Discover more from Dirt Roads
# 31 | Curve Wars: Plata o Plomo
You Mustn’t Be Afraid to Dream a Little Bigger Darling
For a brief, epic, moment, we had learned to love him. It was probably the nuclear bomb that almost shaved Neuer’s left ear off in that Warsaw summer night of 2012, or the legendary early morning drives to distribute cash to the local homeless population, or the defence of the bullied kid too afraid to go to school, or the indoor fireworks and consequent evacuations. In loving him, we were not alone. Mario Balotelli moved to Manchester, and Manchester City, escorted by a nourished group of friends, acquaintances, fans, ladies, and many other archetypes. Some were attracted by its raw football talent, some by the abs, others by the endless waterfall of cash that came with all that. It’s not a rare phenomenon. An agent is able to attract a disproportionate amount of resources thanks to some specific, uncommon, and sought-after abilities, and that attractive power spins an entire ecosystem. It works for the athletes, the musicians, the politicians, the sharks, the hippos, the entrepreneurs. It works for the DeFi protocols as well.
Ahead of the so-called DeFi summer of 2020 a few protocols learned how to attract, rapidly, immense amounts of liquidity thanks to their ability to deliver solutions that were vastly superior to all their rudimental competitors. This was the story of lending protocols like MakerDAO, Compound, or Aave, and of automated market makers like Uniswap or Curve. It didn’t take long for those protocols to find a way to capture part of the value generated (in the form of protocol fees) through the issuance in the open market of governance or utility tokens. Protocols quickly understood that the same governance tokens, originally crypto-proxies of traditional shares due to their ability to coordinate voting and share economics, could have been used as an additional means to attract and trap even more liquidity. Yield farming was born. Investors were parking liquidity in DeFi protocols not to benefit from the underlying protocol profitability, but in order to amass a continuous flow of freshly minted governance tokens to be sold immediately into a hungry market. Value, however, isn’t created from thin air and what was enjoyed by the early adopters needed to be compensated somewhere else. During those times, that somewhere were the deep pockets of overly optimistic new investors (that valued any token at crazy multiples) and the last men standing in the same protocols that were diluting themselves into oblivion via continuous minting. DeFi 1.0 was dubbed the age of decentralisation.
When the flood of liquidity coming into the market started to slow down investors, rather than accepting a yield reduction, opted for a dramatic increase in implied leverage. The ambition was not only to match the yields achieved by the previous generation - yes, generation, even if those were only months apart, but to surpass it by several orders of magnitude. As a consequence, tokenomics got more aggressive - Olympus, margins of safety way narrower - Abracadabra, and farming multi-layered - Curve. DeFi 2.0 was dubbed the age of capital efficiency. Yes, sure.
Curve Wars: There Can Be Only One
One day we will tell our grandchildren the tales from the Curve Wars. And in telling them we will remember whom we fought for, how much money we made, and where we stood when everything fell apart. Nothing characterises the second age of DeFi as good as the Olympus sensation and the Curve Wars; the union of the two might well be what can close it off for ever.
Layer 1: Curve, the Leading Low-Vol AMM
But what is Curve? At the beginning of the stablecoins era Curve, originally named StableSwap, was built as a more capital efficient AMM focused on trading low-vol asset pairs such as stables. The (healthy) coexistence of stables was predicated on seamless fungibility and exchangeability of the tokens, and the existing decentralised exchange solutions at the time (i.e. Uniswap) weren’t providing this at scale. StableSwap was born to offer, on the one hand, an AMM with less trading slippage across stables - by a factor of 100x, and on the other a deposit tool for stablecoin owners able to generate high interest through liquidity providing and lending engines.
The uniform constant product model pioneered by Uniswap, arguably the most impactful innovation in DeFi so far, wasn’t capital efficient and required enormous funds to offer meaningful and tractable liquidity. StableSwap was proposing an alternative that would have been extremely more efficient than Uniswap (in terms of price slippage) unless the relative pair prices would have deviated significantly - something that should not in principle ever happen for stables.
Curve and $CRV → StableSwap, now renamed Curve - which is ironic if you look at the math, launched in January 2020 with immediate success until, during the August of the same year, an anonymous developer took the open code sitting in GitHub and hijacked it to launch his/ her own version of $CRV, Curve’s ERC-20 governance token. The token gained so much traction the Curve team was forced to officially adopt it.
$CRV would become the central tool of the Curve DAO, and the most precious asset to assert control over profits allocation (and therefore liquidity steering) within the protocol. $CRV was to Curve and to the wider DeFi ecosystem what Helen was to Troy and to the Greeks. Then, a month later, the community rallied by $CRV voted on a proposal to self-allocate 50% all generated trading fees. In order to benefit from this 50%, $CRV token holders would have had to stake their tokens in a voting escrow - obtaining $veCRV in exchange. Governance tokens could (and still can) be staked for a period up to 4 years, incentivising engagement. Only in the case of a full 4-year staking a $CRV token could be swapped 1-for-1 with $veCRV. Not much has changed since then.
The perks of being $veCRV → Staked governance tokens do not only grant a slice of the protocol fee pie to holders, they also boost LP returns by a factor of up to 2.5x. Such boost comes in the form of newly minted $CRV - that come to market in a fixed weekly amount, with the allocation calculated through a system of gauges. The parameters governing the gauges are obviously controlled by the DAO, or in other words by $veCRV holders. The circle closes: more liquidity provided into the vaults allows the allocation of newly minted $CRV that can be staked as $veCRV and used to gain further fees and steer the distribution of future $CRV.
Location, location, location → There are currently c. USD 20b (billions) deposited in Curve’s pools, with a daily trading for c. USD 525m, and USD 57m cumulative total fees earned by the protocol. Voting power can exercise control over all this: through weekly set parameters $veCRV owners can 1) steer allocation of newly minted $CRV via gauge boost - token trades at $3.15 currently, 2) put their hands on the 50% protocol tax, 3) and also push new liquidity providers towards certain pools via curated incentives. But what are they interested in the most, is it the returns or the liquidity? It is debatable. Profits are nice but USD 20b of liquidity could change the fate of any project. There are many things you can do with a rich parent, or friend.
So who are the largest holders of $veCRV? There are currently c. 376m tokens and c. 65k holders, but with huge concentration. Another protocol, Convex, currently owns almost half of the total $veCRV supply, as well as a disproportionate power of influencing how future $CRV is allocated and, as a consequence, where liquidity might end.
What is Convex? We need to climb another floor in the crypto-composable pyramid.
Layer 2: Convex, the Ungrateful Son
You don’t need to spend a long time navigating Convex docs to understand there’s no hiding: Convex wants to aggregate liquidity and amass enough voting power to hijack Curve. Curve created a very effective (and profitable) liquidity aggregation machine opting to give the keys to the strongest in the jungle, and Convex organised itself to gather enough brute force to do exactly that. You can think of Convex as an aggressive version of Groupon for DeFi. How does the protocol incentive this? As usual, it’s complicated. If you pay attention, however, some of the mechanisms should sound familiar: Convex is doing exactly what Curve did before.
(1) Users deposit liquidity on Curve. Users deposit liquidity on Curve pools where they accrue LP fees, receiving a fungible LP token in exchange.
(2) Convex incentivises a liquidity transfer. Those LP tokens can be staked in a Convex pool, receiving both Curve’s $CRV incentive and, on top, some $CVX tokens.
(3) Convex incentivises a voting t. Users have the option to convert (irreversibly!) the $CRV owned into a Convex version, $cvxCRV, and to stake the token on the Convex platform. They receive additional yield in exchange.
(4) Convex uses the aggregated voting power. Convex, coordinated by their $CVX token, uses its amassed $CRV/ $veCRV voting power to steer $CRV allocation in its favour, boosting returns further.
Recap → Let me give the man on the street a recap of what is going on. Curve is a great AMM, and because it works so well it attracts a lot of liquidity. The decisions that can steer the control of the platform (and therefore of the attracted liquidity and generated profits) are delegated to a governance token, $CRV. Given the value of $CRV, a protocol has organised to gather as much of it as possible, and it has done it via a similar mechanism now based on a token called $CVX, i.e. a $CRV squared. Did it work? Judge for yourself. TVL in Convex is now c. USD 15b, having briefly passed 20b earlier this month - cred. DeFi Llama. Curve's 3pool ($DAI + $USDC + $USDT) has c. USD 5.4b of value deposited in it, and of this c. USD 1.5b is sitting with Convex. The machine, however, is expensive to maintain. Prior to the most recent downturn the price behaviour of the underlying $CRV token had helped keeping the group in check, but things have changed - $CRV went from $7/token back to $3/token in a couple of weeks. DeFi, however, doesn’t take no for an answer: if Convex has worked so well in hijacking Curve by applying a meta-version of Curve’s same tokenomics, why stop there? Why not simply moving up by one layer?
Layer 2.5: Yearn, the Challenger
Yearn Finance was born to provide users with a simplified access to yield generating strategies within DeFi. A one-stop-shop to aggregate and systematise depositing→claiming→staking→unstaking with great benefits in terms of peace of mind and reduced gas. Yearn’s key product are yVaults, something close to savings accounts for crypto assets. yVaults systematise execution across a wide spectrum of yield generating strategies, and all this in exchange for a mere Mayfair-ish 2/20 fee.
Yearn’s relationship with Curve is a crucial element of the yVault products. ETH is the largest (and simplest) vault on Yearn with more than USD 500m sitting in it: the strategy supplies $wETH to Curve to earn fees and $CRV - the earned tokens are harvested, sold for more $wETH which is deposited into the strategy. The second largest vault, with c. USD 500m of TVL, is Curve stETH: $steCRV (the LP token received when $ETH or $stETH is provided into the Curve pool) is supplied to Convex to earn $CRV and $CVX - and any other available token. Those earned tokens are harvested, sold for more $steCRV and deposited back into the strategy.
Yearn, in other words, operates a hybrid strategy: on the one hand it connects directly into Curve, and on the other it utilises Convex’ boosting engine to optimise returns - but at the cost of further reinforcing the position of one of its most direct competitors.
Layer 3: [REDACTED], Cartel-as-a-Service
There is no reason to stop here either, and we won’t: $CVX has amassed enough liquidity to control $CRV and there is nothing preventing another protocol to amass enough liquidity to control $CVX. Enters [REDACTED] Cartel. Rather than usurping daddy’s place by applying its own receipt the way Convex did with Curve, the Cartel decided to get counselling from the ascending boss in town: Olympus.
Operation Swallowtail → [REDACTED]’s bootstrapping event (Operation Swallowtail) happened on the 15th of December and ran for less than 48 hours, during which the protocol managed to raise USD 73m through the bonding of $OHM - let’s call it a tax to its predecessor, $CRV, and $CVX. It did so by issuing at discount $BTRFLY, its rebasing token, following the Olympus bonding model. The inflation season took off, supporting the aggregation of a treasury chest that peaked at USD 150m in mid-January before crashing by 50% with the rest of the market. [REDACTED] stated mission? Hijack Convex and control Curve.
Swallowtail was only the first of a 4-phase rollout strategy:
Swallowtail: aggregate initial liquidity to put the protocol on the map (done)
Red Admiral: repurchase $BTRFLY LP position on Sushiswap
Holly Blue: transition LP positions into Uni v3, and expand beyond Curve
Orange Tip: migrate to an Olympus v2 model and move fully on-chain
A brief story of the Olympus Family → Ultimately, Olympus sharpened the tools required to attract immense amounts of liquidity while reducing downside volatility by internalising the market making of its own token. There was a flaw in the system, however: with a treasury chest mainly made of stablecoins (or any coin for what matters) it would have been very challenging to generate enough returns to feed a beast that craves 5-digit APYs. The protocol tried to justify its incredible success, and the 20-40x market valuation vs. the value of its treasury, with a not-better-defined currency premium that should have reflected $OHM’s superiority as currency of the future. It might have been more appropriate to call it greed premium. Still, while making few people very rich (and then very poor) Olympus had pioneered a new way for protocols to gather liquidity more efficiently while the bull was roaring - it called its liquidity-as-a-service engine Olympus Pro. Olympus’ success inspired many others, and among the innumerable forks some understood that by cornering a relatively small and profitable market it would have been easier for the teams to squeeze good returns from the treasury assets. Klima did this through a massive purchase of tokenised carbon credits, trying to benefit at the same time from the sustainability narrative that is so pervasive in today’s financial markets. The jury is still out on this, although the project suffered the fate of Olympus during the last months.
[REDACTED] did something different: it openly asserted in its manifesto the intention to champion a portable strategy able to weaponise liquidity to attack any project deemed potentially dominant within DeFi. The mob, rallied by $BTRFLY, could serve any purpose, moving from target to target as a useful and greedy agent of chaos.
Conclusions: Tilting The Pyramid
Nothing, in what has been discussed so far, has to be considered investment advice. But if there’s a personal advice I can give to all investors, through DR, is to follow value flows rather than nominal returns. While the analysis of returns can be polluted by the tectonic movements of macro liquidity - which are very difficult to predict, an analysis of value should focus entirely on the tangible gains provided to the stakeholders thanks to some sort of competitive advantage. In the case of Curve, value is generated through the fees charged to Curve’s AMM users, and its sustainability is predicated on the superiority of Curve vs. its competitors. Value is then redistributed aggressively across the various layers of the pyramid, and the fruit of that distribution is used to steer the allocation further. Yes, I know, the market interact with this system injecting additional liquidity via the purchase of the utility tokens, but as investors we are trying to stay away from analysing liquidity flows. It’s all about value. Liquidity, using Porter’s words - h/t, should remain a means to an end, and not an end in itself.
The whole pyramid depends on the value generated by the base, and with drying value flows buoyant investors can delay its unwinding only temporarily. Curve might have found today its value killer in Uniswap v3. Those are testing times: if the $CRV→$CVX→$BTRFLY→? construct is more than a mere speculative toolkit, it will succeed in rallying the troops of builders and in putting back Curve ahead of its competitors. There is a lot of money at stake. Without this, the Curve Wars might quickly transform into the bloodiest intestine guerrilla in the history of DeFi.