Codename BONSAI! - Proposal for Building out SushiSwap on Solana and Serum

I see, well, that’s certainly one of the important aspects cleared up then. I thought it was still in the works, thanks.

What was the proposed deal from Raydium in terms of token transfer to Sushi exactly, and at what valuation?

Fees, performance and protocol dynamics

Many questions were brought up during the town hall and in the forum about the numbers behind Raydium so we did our best to address them here:

1) How will fees be split between LPs, Sushi and Raydium?

Our basic proposal for Sushi is that Raydium will earn 30bps for all the trades against its liquidity. 20bps will go to LPs, 5bps will go to RAY stakers and 5bps will go to Sushi.

2) How do Serum fees work and how is Raydium able to earn 30bps per trade for Bonsai?

When Raydium sends a limit order to the order book and it’s traded against, its able to receive a rebate of 3bps or 5bps through Serum depending on if its holding the Sushi community’s MSRM (a MSRM is equal to one million SRM) as offered by the ecoSerum Sushi grant discussed in the original proposal. This would be the base rebate for Bonsai and would be sent to a Raydium wallet address to be collected and traded for SUSHI.

Assuming we hold a MSRM, the rest of the fees will be generated through additional slippage from trades. An easy way to do this is to send our orders at a more profitable price. For ease of calculation, let’s assume a trade for SUSHI with a market price of $100 right now. We would decompose the K = X*Y equation to find out how many units we’re willing to sell to the market at $100.01. Assume that number is 1 SUSHI token. What our AMM would do is instead of selling that 1 SUSHI token at $100.01, we would sell it at $100.26, thereby earning an additional 25 cents which equates to about 25bps of extra profit. When put together with the 5bps rebate from Serum, this adds up to 30bps. In this scenario, both a trader buying on Bonsai and a trader buying from SushiSwap’s Ethereum pools would receive an identical number of tokens for the same trade outside of gas fees.

3) So why trade on Bonsai or Raydium if the cost is the same?

Aside from cheaper gas fees, Raydium also leverages Serum’s orderbook. By sharing liquidity using a decentralized central orderbook, SushiSwap traders can also trade against all of Serum’s other market makers and traders. This additional liquidity reduces the slippage of traders and enables them to execute better trades compared to executing the trade within the AMM liquidity pool itself.

4) Won’t competition from other liquidity providers impact Raydium’s volume and profits?

Of course it will but there’s a very easy metric we can use to measure this impact. Daily Turnover / Total Value Locked. Put simply, we want to know how much our liquidity is traded against daily. Currently, our ratio is at about 15% which is slightly higher than Sushi’s 13% for WETH-WBTC. This shows that we’re still getting a good flow of transactions and trading volume which means more fees and profits for LPs.

5) How do you deal with arbitrageurs in a fast moving and low fee environment?

We believe that lower execution times and cheaper gas fees actually increase our earnings. To really understand it, let’s have a look at the charts below:

Figure 1: Chart from BTC on March 3rd

Impermanent losses (IL) occur when the price moves largely in one direction but does not revert. A good example would be if the price of the pool went from point A to point B. The pool would have more BTC in it at point B while BTC is at a lower price. The way an AMM makes money is if the price moves from point A to B then C. At this point, the BTC and USD in the pool are at the same levels as when it was at point A and profit through fees were generated. To summarize, IL occurs when prices move in one direction. When the price reverts, IL cancels out and trading fees to the LP are then realized as profit.

Since both the SushiSwap Ethereum pools and Bonsai Raydium pools expect to start and end at the same price, the question comes down to what happens in the middle and which one reverts more often to capture more profit from fees. This is the point where Raydium really shines. Due to lower gas fees, arbitrageurs are able to trade more often and are able to keep the price more inline with actual market prices and create smaller mean reverting movements.

Figure 2: Chart from BTC on March 3rd

In figure 2, we draw theoretical movements for what would happen in a lower gas fee environment. Arbitrageurs on Raydium would likely be able to capture the price movements from A-B-C, C-D-E, E-F-G and G-H-I. Whereas in a high gas fee environment, an arbitrageur might only be profitable doing the A-B-C movement from figure 1.

6) So if you’re saying traders get better slippage, LPs capture more profits while arbitrageurs can still make their share, then who’s losing?

All this is made possible by super low gas fees. SushiSwap alone currently spends $200,000 a day on ETH gas fees. That’s over $70 million a year going to miners and electric plants. While we do support everyone doing their share and maintaining the integrity of Ethereum, this is also $70 million a year that is being taken out of the Sushi ecosystem. On the other hand, large arbitraging firms also spend similar amounts on ETH. The fees these firms save are also partly earned by Raydium when they keep prices more in line with actual market prices.

7) What is this Fibonacci sequence you’re talking about?

Sorry, that sounds fancier than it really is. This simply describes how far away from the mid-price that we send our orders. Assume mid-price is $100 and we want a base spread of 25bps from the midpoint price. Then we take the Fibonacci sequence without the first number (1, 2, 3, 5, 8, 13, 21, 34, 55, 89) and multiply it by that base giving us (25, 50, 75, 125, 200, 325, 525, 850, 1375, 2225) bps and that’s how far away we price orders from the middle of the orderbook. The prices for an example sell side order book would look like (100.25, 100.50, 100.75, 101.25, 102, 103.25, 105.25, 108.5, 113.75, 122.5). It just works out nicely and squeezes a lot of orders in near the front of the book where they’re more likely to transact while also providing liquidity for a sudden price movement of up to 22% away from the mid-price.

1 Like

The governance proposal is up on snapshot! Voting will be between March 10 1400UTC and March 15 1400UTC:


Thanks for everyone’s feedback on this, the follow up proposal for snapshot can be found here:


Whats up with making the community snapshot? I thought with all the discussion, the two teams could work out a better deal and make it into a core proposal?

Food for thought.

I’m not opposed to cross chain development, but I don’t know how much value Radium brings to the table vs something like Loopring that’s on Ethereum and working now.

The whole point is more decentralization.
More affordable = more people = more decentalized.
tbh i can’t afford use eth, and exploring other solution chain.
and solana is more wise choice between bsc / heco.
please don’t make people hard to join because egostatic minds = which sides is gains more ?
this is for public goods, when publiic goods means nothing, so i’m surrender to convince


Bonsai will be added to a snapshot with some minor tweaks and options for xSushi holders.


I agree with some points by @Fiskantes that it seems this proposal benefits Raydium more than SushiSwap. But I also think partnering with other legitimate platforms is a way to get more users, which in turn is favorable for any DeFi platform.

My question is how does liquidity work for SushiSwap in other chains? I tried using SushiSwap on another chain and the liquidity is almost nonexistent. Even a small token swap causes a large price impact.

If Raydium liquidity can prevent this, then it may be a good decision to help small capital traders and investors participate in the SushiSwap ecosystem.


In my humble opinion, focus on ETHL2’s instead of brancing out to more centralized chains. Im in crypto for decentralization of financial products, not to shift power from one centralizd entity to another. I hate gas fees as much as the next guy, but we have options coming, and exponential scaling when L2+ETH2 is live.