Terra blockchain and $LUNA Crash in Detail
Table of Contents
The collapse of Terra ($LUNA) shook the crypto industry to the core, wiping out $600 billion in wealth. How and why did it happen – and can we expect anything from LUNA 2.0?
TLDR: the dramatic rise and fall of $LUNA
- $LUNA is (or, rather, was) the official cryptocurrency of Terra, an ecosystem that gained popularity for its algorithmic stablecoin UST and the Anchor lending protocol.
- On January 1, 2021 $LUNA traded at $0.65. On April 1, 2022, it was $105.
- In the same period, Terra’s TVL went from $50M to $29 billion.
- Starting in November 2021, analysts raised concerns about a potential attack on Terra and $UST that could lead to a crash, but founder Do Kwon mostly disregarded them.
- In January 2022, Luna Foundation Guard (LFG) was created to safeguard the UST peg and support the development of the ecosystem. Soon, it accumulated $3.5 billion in BTC reserves.
- On May 7 2022, $UST depegged following massive withdrawals of liquidity and dumps on exchanges. $LUNA lost 99.9999% of its value.
- It’s possible that the crash was the result of a well-planned operation by a group of entities that stood to make up to $800 million in profits.
- The crypto market lost $200 billion in a single day, with Bitcoin crashing below $26k.
- Under Do Kwon’s direction, the Terra blockchain was forked to create Terra 2.0. The old Terra was renamed Terra Classic ($LUNC) (though in this article we’ll stick to calling it Terra when describing pre-crash events).
About the Terra blockchain
Terra is a smart contract platform, just like Ethereum, Avalanche, Solana, etc. You can build all sorts of dApps on it, from lending protocols to NFT marketplaces. It has a reasonably fast finality of 6 seconds (the time needed for an operation to be fully confirmed), and the transaction fee didn’t exceed $0.2 even when the price of $LUNA reached an all-time high. The maximum processing capacity for Terra is around 10,000 TPS. While Terra can’t compete with Aptos, Fantom, or Avalanche in terms of time to finality, and its fees are far higher than on Solana, it’s still a fairly fast, scalable, and affordable blockchain.
Terra is part of the Cosmos ecosystem, which we discussed in our article on multichain networks. It’s built using Cosmos SDK, a modular framework that features ‘building blocks’ for blockchains, such as staking, governance, and more. Other projects that use Cosmos SDK are the Cosmos Hub ($ATOM), Kava ($KAVA), Cronos ($CRO), and Osmosis ($OSMO).
Terra also uses the same Tendermint consensus engine as Cosmos and has integrated the Inter-Blockchain Communication protocol (IBC), making it possible to bridge tokens to the other 17 IBC-enabled chains, including Cosmos and Osmosis.
The company behind the Terra blockchain is called Terraform Labs, initially formed in Seoul but now based in Singapore.
The early days of Terra
The first successful project to use Terra was CHAI, a popular Korean payments app launched in June 2019. Users could link an account from one of 15 large Korean banks and pay both online and offline using a CHAI card. While a user would spend Korean won (KRW) from their bank account, the transactions were actually processed on the blockchain using TerraKRW (KRT), a won-pegged stablecoin that preceded UST.
Between November 2019 and December 2020, Terra’s DeFi TVL (total value locked) fluctuated around $2.5 million, then started climbing fast after the launch of Mirror Protocol ($MIR). The TVL hit $50M on January 1, 2021, $200M on March 2, and $1 billion on April 1, 2021. It continued to grow steadily, reaching highs of $29.65 billion just before the collapse.
TerraUSD ($UST): the algorithmic stablecoin on Terra
What is an algorithmic stablecoin?
An algorithmic stablecoin is a token that uses an algorithm to maintain a peg to another fiat or digital currency (usually USD). While regular stablecoins like USDT or USDC are collateralized by actual dollars and other assets and run by a centralized issuer, algorithmic stablecoins normally have no collateral and have no central authority overseeing them. Their supply expands when they are above peg and contracts below peg, often by offering arbitrage opportunities to the users.
The best-known algorithmic stablecoin, apart from $UST, is probably $DAI, created by MakerDAO. Others include Magic Internet Money ($MIM), $FRAX, and Waves’ Neutrino USD ($USDN). We’ve already mentioned Terra’s own TerraKRW (KRT); in addition, Terra has an algorithmic stablecoin pegged to the Mongolian tugrik, $MNT.
UST emission mechanism
To mint one UST, a user needed to burn $1 worth of LUNA. Therefore, if the demand for UST grew, more LUNA was burned, and the supply of LUNA decreased. And vice versa: users could redeem 1 UST to get back $1 worth of LUNA, thus reducing the overall supply of UST.
Whenever UST went above its $1 peg, arbitrage traders would mint more UST at $1 per token using the protocol, then sell it higher on exchanges like Binance and make a small profit. As they sold UST, the price would go back down to $1.
Conversely, if UST dropped below peg, arbitrageurs started buying UST on exchanges to redeem each token for $1 in LUNA – once again, for a small profit. At the same time, they generated sell pressure in the market, driving UST back up to $1.
UST in numbers
UST was launched in September 2020 in collaboration with the Bittrex Exchange. By the end of 2020, it reached a market cap of $180 million. Just before the great crash on May 7, 2022, UST had a $18.7 billion market cap - over double that of DAI ($8.46B). But by June, UST’s market cap had fallen 60 times over to $300 million.
The advantages of UST
- UST’s minting mechanism makes it very scalable, as you can make as many UST as you need simply by burning $LUNA. Scalability is an important feature, considering that the demand for stablecoins from DeFi apps can grow very fast.
- Unlike DAI or MIM, which exist on only one blockchain (Ethereum and Avalanche, respectively), UST is an interchain asset. Thanks to a special bridging protocol called Dropship, UST can be transferred between Terra, Ethereum, and Solana.
- Finally, UST yielded a stable passive income. Before the crash, holders could deposit UST on Anchor Protocol to earn 19.5% a year. DAI and other stablecoins can also be deposited this way, but the interest rates constantly fluctuate.
Let’s take a look at a few of the dApps that made up the Terra ecosystem.
Anchor is a savings and lending protocol where users can deposit UST to earn a stable 19.5% APY. Borrowers need to supply collateral in $LUNA or bridged ETH, SOL, and ATOM. Borrowers also get rewards in the native $ANC tokens.
Launched in March 2021, Anchor reached a TVL of $4 billion by September, becoming the third-biggest lending protocol after Aave and Compound. At one point 56% of the UST market cap was deposited on Anchor.
At the peak in May 2022, Anchor’s TVL reached $17 billion. By the end of May, there was just $25 million left – a dramatic 680x fall in 3 weeks. https://defillama.com/protocol/anchor
Lido is a liquid staking platform that allows users to stake $LUNA and enjoy a 8% APR while still enjoying the utility of their staked coins through stLUNA – special tokens that can be used elsewhere in DeFi.
Other notable Terra dApps
Astroport ($ASTRO) and Terraswap: decentralized exchanges for swapping $LUNA, $UST, $ANC, $MIR etc.
Stader Protocol: another liquid staking dApp
Mirror Protocol: a synthetic asset protocol offering tokenized versions of real-life assets like APPL shares.
Who is Do Kwon, the founder of Terra?
Do Kwon, or Kwon Do-hyung, is a South Korean crypto entrepreneur hailing from Seoul. Born in 1991, Do Kwon got a bachelor’s degree in computer science from Stanford University in 2015. He then returned to South Korea to found his first company, Anyfi, which offered decentralized networking services.
In 2019, Do Kwon was featured on the Forbes list of Asia’s 30 Under 30 in Finance & Venture Capital. He has a daughter, born in April 2022, whom he named Luna after his ‘greatest invention’ – quite ironic, given that $LUNA collapsed just a month later.
Do Kwon is known for his confident and even arrogant attitude. He has repeatedly mocked other Twitter users for being ‘poor’.
In a now-infamous interview given a few days before the $LUNA disaster, Do Kwon also said (laughing) that 95% of tokens would die and that he found it entertaining:
While many in the industry found Do Kwon’s behavior , he was supported by an army of ‘Lunatics’, or fans of $LUNA. Remarkably, many of them remained loyal and enthusiastic even after the crash.
In May 2022, CoinDesk uncovered that in 2020, Do Kwon co-founded another algorithmic stablecoin project called Basis Cash on Ethereum. He used the pseudonym ‘Rick Sanchez’ and apparently used Basis Cash as a test for UST. According to CoinDesk, many of the developers on the team worked for Terraform Labs.
Experts predict $LUNA’s downfall as the price surges by 150%
$LUNA’s meteoric rise started in November 2021 – and as the rest of the market crashed in December following Bitcoin, $LUNA seemed immune, going up 150% from $34 to a new all-time high of $85. In three months, the market cap rose from $14 billion to $36 billion.
Not everyone was impressed, however. Popular Crypto Twitter personality Freddie Raynolds even outlined a possible devastating attack on Terra that would require $1 billion in capital. It mirrored the famous attack that George Soros mounted on the Bank of England and the British pound in the 1990s, earning over $1 billion.
Generally, the attack idea worked this way:
1) Opening a leveraged long position on $LUNA.
2) Buying lots of $LUNA in the spot market and swapping it for $UST via the Terra protocol to accumulate around 500 million UST.
3) All that $LUNA buying would cause the price to rise, so the attacker could now close the long for a good profit.
4) Taking out a big loan ($100 million) in UST through the Anchor protocol.
5) Opening a huge short position on $LUNA.
6) Converting the 100M borrowed UST, plus the 500M UST obtained in Step 2 (600M UST in all) for $LUNA, once again via the protocol. This would have to be done in $100k batches and would take time.
7) Gradually dumping all the resulting $LUNA in the spot market. This would cause the price to go down, so the attacker could now close the big short position for a nice profit.
8) Pay back the UST loan.
9) Repeat steps 1-8 a few more times. According to Freddie Raynolds’ calculations, each iteration could earn the attacker over $50 million in net profit. But at the same time, new $LUNA would keep being minted and enter the market. As the supply rose fast, the whole system would become unbalanced, and the price of $LUNA could collapse.
In response, Do Kwon himself called Freddie Raynold ‘stupid’. He also called upon any billionaires reading the debate to ‘go ahead’ and try to execute the attack. As always, his loyal fans sided with him.
While this is not exactly what happened in the end, it’s important to note that experts were pointing out vulnerabilities in the Terra tokenomics. ‘Lunatics’ believed that the price could keep going up even as the supply rose, simply because the public wanted more and more $LUNA – but they forgot that there is no such thing as ‘up only’ in crypto, same as any other market. (And yes, it seems like some billionaires did indeed take Do Kwon up on his challenge to ‘go ahead’ – read on.)
Luna Foundation Guard is born
As anxiety in the crypto market grew stronger, the management of Terraform Labs took measures to protect $LUNA, especially the $UST peg. On January 19, Terra announced that a new non-profit organization called Luna Foundation Guard (LFG, a pun on the popular crypto slogan ‘LFG’ or ‘Let’s F****g Go’) would be created in Singapore.
The task of Luna Foundation Guard was to accumulate reserves in BTC and other cryptocurrencies to support the UST peg if needed, as well as to support the growth of the Terra ecosystem.
In February, LFT sold $1 billion worth of $LUNA to investors like Three Arrows Capital (3AC) and Jump Crypto in order to buy BTC with the proceeds. The BTC reserve was to be used only as a last-resort measure in case of a serious UST depegging.
By mid-April, LFG had accumulated $2.4 billion in Bitcoin and Avalanche ($AVAX). Unfortunately, as we’ll see soon, even this huge reserve was not enough to save UST.
‘Your size is not size’: Do Kwon rejects another warning and accepts a $1M bet
At the end of January 2022, $LUNA corrected sharply to $50, but then rallied again to a new record of $113. This time, it was well-respected Twitter trader and investor Algod (@AlgodTrading) who predicted a collapse – he even called $LUNA a Ponzi (a scam).
Algod’s reasoning was based on the fact that you could always swap 1 UST for $1 worth of $LUNA through the protocol (just like in the attack described by Freddie Raynolds). As we described above, this mechanism was supposed to help UST return to is the peg, but imagine this scenario:
1) $UST loses peg.
2) People start buying $UST (at $0.99, for example) and converting it to LUNA via the protocol (at 1 UST = $1 in LUNA), making a $0.01 profit off each stablecoin.
3) Then they immediately dump $LUNA in the open market for USDT.
4) The supply of $LUNA in the market grows very fast and starts to exceed demand.
5) Panic starts; everyone sells; the price of $LUNA collapses.
This is also how Ponzi schemes colla[se: once the supply exceeds demand (not enough new users join the Ponzi), it crashes.
On March 9, 2022, when $LUNA traded at $99, Algod tweeted that if the coin were to reach a new ATH, he would short it ‘with size’. This meant he would open a large short position, so strong was his conviction that the price would go down.
Do Kwon’s response became an instant classic: ‘Your size is not size’. This implied that whatever trading capital Algod had was insignificant – that he was basically ‘poor.’
As the debate escalated, on March 13 Algod proposed a $1 million bet that $LUNA would trade lower a year later – that is, on March 13, 2023. Do Kwon happily accepted (once again implying that Algod hardly had the required $1M) and even suggested a bet where UST would still be at peg in March 2023.
<blockquote class="twitter-tweet"><p lang="en" dir="ltr">Cool, I’m in</p>— Do Kwon 🌕 (@stablekwon) <a href="https://twitter.com/stablekwon/status/1503126656545792002?ref_src=twsrc%5Etfw">March 13, 2022</a></blockquote><script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
Algod may or may not have had ‘the size’, but someone else definitely did: GCR, or @GiganticRebirth, another famous retired crypto trader. He suggested a $10 million bet that $LUNA’s price would go down by March 2023. Once again, both GCR and Do Kwon deposited $10 million each in Cobie’s escrow walle
May 7-13: the great UST depeg
Ultimately, what keeps a stablecoin pegged is not the reserves backing it or an algorithm, but trust. If people start dumping it in large amounts and FUD (fear, uncertainty, and doubt) begins to circulate, trust can be lost quickly – then the price goes down.
The peg can be restored through urgent measures that restore trust, such as a highly publicized deployment of capital, but if things go too far and buyers don’t want to step in, the stablecoin’s price will keep decreasing.
The selling started on May 7 with a single account swapping $85 million in UST for USDC on Curve and $108 million on Binance – almost $200 million in total. At the same time, users began withdrawing large amounts of UST from the Anchor protocol: its TVL went down from $14 billion to $11 billion in two days.
The dumping fed the FUD, which, in turn, generated more dumping as Anchor lenders rushed to get their UST out and convert them into safer USDC or USDT. Ironically, Terraform Labs contributed to the FUD by withdrawing $150 million in liquidity from Anchor – though they deposited it back later. As a result of the panic selling, $UST went below $0.986 on May 8.
Some Twitter users also pointed out that large short trades on $LUNA were opened about the same time as the big dumps, and that Twitter was suddenly filled with FUD-inducing posts, suggesting a staged attack.
Luna Foundation Guard announced it would lend $750m worth of BTC to market makers to protect the peg. Meanwhile, Do Kwon remained remarkably calm through all this and even mocked those who panicked about the depegging.
On May 9, UST closed around $0.90 as the TVL on Anchor lost another $2 billion in TVL. Do Kwon tweeted that Terraform Labs was ‘deploying more capital’, but if he hoped this would repair the trust in UST and induce buys, it didn’t quite work. On May 10, UST did briefly go up from $0.8 to $0.93 – only to drop to $0.8 again.
On May 11, UST collapsed to $0.30 – then went up over 100% in a few hours to hit $0.8, though a lot of that gain was probably due to short positions being covered or liquidated. For the next three days, the $UST trajectory was relentlessly downwards until it hit $0.17.
Meanwhile, LUNA was going down with terrifying speed: from $73 to $1.1 in five days. This has hardly ever happened to a top-10 coin. Remarkably, each new drop was followed by a noticeable bounce as people seemingly thought that ‘this can’t go much lower’. It was knife-catching in its purest form: people trying to catch the bottom just to get hurt (or, as they say in crypto, rekt) in the process.
May 12-13: Terra blockchain is halted and $LUNA loses 99.5% in a day
May 12 will go down in crypto history books as the day when a multibillion-dollar asset lost over 99% in a single day. As $LUNA tumbled from $1.10 to $0.005, even the staunchest Lunatics had to change their slogans from ‘$LUNA back to $100’ to '$LUNA back to $1’.
On that day, Terra validators halted the blockchain twice to apply emergency security patches, with no on-chain transactions with UST or $LUNA possible. The second time the network was out for 9 hours.
On the 13th, several major exchanges suspended trading in $LUNA or even delisted it altogether, including Binance (though it restarted trading later), Huobi, FTX, OKX, and Crypto.com. By that time, $LUNA was down to $0.00003, down another 99% in one day.
Luna Foundation Guard deployed all of its reserves to try and save $UST. As revealed by the analytics firm Elliptic, LFG moved virtually all of its $3.5 billion in BTC to Gemini and Binance – presumably to buy UST with it. It was too late: $UST traded at $0.18 by the end of May 13.
Was there a coordinated attack on $UST and $LUNA?
On May 10, crypto analyst @OnChainWizard published a Twitter thread explaining how (in his opinion) an attacker caused $UST to crash and made $800 million. Here’s what it entailed:
1) The attacker built a $4+ billion short position in Bitcoin at around $42k per BTC (that is, they would make a profit if the price of Bitcoin went down from there). It happened just about the same time as LFG was buying BTC to build up its reserves.
2) The attacker also accumulated $1 billion in UST using OTC (over-the-counter) services, so that this position didn’t appear on any exchange order books.
3) They waited until LFG took out part of the liquidity from the UST pools on Curve to move it to a new pool. This was the best moment for the attacker to start selling UST, as with less liquidity the effect on the price was greater.
4) As the perpetrator sold $350M worth of UST, the stablecoin depegged. LFG started spending BTC out of its reserve to buy UST.
5) The attacker began selling the remaining $650M in UST on Binance, causing more pressure on the price, followed by mass withdrawals from Anchor, more depegging, etc.
If it’s true, it was a brilliant plan, as the perpetrator found a creative way to make the price of Bitcoin drop 25%. And indeed, it was similar to what George Soros had done in 1992, when he opened massive shorts on GBP and convinced other financial institutions to do the same. The Bank of England had to start buying GBP using its foreign currency reserves, but it didn’t have enough. The pound dropped 15% as a result, while Soros made over $1 billion in profit on his shorts.
Soros’ attack is well documented, but did the attack on $UST and $LUNA really take place? Experts like the CTO of USDT Paolo Ardoino and IntoTheBlock’s Head of Research Lucas Outumuro agree that the events looked very suspicious. On the other hand, David Z. Morris from CoinDesk’s Money Reimagined stressed that it was not an ‘attack’ – just someone seeing a profitable opportunity and taking it, as always happens in financial markets.
Moreover, blockchain analytics firm Nansen identified seven wallets that played a role in the $UST depegging and suggested that there was no single attacker - rather, a group of entities with a lot of resources who may or may not have worked together.
The aftermath: $200 billion wiped out
May 12, 2022 saw one of the biggest drops in the short 10-year history of the blockchain industry. The total crypto market cap dropped by $200 billion in less than 24 hours.
Of course, this can’t compare to the disaster of May 20, 2021, when the market lost almost $1 trillion in a day. And the great Covid crash of March 12, 2020, while not as impressive in absolute terms ($93 billion lost), was much more serious in percentage terms, as the total crypto market cap was far lower in those days.
Terra 2.0: $LUNA reborn?
The Terra ecosystem itself bled out $27 billion in TVL, most of it from Anchor.
Worst of all, behind these numbers were real people – not just institutions or funds. Many users lost all their savings, having put everything into $LUNA; others went from crypto millionaires to broke overnight. There were even (unconfirmed) reports of suicides committed over LUNA.
Do Kwon didn’t appear apologetic or even particularly shaken, though. Instead, he proposed forking Terra to create a new blockchain, Terra 2.0. The proposal would be put to a vote of the Terra validators.
The ‘old’ chain would keep working and get renamed to Terra Classic ($LUNC and $USTC). The new chain would get the ticker $LUNA – but it wouldn’t have a stablecoin. The holders of LUNC and USTC would be airdropped new $LUNA.
On May 25, 65% of the participating validators voted for the forking proposal. On the 28th, the new network was launched, and on the 30th, the first stage of the airdrop was completed on Binance. The airdrop’s terms depend on how much $LUNA or $UST one held, if it was before or after the attack, and the amount.
Those who had less than 10k LUNA before May 7th were in the best position, as 30% of their airdrop tokens were unlocked immediately, with the rest to become available over the next 2 years. By contrast, ‘whales’ with over 1M $LUNA before the attack will get the first batch of tokens in mid-2023, and the rest will be unlocked over a period of 4 years.
New $LUNA started trading on Binance on May 31. It initially pumped $20, then dropped to $4.
As of the time of writing, $LUNA 2.0 was worth $2 and had a market cap of $246 million - more than assets like Compound, Oasis Network, or Moonbeam – an impressive achievement for a project that was as good as dead a few weeks earlier.
The future of Terra 2.0 remains uncertain, with some experts and traders calling it a Ponzi. Moreover, South Korean authorities have launched an investigation into Terraform Labs after some investors had filed complaints. Do Kwon is already facing a tax investigation, with a potential fine of almost $80 million.
How will the $LUNA crash shape the future of blockchain?
While these two earlier crashes were followed by V-shaped recoveries, the $LUNA crash wasn’t. That’s what makes it particularly worrying. To users outside of the blockchain space, this sends a worrying signal: crypto remains fragile. If a network that just a couple of months ago seemed like the future of the industry can lose 99% of its TVL in a week, can one trust blockchains at all?
For us at Pontem, as we build tools and dApps for Aptos, the latest crash has another important implication: it puts the importance of security and reliability into sharp relief. In the post-Terra landscape, users will seek the safest blockchains with bona fide hardworking founders (rather than flashy figures like Do Kwon), and Aptos will definitely have a competitive advantage in this context.
Aptos has an ambitious vision to introduce the first billion users to blockchain, and it has the technology to match these claims. Can Aptos become the breakout L1 star of the next bull run? It’s definitely possible – and Pontem is getting ready with Aptos’s first suite of foundational dApps, including the browser code editor Move Playground, Liquidswap DEX, and a wallet. Follow us on Twitter and Telegram to stay updated!