From on-chain data to Web3: ETH is leaving the exchange

It is undeniable that the crypto market continues to experience incredible growth and innovation.

In recent months, the market has been turbulent and events have occurred frequently. As the largest cryptocurrency continues to encounter obstacles to its progress, the global digital asset market is still occupied by a bear market. However, it is undeniable that the crypto market is still experiencing incredible growth and innovation.

From the merger of Ethereum to the competition among a number of public chains; from the expansion of the DeFi bubble to the current industry of over $100 billion; and Bitcoin reaching a peak price of $69,000, NFT sales exceeded $22 billion. These historical records tell us that the crypto market has attracted the public's attention in an unprecedented way.

Although the crypto market has been in a deep bear market for a long time, judging from the on-chain data, the web3 ecosystem has gradually begun to occupy market share, and the digital asset industry is still on the road of development at its own pace.

Starting with Bitcoin, as of July 2022, 1 billion Bitcoin wallets have been created. Of these, about 42 million have non-zero balances. It took about 4 years for Bitcoin to see its first 1 million non-zero balance wallets. The last million took six months.

This number could easily be considered “dusty wallets”. Wallets holding small amounts of Bitcoin from random transactions in 2011 could add up. Wallets holding more than 1 Bitcoin set a new ATH, even though it cost more than 100x to acquire one Bitcoin about 5 years ago.

Miners’ exchange trading volume is decreasing

Some of the larger wallets have been pulling funds since the price rally in June 2017. Wallets holding more than 100 Bitcoin peaked at 18,000 wallets in June 2017, while wallets holding more than 10 Bitcoin peaked at 150,000 around September 2019, and they have more or less remained flat since then.

Profitable Supply Percent takes the Bitcoin price when the coin was last moved and the current price to see if the coin is profitable. At the ATH, all UTXOs were profitable. Currently, about 54% of UTXOs are in the green. Historically, the bottom of this metric has been around 40%.

Age groups look at the percentage of network assets that have moved over different periods of time. It is assumed that moving some older coins will result in lower prices as they are acquired at cheaper prices. About 40% of coins have not moved in 3 years or more (including Satoshi’s coins).

One way to measure Bitcoin usage is to count the number of transactions worth more than $1 million. While this depends on the price of Bitcoin, a floor is being established here. About 3,000 transactions worth more than $1 million occur on Bitcoin every day.

Nearly 10% of the Bitcoin supply currently relies on exchanges (about 2.26 million coins) At its peak, this number was close to 3 million Bitcoins. In the past year, about 1 million have more or less flowed to prime brokers.

Miners are a key part of the supply chain for new coins entering the market. Wallets identified as being held by them hold just over 1.9 million Bitcoins (9%) of the supply. They appear to be strong holders — likely due to hedging through off-chain instruments such as options.

At its peak, miners held close to $168 billion worth of Bitcoin. That number is now around $33 billion. As mentioned earlier, it would be unfair to suggest that they lost all of these gains. A large portion of it was likely hedged.

Correlation is not causation and so on - but it is worth observing that last year, miner-owned wallets had the highest number of transactions to exchanges. Miner-to-exchange transactions have been declining since then.

The coinbase premium index looks at the difference between the coinbase pro price (for BTC, in USD) and the price on binance (for BTC, in USD), and is a measure of US investor interest in buying BTC. It's been a bit flat lately, but the discount in May 2022 is worth noting.

Thanks to Do Kwon, this premium gap has dropped from a high of about $160 when Elon was talking about Bitcoin in January 2021 to a low of about $140. Currently, it remains volatile like the rest of the market.

On average, about $350 billion is transferred via Bitcoin each week. To put the scale into perspective, that’s about 5 times the annual inbound remittances to India. This number varies based on the price of BTC, but is still above the $200 billion range in 2017.

Currently, there are about 50 times more tokens on wBTC than the capacity of the Lightning Network.

Of that, about 250,000 Bitcoins, about 60,000 Bitcoins were in MakerDAO’s heyday. Now that number is closer to 40,000. DeFi yields are naturally the key driver of this hybrid between Ethereum and Bitcoin.

Players like Alameda, Grapefruit and 3 Arrows (uh) - have been historically critical to its adoption. Part of the reason June was on fire was that the market corrected itself in May and stress-tested everyone's balance sheets.

In terms of market share, there is nothing comparable to WBTC. The second largest player owns 5% of the BTC-on-eth supply. Given their distribution and reach, I think a similar product from Binance could have huge traction.

That being said, it does look like interest in wrapped variants of Bitcoin is declining. At its peak, we saw nearly 55k wBTC being minted each month. In August, that number dropped by about 80%.

Next, we look at what happened with Ethereum. If we take deployed smart contracts as a measure of developer interest — we can say that activity has trended towards 2019 levels. However, this does not account for the fact that other chains may be attracting developer attention.

Santiment has a developer activity index that takes a few GitHub repositories and their activity to give an indicative measure. Here’s how Solana, Matic, Avalanche are catching up to Ethereum. Note: The data is very rudimentary and doesn’t mean much/has gaps on its own.

Developers are developing in multiple places, but what happened to Ethereum? Let’s start with the basics. Active wallets are a good place to start. What I find interesting is that despite a year of bear market - Ethereum wallets have not significantly fallen back to new lows.

A 2x increase in daily active addresses isn’t much — but it is impressive. Since the last ATH, non-zero wallets on Ethereum have increased from ~8 million to 86 million. Granted, people use multiple wallets — it’s a decent metric for network activity.

People will always argue that these are dust wallets - so we explored every wallet with more than 1 ETH and 10 ETH. Both metrics are at all-time highs. There are about 1.58 million wallets with more than 1 ETH today, compared to 882,000 in January 2018.

The measure of whether people trust the network is the percentage of supply that is in smart contracts. In ETH, 27% of its supply is in smart contracts. The bear market has not affected this number either.

Four of the top ten wallets holding Ethereum are smart contracts or variants of them — it remains one of the few assets that CeFi is losing smart contract share to. So it’s safe to say — people are using Ethereum — and increasingly using its smart contracts.

This chart breaks down the types of transactions on the network. In 2018, when it hit its ATH — about 60% of transactions on the network were just sending ETH from A to B. That number is now down to 30%.

• NFTs account for 15%

• Stablecoins account for 9%

• DeFi 10%

Use cases like NFTs, DeFi, and stablecoin transfers actually far outweigh the number of transactions associated with ERC-20 tokens in a single day.

The majority of balances are held by Binance and Coinbase, which together control about 44% of ETH balances held on exchanges. The second chart here shows how Coinbase has stagnated in market share while Binance has caught up.

ETH is leaving exchanges

The main contributor to this trend is DeFi. Since January 2021, about 14 million ETH has been used to stake ETH2.

There has been a lot of discussion about this, but what’s interesting is that Lido has far outpaced established exchange alternatives when it comes to total value staked, with around 4 million ETH staked through it versus 3 million ETH staked through binance+.

The reason why exchange deposits are used for staking comes from this chart from Nansen - about 40% of the staked ETH can be traced back to centralized entities.

As of this writing, there are over 442,000 validators on the network. Therefore, part of the appeal for ETH holders is that they can see a theoretical yield of around 4.5% on ETH staking in the future.

Nearly 30% of ETH supply has been dormant for 2 years or more. 1+ year hodlwave is ignored here because ETH2 staking affects this metric. This is a large portion of the supply that is willing to withstand multiple cycles.

Part of the reason for this willingness to hold is that much of the network is in relative profit. At $1,290 — close to 46% of ETH holders are still in the green. During the March 2020 crash, the all-time low for this metric was around 18%.

One of the interesting consequences is that miners are holding less ETH. It’s possible that they took their AUM and used it for staking — but since January 2018, ETH miners’ reserves have increased from ~400,000 to 83,000 as of this writing.

One way to measure how much ETH has grown is to look at the number of smart contracts deployed. As of this writing, there are about 24 million smart contracts on the network. That’s about eight times the 3 million during the 2017 rally.

Despite the huge lead, interest in deploying new smart contracts on Ethereum is waning. As a result, the numbers are more or less back to 2019 levels. Developer mindshare is likely being captured by other emerging layer 1s.

Stablecoins have one of the most beautiful growth trajectories of all cryptocurrencies. After a few years, their total market capitalization has exceeded $100 billion. 2022 is the year when USDC finally surpasses USDT in terms of market supply.

Of this ~$100 billion, Binance alone has a balance sheet of $26 billion. Although, if that number is partially pushed up by the 20 billion BUSD they hold. From a purely tether perspective - Binance has 20x more USDT than their North American counterparts (ftx, Kraken).

There may be gaps in this data as it may not include certain L2 and non-EVM chains. But in terms of transaction volume, despite the bear market, stablecoins have grown about 22 times in the past two years. From about 600 million per day to about 14 billion today.

In June 2020, the cumulative volume of stablecoin tx was about 500,000 per day, and today, it is close to 200,000.

What might happen is

1. Whales use staples as holding assets during bear markets;

2. Fewer wallets lead to greater stablecoin transactions.

The reason I suggest this is that transactions with a value of over $1 million account for about 90% of mobile transaction volume today, despite being less than 0.5% of the number of transactions. This is in sync with what we typically see in the traditional economy - so it's not surprising.

One interesting thing I noticed about this data is that transactions under $100 account for a quarter of all transactions done through stablecoins. People still use these networks to send small amounts for a fee.

This number may be lower due to the lack of coverage of non-Ethereum chains.

1. BUSD is the least used stablecoin in smart contracts;

2. 17% of USDT supply comes from smart contracts;

3. Nearly half of DAI and USDC are on smart contracts.

I’m curious how much trading of these stablecoins happens on Uniswap. At its peak — ~60,000+ stablecoin trades were done through Uniswap. Today, it’s closer to 10,000 which is a significant drop (as volume/activity decreases).

As recently as January 2020, we were seeing around $1 billion moving on-chain in stablecoins in a single day. That number is now up ~25x on a random day. In the chaos of May 2022 — ~87 billion worth of stablecoins moved between wallets.

This chart by Richard Chen shows the exponential growth rate of DeFi. From ~4,000 users in January 2019 to 4.7 million today. This does not include all the newer DeFi primitives. DeFi has ~10 million users, which is less than 0.3% of the network’s user base.

Whenever the media reports on DeFi — it rushes to explain how DeFi TVL collapsed without taking into account the drop in the prices of the underlying assets. It took until May 2020 for DeFi to reach its first 1 billion TVL, and we are still 50x.

Likewise, there is a general rush to clarify how MAU on DeFi collapsed. From my observation, in the absence of token rewards, there is a pullback, especially in transaction volume. But the users themselves did not disappear.

That’s not to say volumes haven’t taken a hit — we’ve seen volumes on DeFi platforms drop from a peak of $250 billion to around $100 billion. But here’s the interesting thing — $100 billion is still 5,000x the average monthly volume of $20 million in 2019.

Lower transaction volume + lack of token incentives = lower fees generated

At their peak, prominent DEX platforms generated around $600 million in fees; we are now down to around $80 million in fees.

TVL on these platforms has dropped from $50 billion to around $20 billion — but in terms of the drop in activity here, it seems like we’ve formed a bottom. Those who are still LPing and trading are sticky users and are likely to continue to do so.

The drop in TVL also applies to lending apps — down about 50% — partly due to the May crash. At their peak, lending apps processed about 58 billion, now it’s around 18 billion.

Lack of volatility + demand for lending hit fees on lending apps — cumulative monthly fees on these platforms dropped from ~120 million to ~20 million.

Borrowing demand has increased from ~30 billion to 13 billion - it's safe to say that yield + platform fees have been on a long-term decline. I do find it interesting, though, that there has been an upfront fee to this price over the last few months.

Looking at the price/fee ratio, players like Maple and Goldfinch are still in tradfi bank multiples. Obviously, players like Aave and Compound have a premium associated with them, but prices are more or less in sync with expected fees.

Over the last two years, the lending apps alone have generated $1.2 billion in fees -- this is a cumulative chart. We're still coming down.

Likewise - exchanges collected ~$5 billion in fees over the last 2 years, and fees here are likely a mix of platform revenue + token incentives - so in the future we may see a bit of stagnation here unless there is a major spike in platform activity.

Despite the poor market conditions, parts of the market are still growing. Take GoldFinch for example - their lending volumes remain strong despite the downward trend.

Maple Finance’s lending desk has collected $40 million in fees last year. CeFi lending channels have historically struggled to maintain relevance and customer base.

In fact, under-collateralized lending may be one of the healthiest sectors in this market cycle. The chart below is for Centrifuge - their fees and borrowing volumes are both at ATHs.

GMX (comparable DYDX) has seen a similar trend. Last month, the exchange had a trading volume of around 8.6 billion, which is a new ATH. Fees for the month exceeded $13 million. It is safe to say that the DeFi sector is still healthy and strong.