Parachain Auctions — What We’ve Learned
Hypersphere was formed in 2018 based on the idea that the Polkadot ecosystem would become massive, requiring specific knowledge to navigate and help build the network. Because of our familiarity with the ecosystem, we are able to build conviction, and concentrate on use cases and technologies we believe will offer greater results.
Fast forward two years, and we are pleased to have led the seed rounds and/or advised eight out the first eleven parachains. We wish we had contributed to all eleven because they’re all exciting projects. Our funds have so far contributed more than 35,000 Kusama (or 0.3% of the total supply of KSM and 1.5% of all KSM contributed) to Kusama parachains, and we are excited to support portfolio projects Bit.country and Composable (Picasso) in the continuous slot auctions currently in progress.
If you are new to Polkadot, Kusama, please read the Polkadot blog’s parachain primer.
Parachain Crowdloan Returns
Beating the staking rate is the primary goal of parachain crowdloan contributors. Roughly 2.4M KSM, or $1B USD, is bonded in parachain slots. This is more than 20% of the supply of KSM locked for almost a year! Each parachain team so far has bid for the maximum duration, forty-eight weeks.
Out of the five projects that were awarded parachain slots in the first five auctions, the average return on investment (ROI), measured in KSM, for bonding KSM tokens was an astounding 302%. This compares to staking reward of ~15.9% APY for staking KSM (KSM tokens contributed to crowdloans do not receive staking rewards during the lease period). This staking rate is the opportunity cost for contributing to crowdloans. For long-term holders who don’t intend on selling their KSM, participating in crowdloans has so far yielded a considerably higher return than staking.
However, our sample size to measure return is limited to the five tokens that are currently transferable, and among these there are idiosyncrasies. For example, Phala and Khala will use the same PHA token, and Bifrost and KILT communities are planning to move their parachains from Kusama to Polkadot when their Kusama leases are complete, thus forgoing issuing two tokens. The following table is based on the number of tokens issued per KSM contributed:
Based on the percentage unlocked at launch, if someone had invested in all of these projects equally, she would have earned an 86% return just a few weeks after her initial crowdloan investment, with the remaining 216% of the 302% return coming over the course of the lease. This remaining return unlocks linearly over time and thus the return can fluctuate depending on how parachain token price changes during the vest.
Breaking Down Parachain Candidates Success
Parachain crowdloans are a new innovation in crypto. While it’s difficult to draw causal links, we have analyzed various metrics / qualities to measure why a parachain might garner more support from DOT and KSM holders. For example, we know the community looks at the following characteristics:
Technical solution and how it fits into the ecosystem
Tokenomics of the parachain auction
Community support (size and activity on Twitter, Telegram, Discord; which parachains have contributors with large KSM or DOT holdings)
Accessibility of the auction (which exchanges and wallets support the parachain)
For the first batches of parachains on Kusama, it appeared that the community focused on key infrastructure projects that can grow into ecosystems in and of themselves. DeFi platforms (KAR) and EVM-compatible generalized networks (MOVR, SDN, KAR) received the most support and won the top 3 slots.
As the auctions progressed, other technical solutions with more niche offerings came into play, such as online identities (KILT), privacy (PHA, KMA), and tokenized assets (AIR). We expect this move towards parachain specialization as a necessary evolution of the Polkadot and Kusama ecosystems as more parachain slots come online and cross-chain messaging (XCMP), and eventually SPREE, matures.
Tokenomics of the parachain auction
Each project has the ability to construct the tokenomics of their parachain offerings according to their needs. Projects can choose:
Duration of the lease
Percent of the total token supply allocated to crowdloan rewards (total token rewards ranged from 2% to 30% of total initial supply, with an average of 14%)
Unlock and vesting schedules (percent of token rewards unlocked at launch ranged from 0% to 50%, with an average of 18%. The remaining percent (72% avg) vested linearly, over an average of 55 weeks. The the majority vesting over~48 weeks, or the duration of the lease. The exception here was Bifrost, which vests over 3.5 years.
Implied valuation to break even on rewards versus staking rate (assuming a constant KSM price). There was a fairly large range here, with a low of $39.5M FDV for MOVR, a high of $642M FDV for KILT, and an average of $198M FDV to break even versus staking. Projects which had lower implied FDVs to reach breakeven experienced the highest ROIs — with MOVR contributors earning highest ROI (1,346%), followed by SDN’s ROI of 66%. Note: this is a sample size of only 5 projects due to the other 6 projects not having their token launches yet.
Incentives for early backers, whales, and referrals. Most projects had some sort of incentivization mechanisms — whether it be incentives for early participants, referral rewards, whale incentives.
We thought one of the more interesting mechanisms was Basilisk’s (BSX) crowdloan strategy, which aimed to incentivize the minimum amount of KSM contributions necessary to win while not overpaying for a slot. For example, if you contributed to BSX’s crowdloan when the project was in the lead for a slot, you would have received less HDX (not BSX) per KSM contributed than if you contributed when the project was in second or third place in the auction. The team thus set-up the crowdloan mechanism in a way that created synergies between their Polkadot-based HydraDX community and their Kusama-based Basilisk community while also disincentivizing more KSM contributions when they were the clear front-runner to win the auction.
This HDX bonus is an important point because we’ve witnessed with Moonriver’s launch how a quickly a Kusama-based parachain community can grow. Projects should develop a strategy for creating alignment between their Kusama-based parachain and their Polkadot-based parachain from the outset.
Notes on the table:
Implied token price to reach breakeven on KSM is calculated by analyzing the opportunity cost of contributing KSM (~15.9% APY for 48 weeks), versus the # of tokens received per KSM from participating in the crowdloan
Implied FDV (fully diluted valuation) is the resultant breakeven token price multiplied by the total supply
Implied token price is for BSX only, HDX token price of $0.08059 was used for the HDX tokens (closing LBP price)
Unsurprisingly, projects with strong, pre-existing community support generally performed well. A few factors we noticed that helped to garner community support:
How long the project has been building (most projects that won parachain slots in the first two batches have been building for >20 months, with an average of 23.4 months)
How quickly a parachain launched its candidacy during an auction (Karura was the first to submit its candidacy, and several parachains didn’t submit their candidacy until the first auctions were complete. For the first few days in the continuous auctions, Picasso was the only candidate for the 12th slot, which helped gather initial support)
Grants or stamps of approval from W3F grants, Substrate Builders program, Wanxiang Incubator, Berkeley Xcelerator, Hypersphere as lead investor ;) etc. The first 10 projects that won parachain slots participated in the Substrate builders program.
How many unique contributors were there? We didn’t notice a significant correlation between unique number of contributors and project success Potentially because this metric can be gamed unless there is an anti-sybil mechanism like off-chain KYC.
Size of Telegram / Discord communities and Twitter following. The first six parachain slot winners clearly had stronger social media presences’ than other participants.
Notes on the table:
Khala Network does not have its own Twitter. Phala Network, which is the team’s Polkadot project, has 112.3k followers.
Heiko Finance does not have its own Twitter. Parallel Finance, which is the team’s Polkadot project, has 33.7k followers.
Parachain auctions are a new, innovative mechanism to both raise relay chain tokens (DOT, KSM) to bond for a lease, and distribute tokens to a project’s supporters. Parachain tokens can be thought of as machine-to-machine tokens. Parachains mint tokens to lease security and interoperability from the relay chain.
Initial crowdloans went as smoothly as anyone could have hoped. Projects implemented a variety of on-chain and off-chain incentives and tactics to secure a highly coveted parachain slot.
While we analyzed a number of different metrics here, more than anything we found that the projects with the most community mindshare going into the parachain auctions were generally the most successful in raising KSM. As time goes on and more parachains are launched, we will have more data to work with and will likely be able to draw more significant conclusions. We expect future projects to also have less of a first-mover edge, and thus the project’s technical solution, incentives/implied valuation, and community will be more instructive of which parachains will be most popular.
Looking ahead to the Polkadot parachain auctions, one can expect that crowdloan participants will favor projects that were successful in their Kusama-equivalent auctions. Additionally, many participants will likely be attracted to Kusama-equivalent projects that saw high ROI’s, such as Moonbeam’s Moonriver (1,346% ROI) and Astar’s Shiden Network (72% ROI).
Needless to say, it is an exciting time to be a contributor and we are excited to see what kind of innovative decentralized, cross-chain applications are built atop parachains.
Thank you to Lucas Vogelsang, Marvin Tong, Lurpis Wang, Jakub Panik, and Ingo Ruebe for contributions to this post.