We’ve all seen the headlines: a DeFi protocol loses 40% of its liquidity providers in a single week, yet its total value locked barely budges. Or a Layer 2 chain processes millions of transactions daily, but its native token sits flat. These aren’t random market movements. They’re symptoms of a deeper structural problem—one that the AI industry is now exposing with brutal clarity.
When I first read the analysis of Anthropic’s business model, I felt a chill of recognition. The numbers were jarring: over 40% of its annualized recurring revenue (ARR) came from indirect cloud channels like AWS Bedrock, Microsoft Foundry, and Google Cloud. But the profit per dollar from those channels was far lower than from direct sales. The article—which I’ll refer to as the “Anthropic Analysis”—used a seven-dimensional framework to dissect the risks. Its core finding: channel dependency dilutes margins, masks true scale, and creates a “vapour revenue” illusion.
Sound familiar? It should. Because in crypto, we’ve built the same trap for ourselves.
Context: The Channel Economy
Let’s step back. In both AI and crypto, the “channel” is the middleman who controls access to end users. For AI models, it’s the cloud hyperscalers. For crypto protocols, it’s aggregators, wallets, CEX listings, and even Layer 2 settlement layers. The pattern is identical: a project builds a superior product, but to reach scale, it cedes distribution to a third party. The third party charges a fee—sometimes explicit, sometimes hidden in compute costs or lost user data. The project’s revenue grows, but its profits shrink.
Take Uniswap. The protocol itself is a marvel of automated market making. But the vast majority of swaps now flow through aggregators like 1inch, Paraswap, or Matcha. These aggregators route trades, and they charge a fee. Uniswap gets the volume, but the aggregator captures the user relationship. The same dynamic plays out in Layer 2: rollups like Arbitrum and Optimism rely on Ethereum for security and data availability, but they also rely on centralized sequencers—often run by the same teams—to process transactions. The sequencer is a channel. It extracts value in the form of MEV or priority fees.
The Anthropic Analysis drilled into this with a critical lens. It noted that the 650 billion USD ARR figure (which, let’s be honest, is almost certainly a misinterpretation of a long-term goal or a currency error) was suspect. The real ARR might be in the single-digit billions. But the point isn’t the exact number. It’s the structure: when a company gives away 40% of its revenue to partners, its unit economics become fragile. In crypto, we see the same fragility. A protocol might boast $10 billion in TVL, but if 70% of that TVL comes from a single liquidity aggregator like Curve or a controlled multisig, the TVL is not sticky. It’s rented.
Core: The Macro Asset View
From a macro perspective, channel dependency is a liquidity issue. History repeats, but liquidity decides the tempo. During the 2020 DeFi Summer, capital flowed freely into protocols that offered the highest yields, often through distribution partners like Yearn or Harvest. Those protocols looked like winners—until the liquidity rotated. The ones that survived were the ones that built direct user relationships and community trust.
I saw this firsthand during my time managing a digital asset fund in Mexico City. In 2020, I allocated $2 million into Aave and Compound liquidity pools. But instead of chasing the highest APY, I spent time in the community forums, watching user complaints about interface friction. The pools with the smoothest UX—where users could deposit and withdraw without confusion—retained capital even when yields dropped. The channels (aggregators) were important, but they weren’t the source of loyalty. The community was.
Fast forward to today. The same lesson applies to Layer 2 scaling. Post-Dencun, blob data will be saturated within two years. Then rollup gas fees will double. That’s my technical position, based on analyzing on-chain data and the exponential growth of L2 activity. Most projects are ignoring this, focusing instead on getting listed on the next CEX or integrated into the next bridge. They’re building for the channel, not for the user.
Uniswap V4’s hooks turn the DEX into programmable Lego. That’s powerful. But the complexity spike will scare off 90% of developers. The ones who stay will be the ones who understand the architecture—and they’ll be the ones who build directly, not through aggregators. The channel dependency is a feature, not a bug, for the early adopters. But for the masses, it’s a barrier.
Contrarian: The Decoupling Thesis
The conventional wisdom in crypto is that channel dependency is a temporary phase. “We’ll eventually own our distribution,” the VCs say. “We’ll build our own aggregator, our own wallet, our own chain.” The Anthropic Analysis suggests otherwise. Even a company with massive resources and brand recognition—Anthropic has a valuation of $200-300 billion—cannot escape the channel trap. The cloud providers are both partners and competitors. They offer distribution, but they also offer their own models (Google Gemini, Amazon Titan, Microsoft’s OpenAI partnership). The relationship is a knife-edge.
In crypto, we see the same pattern. The biggest channels—centralized exchanges like Binance, Coinbase, and Kraken—are also launching their own DeFi products, their own wallets, their own staking services. They’re not just distributors; they’re competitors. And they hold the keys to the liquidity.
My contrarian angle: The decoupling thesis—that crypto will “decouple” from traditional finance and become independent—is wrong. More likely, it will become even more dependent on channels. Bitcoin’s post-ETF approval reality is a case study. The “peer-to-peer electronic cash” vision is dead. Instead, Bitcoin is now a Wall Street toy, traded on traditional exchanges, held in ETFs, and controlled by fiduciary custodians. The channel won. Satoshi’s original vision was decoupled from the channel. But the market chose the channel.
Culture is the code that compels human adoption. The culture of Bitcoin originally was about self-custody and resistance to gatekeepers. But the ETF approval shifted that culture. Now, the dominant narrative is “digital gold,” not “p2p cash.” The channel (Wall Street) rewrote the code.
Takeaway: Positioning for the Next Cycle
So what do we do? As a fund manager, I’m looking for projects that recognize the channel trap and actively work to bypass it. That means protocols that invest in their own user interfaces, their own mobile apps, their own direct-to-consumer relationships. It means protocols that treat aggregators as a boost, not a crutch.
In the coming sideways market, the chop will separate the wheat from the chaff. Projects that lose 40% of their LPs in a week are not victims of market conditions; they’re victims of channel dependency. The signals are there: watch for protocols that are paying for volume through incentives rather than earning it through loyalty.
Trust takes years to build, seconds to break. The projects that will survive the next half-decade are the ones that build trust directly with their communities, not through intermediaries.
History repeats, but liquidity decides the tempo. The next bull run will reward those who control their own distribution. The channel dependency trap is not a bug. It’s a choice. Choose wisely.