Kyle Samani Returns to Crypto: How On-Chain Markets Can Outcompete CEXs on Efficiency
Kyle Samani Returns to Crypto: How On-Chain Markets Can Outcompete CEXs on Efficiency
Kyle Samani, co-founder of Multicoin Capital, is back in the public crypto conversation with a long X thread outlining a familiar but increasingly practical ambition: beat centralized exchanges (CEXs) not just on ideology, but on efficiency. In other words, don’t ask traders to “sacrifice UX for decentralization”—make on-chain trading the rational default.
This article distills the core ideas behind that thesis, then expands them into a 2025-ready framework: what “efficiency” really means in crypto market structure, what technical building blocks matter most, and what users can do today to trade on-chain without taking unnecessary risk. For context, see Kyle Samani on X and Multicoin Capital.
Why “efficiency” is the only narrative that can dethrone CEXs
CEXs won the last cycle for a simple reason: they felt faster, cheaper, and easier. Even users who prefer self-custody often keep funds on exchanges because the “trading loop” is frictionless:
- Instant deposits and internal ledger transfers
- One-click cross-margin and liquidation logic
- High liquidity concentration
- Familiar order books and advanced order types
- Customer support and account recovery
But efficiency is no longer a CEX-exclusive advantage. As blockchain infrastructure matured through 2024–2025—especially around high-throughput execution environments, better wallet UX, and more sophisticated on-chain market designs—the gap has narrowed.
The key is to stop treating “DEX vs CEX” as a philosophical debate and instead treat it as a competitive systems problem.
Efficiency, defined: the five metrics that matter
When a trader says “this venue is more efficient,” they usually mean a combination of:
- Execution efficiency: tighter spreads, lower slippage, better fills
- Latency and finality: how quickly you can update or exit risk
- Fee efficiency: explicit fees + hidden costs (MEV, adverse selection)
- Capital efficiency: margining, netting, collateral reuse, funding costs
- Operational efficiency: settlement, custody risk, downtime, transparency
CEXs historically dominated (2), (4), and (5). The on-chain world has always had structural advantages in transparency and composability, but it often gave up speed and simplicity.
Kyle’s point is that the winning strategy is now: make on-chain win on (1)–(5) simultaneously, or at least win hard enough on total cost that traders switch.
The core argument: on-chain can be more efficient than off-chain
CEXs look efficient because they run a private database. But that “efficiency” comes with hidden costs:
- Custody and counterparty risk (users become unsecured creditors)
- Opaque risk management (insurance funds, liquidation engines, internal policies)
- Fragmented trust (proof-of-reserves helps, but doesn’t fully solve liabilities)
- Permissioned market access (jurisdiction blocks, account freezes, delistings)
On-chain venues can flip the model by offering:
- Atomic settlement: trade + collateral updates happen together
- Transparent solvency: risk is auditable in real time
- Programmable margin: smart contracts can replicate (and extend) prime-broker style features
- Composability: routing, hedging, and collateral management can be automated across protocols
- Permissionless innovation: anyone can build better UX on top of the same liquidity
The long-term claim is bold but coherent: if blockspace gets cheap enough and market structure gets smart enough, “CEX efficiency” becomes an illusion created by custody.
What must improve for DEXs to truly surpass CEXs
1) Fast, predictable execution (finality matters more than TPS)
High throughput is useful, but for active trading the real requirement is predictable confirmation and rapid state updates. That’s why performance-oriented chains and scaling systems keep attracting market-structure builders.
Two practical paths emerged by 2025:
- High-performance L1 execution with low fees and fast finality (for example, systems designed around parallel execution and high-frequency transaction processing).
- Rollups / L2s that focus on cost and UX while inheriting security assumptions from their base layer. A helpful high-level view is L2Beat.
The endgame is not “one chain to rule them all,” but execution environments optimized for trading—and bridges / interoperability layers that reduce the penalty of fragmentation.
2) Better DEX market structure: CLOBs, RFQ, intents, and hybrids
AMMs made DeFi usable, but AMMs alone are not the final form of exchange design. Today’s on-chain venues increasingly mix:
- Order books (CLOBs) for tighter spreads on liquid pairs
- RFQ systems for size and better price discovery
- Aggregators that route across venues to minimize slippage
- Intent-based trading where users specify outcomes and solvers compete to execute
If you want a neutral starting point on where DEX design is heading, compare classic AMM concepts with newer execution ideas via Uniswap documentation and broader aggregation / routing approaches in the 0x protocol docs.
3) MEV-aware execution (hidden fees are still fees)
One reason traders stay on CEXs: on-chain execution can feel like it has “invisible taxes” (sandwiching, priority competition, latency games).
By 2025, the industry focus shifted from pretending MEV doesn’t exist to engineering around it:
- Private / protected transaction flows
- Batch auctions or frequent batch execution
- Smarter routing and slippage controls
- More explicit fee markets that reduce manipulation
The goal is simple: make the expected cost of on-chain execution more predictable than a CEX spread + fee schedule.
4) Real capital efficiency: margin, netting, and collateral mobility
CEXs are hard to beat because they let traders do a lot with a little:
- Cross-margin across instruments
- Instant collateral conversion
- Portfolio netting
- Complex derivatives under one roof
On-chain, capital efficiency improves when:
- Stablecoins become ubiquitous settlement assets
- Perps and money markets integrate tightly
- Tokenized real-world assets (RWAs) (like tokenized Treasury exposure) become high-quality collateral
- Smart accounts enable programmable risk controls
For macro context on stablecoins and settlement, see BIS coverage on DeFi and crypto market structure (and related BIS publications reachable from there), and for adoption / ecosystem visibility, DeFiLlama remains a widely used industry dashboard.
5) Wallet UX must stop feeling like a hobby
Even if execution is great, users won’t switch if self-custody feels like juggling knives.
Two wallet-layer improvements are crucial:
- Account abstraction / smart accounts: session keys, spending limits, batched transactions, and safer recovery patterns. A good technical entry point is Ethereum’s EIP-4337.
- Safer permissioning: approvals, revocations, and clear transaction simulation
This is where self-custody products become part of market structure, not just “storage.”
What changed in 2025 that makes this thesis more credible
Without relying on hype cycles, several 2025-era realities strengthened the “DEX can beat CEX on efficiency” argument:
- Perpetuals and sophisticated derivatives became increasingly central to on-chain volume, pushing protocols to invest in tighter execution and risk systems.
- Stablecoins continued to act as the dominant settlement rail for crypto trading and on-chain credit.
- Wallet and UX infrastructure matured: better signing flows, improved simulation, and more widespread smart-account patterns reduced the fear factor of self-custody.
- Tokenization narratives (including RWAs) made on-chain collateral quality and yield-bearing primitives more relevant to mainstream allocators.
- Exchange risk remains a non-zero variable: regardless of which venues are popular, custody concentration is a structural vulnerability that sophisticated users increasingly price in.
The takeaway: even if CEXs remain important gateways, the center of gravity has moved toward on-chain coordination—because that’s where composability compounds.
Practical guidance: how to pursue on-chain efficiency without taking reckless risk
If Kyle’s thesis is correct, more users will trade on-chain more often. Here’s a pragmatic checklist that aligns with that trend:
Keep long-term custody separate from active trading
Use a “vault” posture for long-term holdings and a smaller “hot” posture for daily activity. The point is to limit blast radius.
Be deliberate with approvals
Token approvals are a common failure mode. Revoke what you don’t need, and avoid unlimited approvals unless you understand the trade-off.
Prefer transparent, battle-tested primitives
Route through reputable venues and aggregators, and sanity-check pricing, liquidity, and slippage. If a yield or price is “too good,” assume hidden risk.
Treat signatures like production permissions
The most expensive mistake in DeFi is not usually a bad trade—it’s signing something you didn’t understand.
For consumer-facing security baselines, CISA’s guidance on cybersecurity is not crypto-specific but provides useful mental models; for crypto crime and threat trends, see the Chainalysis Crypto Crime Report.
Where OneKey fits in an “on-chain efficiency” future
If on-chain markets really outcompete CEXs on efficiency, then self-custody becomes the default trading posture, not an edge case. That raises the bar for key management: faster execution should not require weaker security.
A hardware wallet like OneKey can help by keeping private keys offline while still letting you interact with DeFi and on-chain trading workflows. The practical advantage is simple: as you increase on-chain activity, you can reduce exchange custody exposure without turning security into a daily tax—especially when you separate long-term storage from more active wallets and adopt clearer signing habits.
Closing thought
Kyle Samani’s return-to-form message isn’t “DEXs will win because decentralization is morally superior.” It’s more direct: DEXs can win because the best possible exchange is a piece of software that settles transparently, atomically, and composably—without forcing users to take counterparty risk.
The remaining work is engineering and product: better execution, better market structure, better wallet UX, and safer defaults. If those pieces keep converging, the most efficient exchange won’t be a website—it’ll be an on-chain system.
