralization and the coverage):

Degree of decentralization:

Binance Bridge (the “oracle” is centrally controlled by Binance) >>>

Anyswap (the “oracle” is controlled by the consensus accepted by all nodes on Conflux);


wBTC (the “oracle” can only cover one cross-chain channel, i.e., from Bitcoin to Ethereum) >>>

Shuttle Flow (the “oracle” can simultaneously connect Conflux with Ethereum, BSC, etc.)

Asset Minting: (focusing on the degree of decentralization, speed, acceptability of the minted asset)

Regarding the asset minting is more difficult to understand than “oracle”, we have to draw reference from the four options of cross-chain assets proposed by DeFi innovatior, Andre Cronje, in this March (if you already know these stuff, you can jump to the part “single-chain bridge VS multi-chain bridge”).

Take cross-chain transaction of 1 USDC from Ethereum to Fantom as an example:

a. Balance Float: After locking 1 USDC in the cross-chain smart contracts on Ethereum, the “oracle” will notify the cross-chain smart contracts on Fantom, and the contract on Fantom will send 1 USDC to the recipient’s address in Fantom.

Because the total amount of assets circulating in the cross-chain bridge is huge, the cross-chain process will only involve the net changes in the balance of various assets on the cross-chain bridge (the same for in reverse, i.e. from Fantom to Ethereum; if the control of the balance is in the hands of a few individuals, there may be a risk of over-issuance), and does not involve the minting and burning of assets >>> This method is broadly used in centralized cross-chain bridge, e.g. Binance Bridge.

b. Mint/Burn: After locking 1 USDC in the cross-chain smart contracts on Ethereum, the “oracle” will notify the cross-chain smart contracts on Fantom, and the contract on Fantom will mint 1 USDC and send it to the recipient’s address in Fantom (while in reverse, the USDC minted on Fantom will be burned);

The cross-chain bridge itself will not hold any assets on Fantom when there is no asset locked in the Ethereum smart contracts, which means, there will be a 1:1 token holding only after receiving the minting requirement. >>> As a more basic framework, numerous cross-chain bridges employ the mint/burn design, especially single-chain bridges, e.g. Token Bridge crosses Dai to xDAI (The minted tokens has to be accepted by target chain, i.e., xDai can be used directly on the xDAI chain. Thus, this design cannot perfectly support the projects with multi-chain bridge).

c. Liquidity Swap: After locking 1 USDC in the cross-chain smart contracts on Ethereum, the “oracle” will notify the cross-chain smart contracts on Fantom, and the smart contracts on Fantom will send the 1 anyUSDC it holds (assumed using Anyswap) to the recipient’s address on Fantom, and offer an anyUSDC swap for USDC (giving that the applications on Fantom basically only accept USDC rather than anyUSDC).

The above process, except for the last step, is the same as the option a (i.e., there may be a risk of over-issue). For the last step, compared with the two options mentioned in a and b, it proposes that there must be a liquidity pool for each chain within the cross-chain framework >>> The example of this option is Anyswap v2.

Starting from this scheme, since the assets that users ultimately use in the target chain are not wrapped tokens minted by the cross-chain bridge, but the tokens that are universally accepted on the target chain, the transferred assets is absent from the endorsement support of the cross-chain bridge. AC reckoned it as an option “does not require custody”, while the security has been highly improved than before (the possibility of USDC price collapse is much lower than anyUSDC).

d. Wrapped + Mint/Burn: This option is a combination of b and c, that is, on the basis of option b, every anyUSDC on Fantom is minted based on USDC locked in other chains at a 1:1 ratio;

In addition to its inheritance of the security from the non-custodial solution, since anyUSDC is generated from the decentralized smart contracts at a 1:1 ratio, the risk of over-issue is avoided >>> Currently, this is my most favorable option while it is also the multi-chain asset generation option with the highest security. Such option is adopted by Anyswap V3 and Shuttle Flow.

Single chain bridge VS multiple chain bridge:

Based on the summary and comparison of the current decentralized cross-chain bridges on the market that support or do not support multi-chain and cross-chain function, we can conclude these differences at the institutional level are drawn:

Single-chain bridge: Most of the single-chain bridge projects put the “oracle” + consensus structure (for cross-chain process) in the original chain or the target chain, which is highly bound to one of those two chains, that is, only to facilitate cross-chain process in one channel.

Multi-chain bridges: Compared with single-chain bridges, multi-chain bridges mostly make the structure of the oracle + consensus structure (for cross-chain behavior) independent from any target chain or original chain (essentially, it does not need to highly depend on any chains), so that in the case of crossing structure with N chains, the number of bridges that need to be built is N instead of N*(N-1)/2.

(P.S. Indeed, this structure is very similar to the Relay Chain concept from Polkadot and Cosmos, and I have to admit how farsighted these “cross-chain OG” are)

I know from this article, many folks just want this table:

If you already have an understanding of “oracle”, asset minting, and single/multi-chain, then the following mapping will be easier to understand:

Before we dive deep into the technology implementation, we can look at the actual data of each bridge, and it is not difficult to draw two conclusions:

a. The technical implementation is not the most important. To attract more cross chain asset, the key is to have the target chain attractive for the asset/capital.

b. To be honest, despite the difficult using a bridge, if the Dapp in the target chain is attractive, the official bridge (no matter centralized or not) is the most popular one.

Every bridge has its own features, which are highlight in the table. We will give you folks a summary below:

1. The degree of decentralization >>> This is straight-forward, doesn’t need further explanation.

2. Multi-Chain bridge >>> At present, the relay chain seems like a necessity, it would be great if there is an another interface can transfer across different chains;

3. Decentralization of the “oracle” >>> It would be better to have a chain-level consensus to ensure the correctness of recording of cross-chain behaviors. If it is not convinced enough when the consensus is based on several groups of individuals, issuing governance token to improve the decentralization of governance is also proper solution.

4. Pledging + Minting/burning is the best “asset minting” technical solution currently >>> If you don’t fully understand this point, I highly recommend to go over the original text of AC with a link is attached at the end; (if the tokens minted by the cross-chain bridge is in the target chain are as common as USDT, then the step of building liquidity pools can be omitted)

5. Improve the utilization of locked assets >>> Let the assets flow in to DeFi to earn interest rather than merely be locked.

6. What chains the bridges can connect to (permissionless chains are better) >>> Essentially, someone need to provide liquidity, just like Uniswap’s listing new tokens.

7. Competence to support the public chains that are not compatible with EVM >>> The development will be more complicated, the new teams should be careful.

8. Competence to support NFT.

9. It would be best if a single cross-chain transfer can be completed within a few minutes.

10. Eliminate the obstacles for users of reserving gas fee for each chain.

11. Competence to sent to any address, not only to the target chain address binding with Metamask >>> In theory, it is feasible to realize this feature by changing the token transferring and receiving address of the smart contracts to the asset receiver address.

12. No transfer limit.

13. It is very difficult to realize all 12 features above. The proper solution is to make a trade-off.

Further requirements and exceptions:

Regarding the industry with high creativity, I always believe that the projects can learn from their competitors. Thus, I want to put down some irrelevant stuff here but can stimulate you guys’ further thinking. After all, no one could figure out the potential of AMM before Hayden Adams pointing it out.

Further requirements: cross-chain liquidity:

The high-quality assets outside Ethereum are continuously increasing. And the emergence of DeFi that focusing on cross-chain liquidity (e.g., Linear, O3Swap) indicates the further requirements for cross-chain information and cross-chain assets: cross-chain bridge should embrace a closer connection with DeFi, not only just fulfill its basic function, i.e., transferring assets across chains.

Further requirements: cross-chain for Layer 2

In order to prevent liquidity fragmentation and destruction of DeFi’s composability, four commonly used Layer 2 solution including Optimism will certainly need a expressway for asset transfer, or even a channel to let assets flow between Layer 2 and heterogeneous chains.

The efficiency of routing from Layer 1 is too low, and the consensus mechanism of Layer 2 is also born to be unable to carry the aforementioned “pledging + minting/burning” feature in the short term, so the state channel seems to be the most anticipated of current feasible solutions, e.g., Celer Network, Connext.

Exceptions: Starkware Caspian & ThoreChain:

Although Starkware Caspian is a mind-blowing solution based on Layer 2 and ThoreChain is a solo player in the cross-chain world of Layer 1, to a certain extent, the goal they want to achieve are the same: to complete the settlement without influencing the integrity of the original base layer liquidity pool.


Now let’s come back to the answer you kept in your mind at the beginning. You should have an answer about what an ideal cross-chain bridge should look like.

Have a nice day!

We are looking for fundamental blockchain infrastructures and innovative business models to accelerate the deployment of blockchain technology of our future.