• 08 Dec, 2025 *

Or, What Farcaster teaches about crypto’s broken promises

I joined Farcaster three years ago, excited to be an early part of a community that believed in decentralization, censorship resistance, and distributed ownership.

The premise was intriguing:

• Farcaster (the protocol) would enable users to own their social graphs and facilitate permissionless development
• Farcaster (the application / ‘client’) provided a Twitter-like interface to ease the transition away from ‘Web 2.0’ social media

Today, Farcaster has acquired a token launchpad and positions itself as a wallet. The team behind it is chasing the latest meta of ‘social trading’ in an attempt to find product-market fit, and a path to grow into its $1B valuation.

Farcaster’s arc provides a lens to take stock of crypto—where it’s been, where it is, and why I’ve grown disillusioned with the space.

From fat protocols to aggregation theory

Nine years ago, Joel Monegro advanced his fat protocol thesis, arguing that blockchain inverts the traditional economics of internet-enabled businesses.

Rather than a ‘thin’ protocol layer and ‘fat’ applications built on top of it, a shared data layer (‘state’) would reduce application moats while network tokens (e.g., ETH) enabled direct value capture at the protocol layer.

In practice, application moats continue to exist and token mechanics haven’t automatically translated protocol usage into value capture at the protocol level.

Ben Thompson’s aggregation theory states that in digital markets, value accrues to the owner of the user relationship. The pattern is:

1. Platform owns demand (user relationships)
2. Platform aggregates supply (commoditizes providers)
3. Supply competes for access to demand
4. Platform captures value

In crypto, the same dynamic is emerging:

• Protocols are supply (infrastructure anyone can use)
• Applications own user relationships
• Supply is commoditizing, application layer captures value

The fat protocol thesis assumed tokens would prevent this—that protocol-level value capture would resist aggregation.

But crypto ran into a problem: the only way to capture net new users is to abstract away the crypto.

As a result, fewer new users know which protocol (‘chain’) they’re touching. The token accrues fees, the application accrues users.

Over time, users matter more.

The aggregator stack

Consider the stack Coinbase is building. Base is their L2 chain—Ethereum security with Coinbase-controlled sequencing. On top of it, application-specific protocols provide functionality:

• Farcaster: Social graph and identity
• Zora: Content tokenization
• XMTP: Messaging
• ENS: Naming and identity

These protocols are open, composable, available to any integrator.

That openness is the problem.

Base App assembles them into a unified experience: trading, payments, social, content, messaging, mini-apps. Users see ‘Base.’ They don’t see Farcaster, Zora, or XMTP. The protocols provide functionality; the aggregator owns the relationship.

The same dynamic that made Google the value layer for websites and Facebook the value layer for content is making Coinbase the value layer for crypto social. Supply (protocols) competes for access to demand (users). Demand captures value.

Farcaster built a protocol. The question is whether that matters.

The decentralization paradox

Crypto’s founding ethos emphasized:

• Distributed ownership: No single entity controls the network
• Censorship resistance: Transactions cannot be blocked or reversed by authorities
• Permissionless access: Anyone can participate without gatekeepers
• User sovereignty: Individuals control their own data and assets

These properties require genuine decentralization. Not as an aesthetic preference, but as a technical requirement.

However, across the stack we see increasing centralization.

• Infrastructure concentration: nodes on AWS and Google Cloud, RPC traffic via Infura and Alchemy, validator share via Lido and Coinbase staking
• Token ownership: VC-backed projects launch with heavy insider allocations
• Governance: Foundation-controlled multisigs, upgrade authority residing with core teams

Centralization creates concentrated value capture, vectors for censorship, gatekeeping, and single points of failure.

The uncomfortable truth is that much of what calls itself ‘decentralized’ is a distributed system with centralized control points. The architecture enables the same capture dynamics as traditional tech.

Farcaster as a case study

The thesis

Merkle Manufactory ‘launched’ Farcaster in 2020, betting that a decentralized social protocol could succeed where centralized platforms had captured value and compromised users.

Own your social graph. Censorship resistance. Permissionless development. User sovereignty.

This attracted ideologically motivated users willing to tolerate UX friction to build a new network.

The curated era (2022-23)

@dwr ran an invite-only onboarding process, focused on ‘quality’ daily active users (qDAU): founders, developers, thoughtful ‘builders’.

The ethos was explicitly ‘computer not casino’ in Chris Dixon parlance.

There was an implicit decision embedded in the curation: own the existing crypto user base, don’t pursue adjacent audiences. The network grew dense with crypto-native users but never built a bridge to the speculators of Crypto Twitter, let alone mainstream adoption.

The community was high-quality and engaged, but it was small (~1,000-2,000 DAU).

The Warps experiment

In October 2023, Farcaster (then operating its app under the brand ‘Warpcast’) launched Warps—an in-app currency.

The design choices were revealing:

• Off chain: Not a token, no blockchain settlement
• Non-tradable: Couldn’t be swapped or sold
• Centrally administered: Controlled by Merkle
• Anti-speculation: Structured like a stablecoin

Warps could be earned through engagement or purchased directly for ~$0.01 each. Uses included paying mint fees—the flow was extremely smooth—gifting to other users, and creating channels. (I created /beer, /grappling, and /puns).

The rationale was partly practical: Apple’s App Store rules constrained in-app crypto transactions.

But the architecture went beyond App Store compliance. Warps represented a closed-garden approach that contradicted Farcaster’s ‘sufficiently decentralized’ narrative.

Warps weren’t fungible. As I wrote at the time, ‘They are Chuck E. Cheese tokens for nerds using Warpcast.’

A protocol promising user sovereignty implemented a ‘currency’ where the operator controlled everything—issuance, utility, transferability.

As memecoins gained traction in the Solana ecosystem and Rainbow released points rewards, two bottom-up, community-launched tokens emerged from the /degen channel: $POINTS and $DEGEN.

Matthias has written the story on $POINTS.

At the time, I thought $POINTS was brilliant. As a token on Ethereum mainnet, $POINTS was a jailbroken warp: a community token that could be used across clients built on Farcaster (the protocol) and with the Ethereum economy writ large. Its dramatic demise was disappointing.

But $DEGEN took the baton and was a phenomenal lesson in how tokenomics could drive engagement for Farcaster. Jacek architected a series of airdrops based on participation in the ecosystem, ownership of community NFTs, tipping mechanisms that distributed $DEGEN to users.

It created an explosion of engagement, including commerce in the real economy. I was the first buyer of a DEGEN sweatshirt using $DEGEN. People sold an incredible amount of Girl Scout Cookies for a local troop.

The airdrop was a wealth creation event for active members of the community.

If Merkle wanted to lean into the ‘casino,’ they had a grassroots, community-generated opportunity. The team could have harnessed the energy, dropped a protocol token that would have captured value for Farcaster itself while rewarding the early users who had built the culture.

They chose not to do so.

A ‘steel man’ of the team’s position would say $DEGEN embodied unsustainable speculative energy that would attract mercenary users and repel the builders the team stated they preferred. The token’s subsequent collapse—down 90%+ from ATH—arguably vindicates this view.

But the ‘steel man’ misses the point. The team didn’t reject casino energy on principle; they rejected community-generated casino energy, then pivoted to corporate-controlled casino energy eighteen months later.

They didn’t protect the ‘computer’ ethos.

They ensured that when the casino came, they and their investors would own the value, and early users wouldn’t benefit from it.

Frames

As $DEGEN was driving engagement, Merkle released Frames—interactive mini-apps embedded in posts—in January 2024.

They were slick and opened an interesting design space for developers. They’ve only gotten better.

On the back of a spike in user growth, Merkle raised a $150M round at a $1B valuation.

The team attributed growth to Frames, but $DEGEN was almost certainly the primary driver. The attribution to product innovation created a disconnect between the investor narrative and the user reality.

This mattered because it shaped what came next. If Frames drove growth, double down on developer tooling. If $DEGEN drove growth, embrace tokens.

Merkle advanced the former interpretation, raised a war chest, and watched DAU collapse when the $DEGEN energy faded.

The collapse and pivot

In the space of three months, Farcaster DAU cratered from ~70,000 (July 2024) to ~40,000 (October). (I’m using a fork of @pixelhack’s Dune dashboard to run my own query).

Farcaster became an increasingly noxious and antagonistic place to spend time. The U.S. election contributed to it, surely, but the tone from the top was increasingly patronizing of users and permeated with gaslighting.

DAU continued a secular decline to ~25,000 in July 2025.

I left for an extended period at the beginning of the year, returning recently to check out a viral NFT collection (warplets).

The mood music has changed and the scene is full casino.

No longer positioning itself as ‘decentralized Twitter’, Farcaster is now a ‘wallet with social layer’ across Base, Monad, and Solana. (>90% of activity is on Base).

In October, it acquired a token launchpad (Clanker) that generates c. 400,000inweeklyfees,comparedtoFarcasterweeklyrevenueofwhat, 10,000?

The contrast is stark. The acquisition of Clanker appears to be a functional decision to secure a revenue model.

But the pivot embraces everything the 2022-era Farcaster rejected. Tokens. Trading. Casino energy.

But it does so while abandoning the community that would have made it work.

The structural problems

The problems aren’t independent failures—they’re a chain that leads from governance to commoditization.

Governance. Farcaster isn’t permissionless. Snapchain runs on 11 validators elected semi-annually by the team. This isn’t decentralization with training wheels; it’s a permissioned consortium that calls itself a protocol. The architecture enables everything that follows: if the team controls consensus, they control the network.

Control enables capture. With governance centralized, the team chose who would benefit from the network’s growth. Standard practice—Uniswap, Optimism, Arbitrum, Blur, even DEGEN—is to airdrop tokens to early users who provided the network effects that created value. Farcaster’s early users, who built the culture and justified a $1B valuation, receive nothing.

Presumably, the allocation is reserved for insiders and investors. The community contributed attention and content; the equity accrues to those who contributed capital. This is the economics of Web 2.0 with a decentralization narrative layered on top.

Capture creates dependency. Farcaster raised $150M from Paradigm, a16z, Haun, USV, Variant, et al.

But VC rounds typically involve milestone-based tranches. With DAU down 75% from peak and engagement down only, any performance gates are likely unmet. The Clanker acquisition—buying $400k/week in revenue rather than building it—suggests a team managing toward metrics they can no longer grow organically.

What happens when the next tranche doesn’t arrive? The protocol becomes more dependent on its largest distribution partner.

Dependency completes commoditization. Inside Base App, Farcaster provides the social graph layer. Users see ‘Base’, not ‘Farcaster’. Coinbase owns the user relationship; the protocol becomes infrastructure.

This is the aggregator endgame. Farcaster’s openness—the same property that enabled integration—enables replacement. If Farcaster becomes a liability (regulatory scrutiny, technical debt, community toxicity), Coinbase can migrate to alternatives. The protocol persists as a feature until it doesn’t.

The fat protocol thesis promised that owning the data layer would capture value. Farcaster owns a data layer but the value appears to accrue to Coinbase.

The wallet pivot is an attempt to escape this trap by becoming an aggregator themselves: own the user relationship directly, not just the infrastructure beneath it.

But this requires competing against better-capitalized wallet players (Phantom, Rainbow, Coinbase Wallet) with a social graph that’s largely dormant and a revenue model dependent on the memecoin meta continuing.

It’s a bet that aggregation can happen at the wallet layer, and that Farcaster can win that fight.

The fat protocol thesis didn’t fail because protocols can’t capture value. It failed because Farcaster didn’t lean into tokenization when the door opened, and they built a protocol when they needed to build a product.

By the time they pivoted, the users who would have made it work were making for the exit.

What Farcaster teaches

Farcaster is a case study in the collision between protocol idealism and aggregator economics. But it’s also one about misreading your own growth and missing your market timing.

The team built a curated community, watched that community generate organic token energy, dismissed it, attributed growth to their own product work, raised $150M on that story, and then pivoted to tokens anyway—but only after alienating the users who would have made a token strategy succeed.

The fat protocol thesis suggested owning the protocol layer would capture value. Farcaster built a protocol. But users are increasingly likely to interact through intermediaries that obscure it.

Meanwhile, Bluesky proved the demand existed. Bluesky captured 30 million users with $23M in funding while Farcaster stalled at 25,000 DAU with $180M. The demand for decentralized social was real; Farcaster just wasn’t building for the people who wanted it.

The aggregator’s victory

The last three years of Farcaster offer a verdict on the fat protocol thesis.

The theory was that value would accrue to the shared data layer. The reality, demonstrated by the divergence between Farcaster (the protocol) and Farcaster 2.0 (the wallet), is that value accrues to the owner of the user relationship.

The vision of decentralization remains illusory. There are distributed databases, but we access them through centralized gateways that dictate the economics. The infrastructure is permissionless but the user experience is curated.

This leaves the original ethos—censorship resistance and user sovereignty—in a precarious spot. They are no longer the default settings of ‘Web3’, but niche features for those willing to bypass the aggregators.

The hope for crypto lies with developers who are indifferent to valuations and allergic to VC term sheets.

There are pockets, but hope is thin on the ground.

As the Silicon Valley bros say: exit.

_

Previous