Hook
The code says switching cost, but the liquidity says exit friction. CLSA’s latest deep-dive on six SaaS giants—Salesforce, ServiceNow, Oracle, Microsoft, Workday, Adobe—calls their moats "still strong" against the AI wave. The report is a punchy rebuttal to the ‘vibe-coding’ narrative. But here’s the part the analysts didn’t articulate: the exact same structural defensiveness lives in DeFi’s core protocols. Uniswap’s liquidity depth, Maker’s collateral engine, Aave’s interest rate models—they all exhibit the same ‘organizational embedding’ CLSA glorifies. Only difference: crypto’s moats are transparent on chain, not hidden in a blog post.
Context
CLSA’s thesis rests on three pillars: deep process integration (workflows, compliance), switching costs (data, configuration, training), and ecosystem lock-in (app stores, partner networks). They rate Microsoft and Adobe as Outperform because AI monetization is visible (Copilot, generative fill), while ServiceNow and Workday are Underperform because their AI story is newer and TAM expansion is less obvious. The report is a masterclass in ‘counter-hype’ reasoning—reminding investors that incumbents don’t die just because a shiny agent pops up.
But the report is silent on blockchain-native protocols. That’s a blind spot. Because if you take CLSA’s framework and apply it to Aave, Uniswap, or even Ethereum itself, the moat metrics are not just comparable—they are harder to breach.
Core: DeFi’s Moat is Code-Deep and Liquidity-Wide
Let’s start with process integration. CLSA claims Salesforce’s sales association is sticky because it’s embedded in how a company manages leads. Replace ‘leads’ with ‘liquidity positions’ and you get Uniswap. Traders don’t leave a pool with $500M TVL for a new AMM that has $5M, even if the new one has zero slippage. The act of moving liquidity is itself a transaction that incurs price impact, signaling costs, and reputation risk. Uniswap’s ‘workflow’ is the swap route; the ‘compliance’ is the immutable smart contract rules. No amount of AI can rewrite that without forking the chain, which fractures liquidity.
Switching costs in crypto are often dismissed as low because ‘code is forkable.’ That’s a surface-level take. Forking a protocol doesn’t fork its user base, its integrated front-ends, its oracles, or its governance reputation. Look at SushiSwap’s attempted vampire attack on Uniswap—it worked temporarily because of incentive alignment (SUSHI rewards), but the moment rewards dried up, capital flowed back. The real switching cost is not technical; it’s network density. Uniswap’s 10+ billion in cumulative volume across thousands of pairs creates a data network effect that a new fork cannot replicate overnight. CLSA calls this ‘data moat’; on chain we call it ‘liquidity depth as a barrier to entry.’
Ecosystem lock-in is even stronger in DeFi. Aave’s aTokens, Compound’s cTokens, and Maker’s DAI are all composable with dozens of protocols. A user holding aDAI can use it as collateral on Curve, deposit into Yearn, or trade on Uniswap. Leaving Aave means breaking all those integrations. That is the equivalent of a Salesforce user losing their AppExchange plugins. CLSA would call this ‘platform ecosystem.’ In crypto, it’s ‘money legos.’ And it’s more resilient because the composability is permissionless—no single entity can turn it off.
Contrarian: The Fragility that CLSA Overlooks (and DeFi Embodies)
CLSA’s blind spot is assuming that moats are static. They are not. The same data network effect that protects Uniswap also makes it vulnerable to a sudden compression of liquidity during a bear market. When TVL drops 60%, the moat shrinks proportionally. That’s not true for Salesforce—even if a customer reduces seats, the data remains. In crypto, capital can leave overnight. The moat is a river, not a pond.
Furthermore, CLSA’s report highlights the importance of “compliance as a moat.” DeFi protocols have no compliance layer—that’s both a weakness and a strength. It means they can’t be regulated out of existence, but also that they lack the bureaucratic friction that locks in enterprise customers. For retail users, switching costs are lower because there’s no onboarding, no contract negotiation. The barrier is purely economic (gas fees, slippage, risk perception). That’s a more elegant moat, but also one that can be arbitraged away by a more capital-efficient competitor.
Takeaway
CLSA is right: SaaS incumbents have deep moats that AI alone won’t topple. But they should look at DeFi. The same framework—process integration, switching costs, ecosystem lock-in—applies with even greater transparency. The difference? In SaaS, the moat is built by time and sales. In crypto, it’s built by liquidity and code. Both are hard to cross, but only one is open for anyone to verify.
The question isn’t whether AI kills the moat. It’s whether the moat can survive a 70% drawdown in capital. Code doesn’t care about your lock-in. Volatility is just interest for the impatient.