In the early 2010s, the crypto market consisted of a handful of small, retail-focused brokers. Over the last decade, the market has grown to over $15B of daily volume across spot, derivatives, and on-chain markets. Today, BTC sits among the most liquid assets in the world.

The quirkiness of crypto-assets has created a novel market: Crypto markets are open 24/7. Users all over the world can access completely fungible assets. Public blockchains allow frictionless transfer of crypto and fiat in minutes. Individuals can custody assets themselves as securely as the banks. Unlike the NYSE or Nasdaq, retail users can trade on the largest crypto exchanges directly without intermediation.

Even with these advantages, the crypto market structure is opaque and poorly understood due to global fragmentation, rapid change, and the diversity of participants.

This essay outlines the two significant evolutions in the crypto market over the last decade and articulates a vision for what the future could hold.

Market Structure 1.0: 2010 to 2017

The prehistory of the crypto market began with p2p trading over Bitcointalk and IRC. The first real market structure began in earnest in July 2010 with the launch of Mt. Gox. Over the next five years, many early exchanges launched fiat on-ramps to service individuals.

With bitcoin’s nascency, accessing stable banking channels was a challenge, leading to the rise of stablecoins like Tether. As adoption grew, over-the-counter (OTC) desks began servicing the handful of early institutional firms. Without sophisticated market makers, liquidity was poor. Cross-border and cross-venue spreads were often measured in single-digit percentages.

The influx of new market participants in December 2017 overwhelmed exchanges daily and marked the end of an era. It’s not worth reflecting on this bygone time further; thank you to the early builders who got us here.

Market Structure 2.0: 2018 to Present

Since December 2017, the crypto market evolved to a 2.0 iteration, growing from a market designed for individuals to one that’s institutionally accessible. Over this time, derivatives volumes grew over 25x while bid-ask spreads fell 10x. The market matured from manual, expensive, and BTC-denominated to fully electronic, cheap, and stablecoin-denominated.

The major drivers of this shift are:

• Derivatives liquidity eclipsing spot: Since 2017, the epicenter of crypto liquidity shifted from the spot market to derivatives. Crypto derivatives lagged spot volumes in 2017 but now trade 3-5x as much at over $10B/day. As two-sided liquidity increased, Bitcoin volatility fell materially, with 60d volatility ranging from 2-4% currently, from 4-8% in 2017-18, and 7-10%+ in 2013-14. Today’s derivatives landscape is diverse, including the US-regulated market (CME, Bakkt), global players (FTX, Deribit), and offerings from international spot exchanges (Binance, Huobi, OKEx).
• Electronification of OTC execution: Since 2017, OTC spreads have compressed an order of magnitude, from 50-200bp to 5-10bp today for an 8-figure BTC trade. In 2017, OTC trading was primarily over voice and chat. Today, it is fully electronic, dominated by quantitative trading firms like Jump, B2C2, Amber, and Alameda Research. Rather than firing up Skype, clients can connect directly to platforms hosted by market makers and stream quotes, execute trades over API.
• The emergence of lending: In 2017, there was nearly 0 credit in crypto. Today, trading firms can access over $2B of BTC and stablecoin borrow, with desks intermediating both retail (BlockFi, Celsius, Blockchain.com) and institutional (Genesis) lenders. The lending market has lowered market makers’ capital costs, benefiting institutional clients with tighter spreads and retail users with high single-digit yields.
• Stablecoins as the crypto-finance reserve asset: In 2017, nearly every crypto-assets’ dominant trading pair was BTC-denominated. This resulted in significant price dislocations and evaporating liquidity during high volatility periods. Today, the most liquid trading pair of almost every top-30 crypto asset is stablecoin-denominated. Stablecoins have replaced BTC as the crypto market’s reserve asset, reflected in 10x growth in total stablecoin issuance from January 2018 ($2B to $20B).
• Institutional services and products: In 2017, institutions were limited to accessing the market through retail channels. Today, institutions can work dozens of qualified custodians, electronic execution, and lend/borrow markets new entrants like Tagomi, Fireblocks, and Anchorage in addition to incumbents like Coinbase and Genesis/BitGo. ECNs like Paradigm.co have improved block trading workflows while specialized venues like LMAX Digital now lead global BTC:USD liquidity.

Market Structure 3.0: ~2020 to ?

We’re in the early stages of the next structural evolution, from 2.0 to 3.0.

At maturity, the 3.0 iteration will (1) be radically more capital-efficient and (2) bridge centralized markets to emerging decentralized finance (DeFi) markets.

1. Capital efficiency

Crypto trading remains capital inefficient due to market fragmentation and a lack of industry-wide credit assessment.

Today, exchanges have high margin requirements, and firms cannot cross-margin, i.e., use margin posted at one broker or venue to collateralize a position elsewhere. This forces firms to fully fund nearly all their trading activity and subjects trade settlement to many block confirmations (at-minimum). Full funding is especially taxing in highly volatile environments when on-chain congestion is most significant.

As a result of these inefficiencies, the perpetual swap has become the dominant source of short-dated funding. Crypto perps: “Came for incremental capital efficiency, stayed for the 20x leverage.” Dependence on the perp market as the backbone of market funding is very sub-optimal: cascading liquidations in March 2020 resulted in over $1.6B notional liquidations on BitMEX alone, much of which should be preventable in a more capital-efficient market.

Credit creation and shortened trade lifecycles drive capital efficiency. Credit creation allows firms to get more than $1 of buying power per $1. Shorter trade lifecycles mean the same $1 can go around more times, increasing liquidity per dollar.

Some specific ways we’ll see credit creation and shorter trade lifecycles:

• Dedicated prime brokerage: Large exchanges have the most robust balance sheets in crypto, allowing them to extend significant credit directly. Indirectly, PBs like Coinbase enable clients to margin across venues and cold storage speeds up trade lifecycles by moving transfers off-chain.
• Crypto-native derivatives clearing: In traditional markets, exchange-traded derivatives are cleared by central clearinghouses who maintain the ledger of margin calculations. Over the last few years, firms like Zero Hash and ErisX have attempted to port this model to crypto directly. Alternatively, approaches like X-Margin could achieve this crypto-natively by making cryptographic collateral attestations on-chain.
• A formal repo market: The $2-4T/day repo market allows institutions to borrow cash on a secured short-term basis. Crypto already has an informal repo market through the perpetual swap funding, and bi-laterally settled OTC repo ($50M/day). A formal institutional repo market could enable sizable near-term borrowing without exposure to perp venues.
• Lower on-chain confirmation thresholds: Superior understanding of public blockchain settlement assurances drives faster deposit times between known counterparties. Fireblocks’ Digital Asset Transfer Network settles over $25B/month on-chain and allows members to opt into “instant” deposits with each other. Fireblocks-initiated transfers are SGX-signed and ensure that an (unconfirmed) transaction is the only one signed for a given UTXO. These guarantees allow exchanges like FTX to accept deposits from other Fireblocks users, as soon as the transaction hits the mempool.

2. CeFi ↔ DeFi convergence

Decentralized finance (DeFi) first appeared in late 2017 and grew parallel to Market Structure 2.0. As DeFi continues to gain structural importance, CeFi and DeFi will converge as they see overlap in market participants, liquidity pools, and product UX.

Even in its 1.0 iteration, DeFi has already begun disrupting “CeFi.” Some examples:

• Liquidity is built on AMMs first: In 2017-18, building liquidity in a nascent asset required working with exchanges and market makers. With AMMs enabling permissionless listings, passive retail LPs ca create market depth before a professional LP even has inventory. Once the starting point of liquidity, major global crypto exchanges are now a late mover in the long-tail.
• Best execution requires on-chain interaction: AMMs like Curve now hold nearly $1B in stablecoins. The direct implication is that brokers unable to access on-chain liquidity have an immediate disadvantage versus those who can. This shift happened quickly; the impact of AMMs on centralized liquidity was visible as recently as March 2020. For many MMs, this has been a real forcing function to onboard to DeFi.
• Crypto-native cross-margining: Margin positions in DeFi have been tokenized since the start, taking forms like Uniswap LP shares, Compound cTokens, and Synthetix synths. DeFi margin tokens are fully-collateralized and callable on-chain, enabling transparent rehypothecation when used composably (e.g., using Uniswap LP shares as Maker CDP collateral). While FTX has already pioneered tokenized margin on centralized venues, the design surface is only now opening up.
• DeFi’s frictionless UX: DeFi UX is better than CeFi in many ways. While critics often focus on gas fees, DeFi offers users superior security UX (non-custodial) and frictionless access. Scanning a QR code and signing a MetaMask transaction is accessible and closer to using Snapchat than fiddling with a traditional brokerage.

Though DeFi already has deep spot trading and lending markets, DeFi hasn’t “eaten” CeFi yet. Throughput and high gas fees remain significant structural barriers. As L2 solutions come to market, DeFi presents a real threat to centralized venues as applications like Synthetix, which can replicate BTC-margined synthetic assets with UX that’s comparable to using FTX.

What does this mean for CeFi players? A few implications:

• Better DeFi interfaces: Exchanges will lean on their scale efficiency to intermediate DeFi for users the same way they have with staking services. Offering interfaces to access DeFi is a natural way to prevent capital flight on-chain. Providing liquidity of locked assets, lower fees (via pooling), and additional off-chain margin are some ways exchanges can incentivize users to access DeFi through their interface. For many users, an exchange account may be the most convenient way to access on-chain protocols as the default wallet.
• Normalization of non-custodial trading: Non-custodial exchange offerings are another natural “defense” against capital flight. Binance and FTX have fully leaned into the challenge, building non-custodial DEXs Binance Chain (a Cosmos zone) and Serum (on Solana). Protocols like Arwen can enable non-custodial trading for exchanges pursuing a hybrid approach. Newer projects like dYdX and DeversiFi (originally via Bitfinex) are building to scale and aim to compete with centralized UX, powered by StarkWare’s ZK-based L2.
• “CeDeFi” is real: Beyond non-custodial window dressing, every major CeFi player will attempt to capitalize on “CeDeFi” unironically. Low effort solutions could be structured products trying to mimic on-chain yields. A more comprehensive solution could include full EVM-compatible chains that can port DeFi, e.g., Binance SmartChain.
• Institutional DeFi support: Retail users have lower frictions to access DeFi relative to large institutions, subject to compliance or regulatory hurdles. As larger firms are beginning to view DeFi markets as a first-class citizen, service providers will seek to enable frictionless access for their clients without compromising existing workflows. Over time, we may even see some projects launch whitelisted (KYC’d) liquidity pools over time. Off-chain DeFi insurance could alleviate mechanism and contract risks far more coverage on a un/under-collateralized basis.

As scalability improves, on-chain financial infrastructure will begin competing with centralized infrastructure. However, the diversity of users and the importance of fiat on-ramps means centralized venues aren’t disappearing any time soon. The long-term winners are users, who can explore the spectrum of options across trust, price, risk, and UX.

Conclusion

The crypto market structure has gone through two significant evolutions over the first decade. Despite rapid innovation, the market is far from maturity or the scale needed to support a multi-trillion dollar market cap.

Crypto market change has always been grassroots, driven by entrepreneurs and user demand. While this has resulted in some wheel re-invention, the optimistic view is that innovation wins long-term—the crypto sandbox of today is the design of every major market in the future.

Historically, the crypto market has been opaque and poorly understood. We hope this essay provides a useful starting point for future dialogue as we build an open and efficient financial system over the coming decade.

If you are focused on Market Structure 3.0 (or 4.0+, we won’t discriminate), we’d love to explore it together. Please reach out at any stage, the earlier the better: arjun@paradigm.xyz

Thanks to the builders and size traders for valuable conversations on market structure—Tiantian Kullander at Amber, Kevin Johnson at Tagomi, Kyle Davies at Three Arrows, Michael Shaulov at Fireblocks, Josh Lim at Genesis, Tarun Chitra at Gauntlet, Darshan Vaidya at X-Margin, Dave Farmer + Sam McIngvale of Coinbase—and to my partners at Paradigm, particularly Fred Ehrsam, who’s spent most of the last decade building robust crypto markets.

Important Disclosures: The content of this post is provided for informational purposes only. Nothing herein constitutes investment, legal, or tax advice or recommendations. Nothing on this site should not be relied upon as a basis for making an investment decision and is not an offer to provide advisory services. It should not be assumed that any investment in the asset class described herein will be profitable and there can be no assurance that future events and market factors would lead to results similar to any historical results described in this paper. The assets discussed herein are not representative of all assets in which Paradigm invests. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in this paper are based on the subjective views of its author, are subject to change without notice and may differ or be contrary to views expressed by others. Certain information contained on this site has been obtained from third-party sources. While such information is believed to be reliable for the purposes used herein, Paradigm has not independently verified such information and makes no representation or warranty, express or implied, as to the accuracy or completeness of the information contained herein.