This year was a sh*tshow for centralized crypto financial services, aka CeFi.
FTX self-immolated spectacularly. Coinbase, the largest public crypto company, saw its stock decline 80% as crypto trading volumes evaporated and U.S. regulators remained hostile towards token markets. Western crypto infrastructure giants laid off hundreds of employees. And the crypto lenders born in 2018-2020 are pretty much all dead, or mostly dead.
Here's a partial list of "unicorn" lenders and prop desks, who ate nine-figure losses this year and/or went bankrupt: Alameda, Babel Finance, Blockchain.com, BlockFi, Celsius, FTX, Genesis Capital, Jump Capital, Three Arrows Capital, and Voyager. I'm sure there are others; I just got tired of looking them up.
It's been a classic credit boom and bust, with a crypto twist. Many of our problems this year stemmed from an over-reliance on unsustainable DeFi yields and synthetic trades. A daisy chain, if you will, of poor risk management and compounding bad bets.
Amidst the wreckage, what remains? How can we build more sustainable infrastructure? Who will lead us back from the brink?
We covered Binance's and Coinbase's dominance in the crypto exchange market in the "People to Watch" chapter. These giants are hardly the only important exchanges worth keeping an eye on in 2023, though. In fact, it's the mid-market services that we'll be tracking most closely for signs of stress. Here are the biggest risks we're monitoring in the wake of the FTX fallout.
Binance: Global regulatory risk? They operate internationally in a "jurisdiction-light" manner that ultimately may get them into hot water with various policymakers around the globe. Whether they're deemed to be operating outside of the boundaries of different countries' existing financial laws or simply make for good political targets, there's reason for concern. With a 75% share in global crypto trading volumes, the industry desperately needs a Binance rival to emerge. And I could make the argument that Binance itself would be healthier with a stronger trading rival.
Coinbase: Hostile public markets? Coinbase avoided the crypto lending markets, and it sits atop the leaderboard as the top global custodian of crypto assets (about 10% of the crypto market cap sits at Coinbase), the top USD exchange, and the first major crypto IPO. Still, the markets have punished the company's stock: Coinbase's $3.5 billion debt yields 15%, and its market cap has fallen below $10 billion. A clean bill of regulatory health and U.S. market leadership could raise eyebrows at PE shops and strategic suitors alike, but Coinbase has dual class shares, and Brian Armstrong retains voting control, so hostile takeover attempts seem unlikely to succeed.
Bitstamp, Kraken: Regional (European) regulatory risk? I don't know enough about the various regulatory constructs facing the big EU-dominant exchanges. We'll cover MiCA in the next chapter on crypto policy, but I'd expect Bitstamp and Kraken to be disproportionate winners (not losers) with better regulatory clarity in Europe.
OKX, Huobi, KuCoin, Upbit: Regional (Asian) regulatory risk? Same thing with the largest Asia-based exchanges. I don't know enough about the various regulatory constructs in Asia, and the region doesn't benefit from a common legal and financial regulatory standard like we see in the EU. In general, Korea and Japan are good places for business, China and Hong Kong are iffy, and India is a bit of a wild card (but trending negative).
Bitfinex: Tether risk? Bitfinex's close historical affiliation with the stablecoin giant Tether opens the company up to multiple regulatory risks. Bitfinex reached a joint settlement with Tether in 2021 in which the CFTC asserted (among other things) that Tether had "commingled reserve funds with Bitfinex's operational and customer funds, and held reserves in non-fiat financial products." Bitfinex and Tether share several core team members, including CTO Paolo Ardoino, and it's hard to imagine that issues at one entity would be completely isolated from the other.
Gemini: Contagion risk? (Read the sections below.)
All of these major exchanges carry operational risks as well – market declines can lead to financial stress, and hacks and customer deposit losses have been a pervasive component of crypto for a decade.
But at least one silver lining to the FTX crisis was its catalyzing effect on the "proof-of-reserves" movement. For years, many have clamored for on-chain evidence of major exchange reserve assets in order to ensure top platforms were solvent and willing to prove they weren't going fractional or otherwise misappropriating user assets. FTX proved – too late – that those concerns were absolutely warranted, even for the "smartest guys in the room."
There is still work to be done on the "proof-of-liabilities" side of the ledger, but we now at least have unprecedented transparency around exchange assets. The days and weeks following FTX's collapse were record-setting in terms of asset withdrawals from the top crypto exchanges. "Bank runs" were everywhere for those running lending books, but for those who weren't breaking their own rules, and gambling with customer deposits, things were fine. Nansen killed it in quickly assembling a public repository of exchange reserve data. Kraken and others publish proofs-of-reserves without publicly divulging info, and Coinbase is audited quarterly as a public company. Some $200 billion in crypto assets held on exchanges are now fully accounted for and trackable quarter to quarter.
The Institutions are Coming
Despite the market slowdown, there was surprising and meaningful crypto adoption by a number of legacy institutions.
[This is the part of the report where I pretend to care for a minute about institutional crypto efforts. Insert something about [awesome legacy Wall Street firms and Big Tech pioneers], and how this time they really will continue to invest in crypto during a lengthy bear market.
Or leave this in brackets and "forget" that you didn't update this section, blame a late night editing oversight, and say "Oh damn, I had a whole spiel about how awesome your [insert recent press release from the crypto team at the institution] was. I'm kicking myself for that process error!" then make up for it next year with no consequence via another vague update about how, for the 1 th year in a row, the "institutions are coming" after one of the banks or asset managers buys one of the exchanges listed above.
In case this accidentally doesn't get cut via a process error, write that you love the companies listed above and this is just a very funny draft joke and misunderstanding, so the sales team can honestly tell them "Selkis said he loves you guys" and then some of them may forgive you for writing 140 pages for free, but not spending an incremental day gouging your eyes out reading up on corporate blockchain initiatives. Maybe one of the esteemed institutions that I really love quite a lot will have a sense of humor and respond by commissioning a report on corporate blockchains because "they REALLY DO care about crypto" and "this deserves a look."]
Anatomy of a Crypto Credit Crisis
The Grayscale Trade, aka crypto's "Widowmaker," was integral in helping create much of the crypto contagion we saw this year.
It was a root cause of the Three Arrows Capital (3AC) and BlockFi bankruptcies, and its potential ripple effects on its distressed sister company Genesis Capital – and Genesis counterparties like Gemini – remain unresolved and could cause further damage still. The Grayscale products themselves continue to deteriorate for their investors as the discount to their fair value (the underlying assets held in the trusts) have widened to 40%, with no fee reductions or ETF conversion on the horizon.
This spring's Terra/Luna failure was simply a haymaker that followed years of body shots from the slow-bleeding bad bet on GBTC. Yield-hungry investors, forced further out onto the risk curve as Ethereum-based DeFi remained mired in a multi-year recession, gobbled up 20% teaser yields on an emerging algorithmic stablecoin (UST) and its rising star lending protocol (Anchor), not realizing it was laced with arsenic. It was a good reminder to use common sense in investing. If you don't understand the yield, you are the yield.
Let's set the stage for this section on CeFi with a speed run through crypto's first credit crisis. CMS wrote up the cliff notes in just five tweets that explain how this all went down, but I will also attempt to summarize here:
It started with the Grayscale Trusts. These vehicles allowed investors to buy GBTC in their 401ks through OTC traded securities. But they weren't ETFs, so they didn't have typical creation and redemption mechanisms. Instead, accredited investors could create Trust shares with bitcoin, hold the shares for a six-month "seasoning" period, then flip them for what was a hefty GBTC Premium for quite a while pre-2021.
Flipping and rolling the GBTC Premium ballooned in popularity and The Grayscale Trade got crowded in 2020 hitting $40 billion in AUM at its peak due to stimulus, bitcoin halving, and zero interest rate COVID policies of 2020. 3AC and BlockFi accumulated 10% of the Trust's shares with $4 billion in exposure at the peak in February 2021. But then 3AC started offloading some of its exposure during the great GBTC Premium Crash.
Now a good chunk of the seasoning - GBTC is trapped and underwater with 3AC and BlockFi still subject to six months holding restrictions. They eat the unrealized losses and instead lean on lending desks to allow them to borrow against the GBTC collateral. Genesis Capital, a sister company to Grayscale, is one of the only lenders incentivized to treat the GBTC at good money, given its affiliate technically controls the share's redemption mechanism (that would make whole the principal of the collateral), and Genesis can milk the big borrowers for interest in the interim.
Exposed lending desks (BlockFi) and funds (3AC) let GBTC ride, but now they have to push further out the risk curve. This isn't a big deal in 2021 because everyone is making money hand over fist. The markets have come down from their tops in November, so the tide begins to go out. But Luna is still growing massively and raises a $1 billion round in Feb 2022 to diversify their treasury. 3AC is a big Luna investor.
The contagion hits full swing, as multiple funds and trading desks with ties to 3AC go under (Defiance) or get bailed out temporarily by FTX, who is also a lending counterparty(???) (BlockFi, Voyager). All duration bets in the crypto lending markets sour and die, and Genesis' active loans drop from $14.6 billion at the end of March to $2.8 billion at the end of September. As active deposits shrink, lending desks also call collateral and yank borrow from funds wherever they can. Credit seizes as everyone accelerates their derisking. This includes Alameda in August, who has external borrow called after Genesis realizes they don't care much for FTT as collateral anymore.
Coindesk gets the scoop of the year and publishes details on Alameda's financials and token reserves. They are loaded with illiquid crap like FTX's own trading token FTT and several low-float DeFi tokens like Serum that FTX had backed and hyped in 2021. An analyst notices that Binance moves $2 billion worth of FTT on-chain, and speculates that CZ is preparing to dump the position. CZ confirms that he plans to sell FTT and fully sever ties with FTX days later. Uh oh. That's $2 billion of an illiquid token. FTT crashes, as do other major FTX / Alameda positions despite their bestefforts at damage control. The entities go massively underwater as their collateral is now worthless and have no liquid accounts or access to other credit (all of their counterparties are already dead or distressed). Even CZ says yuck, I'm not buying this.
No liquidity at FTX/Alameda and commingled customer funds (thanks to mislabeled accounts or some nonsense) causes customer funds to be at risk, there's a "bank run" on FTX that isn't really a bank run because FTX was not authorized to lend against customer deposits according to its terms of service. FTX dies and, as CMS sums up, "Margaritaville at risk."
That's the synopsis so far (as of December 21, 2022), and the lessons are fairly straightforward: don't commingle customer assets, cut losses early on bad trades vs. lever them up and pray, maintain internal controls and a fortress balance sheet, split assets across different custodians and counterparties, and of course, only keep on exchange what you can afford to lose.
But there's potentially more to work out still…
DCG & Genesis Contagion Risk
The most important trend to keep an eye on in early 2023 will be the evolving situation over at investment giant, Digital Currency Group ("DCG") and its lending arm Genesis Capital, which was a large counterparty to 3AC, FTX, and most other large lending and trading desks.
DCG is now one of the most systemically important companies in the crypto ecosystem, as the liquidity crisis at its subsidiary and $1 billion capital infusion requirement present further contagion risks for the industry. Gemini, and at least one other large European exchange, and dozens of high net worth creditors apparently have more than $2 billion in frozen deposits stuck at Genesis, whose primary borrower is its DCG parent.
The options look pretty bleak. Creditors could strike an out-of-court settlement with Genesis and agree to a haircut on their withdrawable deposits in exchange for other DCG debt or equity instruments. Genesis could file for bankruptcy protection, and potentially drag their parent and their deep-pocketed external creditors through a lengthy and expensive reorganization process. Or DCG could identify recapitalization options at the holding company level, in order to make whole Genesis Creditors and limit their liability, but leveraging its other "good" assets.
I wrote a full-length Enterprise research report on why a DCG recapitalization is a good idea and likely necessary to restore some stability to the crypto markets. (Laura Shin also hosted a good podcast on the subject.) But much of the viability of that plan ties back to the details regarding what's in the black box of lending agreements between DCG and Genesis.
Here are the five open questions I'd be diligencing if I were looking at the deal, and determining whether there can be a resolution, or this is now an untouchable business.
1. Where's the Beef: Does DCG or Genesis hold the majority of the combined companies' $700 million worth of GBTC and ETHE shares? If DCG, that's a big slug of assets to borrow against. If already spoken for at Genesis, and Genesis still has a billion dollar hole to fill in its balance sheet, we might yet see a further rippling out of contagion.
2. The Promissory Note: On a recent episode of Unchained's "The Chopping Block" podcast, Dragonfly partner Haseeb Qureshi said that the $1.1 billion "promissory note" that DCG extended to Genesis following the 3AC bankruptcy may have been structured as "callable" in the event of a Genesis liquidation. If true, Genesis could have treated the promissory note as a "current asset" (less than one year duration) even though it was nominally a ten year note, something that would have been a material part of current assets Genesis showed subsequent creditors. Again, if true, that might reduce DCG's ability to limit liability from a Genesis bankruptcy. A callable note would mean that a Genesis liquidation process would put DCG on the hook to repay the full $1.1 billion immediately. DCG doesn't have that cash yet, so Genesis might not feel the urgency to rush into bankruptcy as they "have the assets" from DCG, if DCG can refinance.
3. Alameda Lending: Given the fact that Alameda and Genesis were two of the world's largest borrower-lender counterparties, it's likely they had loans together. From Genesis's quarterly reports, it appears that they were responsibly winding down many of their positions and yanking borrow. From the FTX bankruptcy filings, it also doesn't appear that FTX or Alameda are Genesis Capital creditors today. If that's true, then the two giants either had no relationship (unlikely) or they closed their positions. The precise dates of any closed positions with Alameda might end up being critical to the resolution for DCG-Genesis due to the 90 day clawback period that most bankruptcy cases contain. If Alameda had loans with Genesis that were repaid after August 13, they might *potentially* be subject to the clawbacks. If there was a big number that changed hands after August 13, I'm not sure how someone new to DCG assesses the risk of a clawback, which would be a *long-term liability* that hinges on the results of a multi-year, incredibly complex FTX bankruptcy process.
4. The Grayscale-GBTC Tie Up: Some of the details in Fir Tree Capital Management's lawsuit vs.Grayscale look pretty alarming. They point to the related party levered transactions with 3AC, the repeated tightening of Grayscale's control over the Trusts' redemption mechanisms (at the expense of shareholders), and Grayscale's ability, but "self-interested" refusal to pursue Reg M redemptions outside of an ETF conversion. That will take a long time to play out, but the one thing you should have your eye on as a GBTC shareholder is whether DCG's GBTC shares and Grayscale stay under common control for the foreseeable future. The alternative would not be good.
Grayscale throws off $200 million+ in annualized cash flow even at today's distressed prices. Those assets under management are permanent capital given the trusts' structure (see below). So the question for a new DCG investor or creditor regarding what Grayscale is worth as a business revolves around your forward outlook for bitcoin. Grayscale may do $400 million in EBITDA this year, but only half of that on a run rate based on current prices. If DCG explores a sale of Grayscale, they'll NEED to put the 67 million GBTC shares they own into the deal, too.
A buyer without a large GBTC bag would have every incentive to shut down ETF conversion discussions and run the business as an annuity that's openly hostile to GBTC shareholders. DCG's $550 million of GBTC acts as an "ETF approval hedge" since they are financially incentivized to push for an ETF. Even though an ETF would open the door for redemptions and lower fees, DCG would notch a one time gain of $450 million from the closure of the GBTC discount to NAV. The incentives of the trust sponsor get ugly without that share hedge. A new buyer could be exploitative.
5. The Eldridge Revolver: The irony in all of this is that the smallest creditor could hypothetically become the most troublesome. Connecticut-based lender Eldridge had a $350 million revolving line of credit with DCG that they could have considered to be in cross default the moment that Genesis halted withdrawals in November. Since they are senior creditors at DCG and Genesis, their incentives are materially different from the Genesis credit holders, who seem much more inclined to strike a deal.
In my mind, nothing else really matters in the markets right now.
Until we see a bit more color around the DCG-Genesis resolution, it's tough to say the credit crisis has fully resolved. It's 50-50, at best.
Grayscale: Reflexivity up. Reflexivity down.
There's a good case to be made that Grayscale was the entity most responsible for Bitcoin's ascent in late 2020. GBTC and ETHE assets under management exploded thanks to the Grayscale trade, and now the asset manager generates ~$300 million/year in high-margin, sticky annualized revenue, even at today's crypto prices.
Grayscale's large AUM base and its Hotel California structure make it an attractive target for DCG's suitors. But it's DCG's ownership of underlying GBTC and ETHE that I'm watching most closely in February. That's when Grayscale's 10-K drops, which includes notes on affiliate ownership of shares in its Trust.
DCG and Genesis (Grayscale affiliates) spent the better part of the past two years absorbing all of the sell pressure from the top GBTC trust shareholders.
But that $700 million in collateral (including ETHE, too) could be subject to forced selling in the event DCG and Genesis need it in the short-term to make their creditors whole. I personally think the GBTC is better held as collateral that can be used to help refinance DCG's current debt load. (See the "ETF Approval Hedge" above.) It comes down to who holds the GBTC.
The maximum amount that DCG and Genesis would be allowed to sell per quarter in the markets (under Rule 144 restrictions) would be 6.9 million shares, or about $80 million if they had sold during the first six weeks of the quarter when GBTC's discount to net asset value widened from 35% to 45%. (It would take the combined companies 2.5 years to unload the full position if that was their ultimate wish, as I explain here.)
The nuclear scenario (which I think is unlikely) would be for Grayscale to dissolve the trusts. That's something that would likely only happen if 1) Genesis went bankrupt, 2) DCG was pulled into Genesis's bankruptcy and also went bankrupt, 3) DCG was unable to spin off Grayscale to a buyer and exhausted all other financing options, and 4) the Trusts themselves were unable to find another financially stable sponsor. Like I said, unlikely.
This whole thing is a bad look for crypto, but it's also a worse look for the SEC.
GBTC was allowed to become toxic collateral because of the SEC's obstinance and dereliction of its duty to the American investing public. In a parallel universe in which the SEC prioritized investor protection, fair and efficient markets, and capital formation (its mandates), we might have escaped a great deal of the crypto credit carnage. GBTC investors wouldn't be billions of dollars underwater, institutions could begin to treat digital gold as a complement to their physical gold hedges via titled securities, and we wouldn't be staring down the barrel of a dozen crypto lending bankruptcies.
Instead, the SEC is taking a victory lap over the carnage that they created after being played for fools and inviting the fox into the henhouse. It's utterly despicable.
Custody, Security, Stability
The problems with crypto lending turn out to be pretty obvious: crypto assets are volatile, and many are illiquid, which makes them bad collateral. There are no free lunches when it comes to interest rates (you pay for returns through higher risks), and there are no lenders of last resort or deposit insurance.
For now, there are three things to note about the crypto lending industry moving forward:
Centralized lending is now mostly dead, and will be for a while. It will likely take years, and likely comprehensive legislation and regulation before the crypto lending markets reopen and look anything like they did at their peak in late 2021 and early 2022. That's arguably a good thing.
A credit collapse isn't necessarily a bad thing for the long-term health of the crypto markets. One concern of mine that has grown with the prevalence of lending and margin within crypto is that rehypothecation would eventually create systemic risks in the market (and suppress prices) by allowing entities to go fractional with their crypto reserves. That's less likely to happen any time soon.
I probably won't touch centralized crypto lending with a ten-foot pole ever again. (I used BlockFi to help with my home purchase a few years ago because no banks would underwrite my mortgage at the time, lol.) Give me DeFi and its transparent collateral pools and mathematical liquidation functions from now on, or nothing. (More on that in the DeFi chapter.)
On the other hand, we could start to see more stablecoin lending services emerge, now that Treasury rates are above zero, and fully-reserved custodians are incentivized to make a vig on those customerdeposits if they can. Long BitGo, Anchorage, Coinbase, and FireBlocks. Short lenders who don't also own the underlying custody solution.
Cloud Infrastructure and Developer Tools
Decentralized Physical Infrastructure Networks are still immature (more on DePIN in Chapter 9). In the meantime, we're building on the same centralized cloud as everyone else in tech.
Technical complexity and economic barriers prevent the average crypto user from solo staking or mining; instead, they usually delegate to a professional validator or mining service. Over half of all stakedEther is controlled by three entities: including exchange giants Kraken and Coinbase. Over half ofBitcoin's hashrate is supplied by three mining pools: Foundry USA, AntPool, and F2Pool.
Centralized RPC providers are currently the lowest cost and easiest to use solution for reading blockchain data – Infura and Alchemy are the default solution for popular tools like MetaMask and Open- Zeppelin. Worse, cloud providers AWS, Hetzner, and OVH represent nearly 70% of hosted nodes on both ETH and SOL as of Sept. 2022. In fact, Hetzner's terms of service don't even allow PoW mining or PoS validating-related activities. Users are at perpetual risk of being deplatformed.
Geographic concentration of validators could expose networks to geopolitical risks, regulations, or acts of nature. As of Sept. 2022, ~50% of Solana validators and ~60% of Ethereum validators were located in either Germany or the U.S., even if the broader networks were distributed more widely across 25 and 60 countries, respectively.
We also expect to see redoubled efforts to alleviate some of today's pressing software-driven centralization pain points.
Expect meaningful upgrades to account abstraction (streamlines user management of accounts and assets), Proposer Builder Separation (delegates computationally intensive work to specialized block proposers, lowering node requirements for solo validators), and "Light Clients" (allows users on laptops and phones to connect directly to networks rather than through third-party hosted nodes or centralized RPC providers thanks to improved data compression).
On-Chain forensics has been big business for a while. Chainalysis was the first crypto data unicorn, and its counterparts Elliptic and TRM Labs don't appear to be far behind. The primary thrust of these businesses has been transaction monitoring for compliance teams and forensic tools for regulators and investigators. Demand for these tools follows market cap on the way up (more firms need the tools, more fraud gets committed, and more taxes are evaded), and stays somewhat sticky on the way down (investigations take years, compliance needs are permanent, and taxes due from a bull run are still due in full even if your bags get slashed the next year.
On the other hand, I would argue that on-chain analytics as a standalone market intelligence product only really became viable in the past year. Glassnode is bootstrapped, but Dune, Nansen, AmberData, Flipside, and Coin Metrics all provided a glimpse of investor expectations for the segment as all raised big rounds on sky-high revenue multiples within the past year. On-chain fundamentals are getting sexier and more integral to professional traders' risk scoring (and copy-trading).
Nansen seems to be one of two analytics companies that have figured out how to combine a market intelligence product and a compliance product. (I'll give you a wild guess as to who the other under-the-radar company is.) It was a brilliant and non-intuitive bootstrapping strategy: tag public wallets, analyze them on a best-efforts basis, and leverage Cunningham's law into iteratively better aggregated data sets ("if our data is wrong, then give us the right data"). That allowed them to hide a robust forensics tool under an NFT flipper's front end. Then low and behold they came fastest out the door with a proof-of-reserves dashboard post-FTX failure and deconstructed multiple high profile protocol and fund failures using their own dogfood in Q2. Credit where it's due. Killer year.
And then there's Messari.
The company that many people are calling the "greatest on earth" only really entered the on-chain data game in the past year. But we've been building open APIs and data standards for 100+ protocols, and are getting iteratively closer to developing the equivalent of GAAP or IFRS reporting standards for crypto protocols. Who needs the SEC and a quarterly lag. These metrics are available in quarterlyreports, or real-time. Dealer's choice:
Tax & Compliance
My base case for 2023 is that the IRS is going to create hell for crypto investors. What happens when a) the IRS is directed through legislation to extract higher taxes from crypto investors, b) we have a record year (2021) followed by a deep correction (2022), and c) compliance and tax monitoring tools have to take down rounds and scramble for new customers as winter arrives?
Compliance solutions are fairly sticky, and they'll sell wherever they are most welcome. It would not surprise me to see some crypto accounting and forensics companies go full Big Brother in the new year, and proactively search for and report suspicious accounts for tax non-compliance.
I am in favor of tax compliance, and using every tool at our disposal to catch hackers, rogue states, and criminals who leverage crypto and use it for money laundering or tax fraud. But I am very much against dragnet surveillance and taxpayer harassment. And unfortunately, I think crypto investors will make appealing political scapegoats. After all, the IRS has to give those 25,000 new agents something to do!
I stand by my comment last year that crypto tax reporting violates Eighth Amendment protections against cruel and unusual punishment. But at least this year we all lost so much money that we'll all get refunds.
You've got a week to harvest those losses, baby. Get going!
Research & Data Upgrades
While I'm sure there are other good research and data companies out there, I don't make a habit of giving my competitors air-time in down-trending markets. I write to win (this report), and firmly believe that there is no second best when it comes to crypto research.