$360.
That is the combined daily fee revenue of six blockchain projects. Berachain. Celestia. Scroll. Eclipse. Sonic. Manta.
Not per chain. Total. All six together generate less revenue than a single Uniswap pool on Arbitrum during a sleepy Sunday afternoon.
These projects raised over $5 billion from top-tier venture capital firms. Their tokens are down 98% on average. Their total value locked (TVL) has collapsed by 75–99% after airdrop campaigns ended. Their core developers have either left, pivoted to AI, or stopped blogging altogether.
This is not a crash. This is a complete evaporation of value.
I have spent the last eight years auditing smart contracts and analyzing protocol economics. I have seen failed launches before. But this collection of failures—funded by the same VCs who now quietly write off their positions—represents something more systemic. It is the corpse of the infrastructure narrative, lying in plain sight.
Inheritance is a feature until it becomes a trap. Execution is final; intention is merely metadata.
Let us walk through the graveyard.
Context: The Hype Machine
Between 2021 and 2024, the crypto market was flooded with capital. Interest rates were near zero. Every VC fund needed to deploy into the next paradigm shift: scalable Layer 1s, modular data availability layers,zkEVMs, SVM-based rollups. The pitch was always the same: “We will onboard the next million users.”
The six projects examined here are representative, not exhaustive. Berachain promised a novel consensus mechanism called “Proof of Liquidity” (PoL). Celestia offered modular data availability (DA) decoupled from execution. Scroll built a zkEVM aimed at full Ethereum equivalence. Eclipse claimed to be “Solana on Ethereum” using the Solana Virtual Machine (SVM). Sonic (formerly Fantom) rebranded and attempted a DAG-based L1 revival with Andre Cronje at the helm. Manta Network aimed to be a universal ZK layer for privacy and compliance.
Each raised hundreds of millions. Each launched a mainnet. Each boasted impressive testnet numbers before mainnet—often inflated by sybil attackers and airdrop farmers.
Then the music stopped.
Core: Forensic Dissection of the Six Chains
1. Berachain
Berachain raised $142 million across two rounds, with a public sale that valued the project at over $1.5 billion. Its core innovation—Proof of Liquidity—was supposed to align validator incentives with liquidity provision. Validators stake BERA, but also must bond liquidity tokens from partner protocols. In theory, this creates a symbiotic flywheel.
In practice, the chain launched in early 2025. Within three months, BERA had dropped 98% from its peak of ~$15 to $0.30. The TVL? Not published, but daily fee revenue sits at an infinitesimal fraction of costs. The chain experienced a network pause after being impacted by a Balancer hack. Brevan Howard, a $40 billion macro hedge fund, participated in the B round and secured a one-year, no-risk refund clause—meaning they could pull their capital out if the project underperformed. Retail had no such option.
“Chain business growth engine” was the tagline. No business ever materialized.
2. Celestia
Celestia raised $55 million and pioneered the modular blockchain thesis. It launched its mainnet in 2023, offering a data availability layer that rollups could use instead of Ethereum’s expensive calldata. The promise was that Celestia would become the backbone of a new ecosystem of sovereign rollups.
Instead, Celestia’s native token TIA fell from $20 to $0.30—a 98.5% decline. The narrative around modular DA has collapsed. Competitors like Avail and EigenDA emerged. The market realized that crytpo’s TAM for DA layers is far smaller than the hype suggested. Celestia’s blog last active. Daily fee generated by the Celestia chain? Less than $50.
Data availability is necessary. It is also a commodity.
3. Scroll
Scroll raised a total of $80 million across multiple rounds. It is a zkEVM designed to be bytecode-compatible with Ethereum. The team executed well technically—Testnet to mainnet transition was smooth. ZK proofs were verified on Ethereum. Gas costs were competitive.
Then came the airdrop in early 2025. TVL peaked at around $150 million before the snapshot. After the token distribution, TVL collapsed by 75% to under $12 million. Daily fee revenue? $24. Yes, twenty-four dollars.
Scroll’s SCR token has not yet listed or listed at a fraction of the private valuation. The project still has a team, but the user exodus reveals a harsh truth: users came for free tokens, not for the technology. Without incentives, the chain is a ghost town.
4. Eclipse
Eclipse raised $50 million from Placeholder, Hack VC, and others. Its unique selling point: a rollup using the Solana Virtual Machine (SVM) but settling on Ethereum. In theory, this combines Ethereum’s security with Solana’s performance.
In reality, Eclipse TVL peaked at just $115 million post-airdrop and has since dropped to near zero. The team’s last blog post is over a year old. The founder has already pivoted to a new AI project called “The Human API.” Eclipse is effectively abandoned.
5. Sonic (formerly Fantom)
Sonic is the rebranding of Fantom, a DAG-based L1 that once had a peak TVL of $12 billion. After the collapse of the Terra ecosystem and the departure of Andre Cronje in 2022, Fant bled value. The rebrand to Sonic in 2024 and a massive funding injection did not help. Cronje returned briefly, then left again to build “Flying Tulip,” a separate DeFi project. Sonic currently claims $1.6 billion in TVL—but that number is suspect. Daily fees, if any, are not published. The token S has fallen 99% from its all-time high.
Sonic is a cautionary tale of a chain that once had network effects and lost them all. Reputation cannot be rebooted.
6. Manta Network
Manta Network raised $30 million and positioned itself as a ZK generalist, offering both a layer 2 and a set of ZK tools. Its TVL peaked at $650 million during a liquidity mining program. After the program ended, TVL collapsed to $4 million—a 99.4% drop. The Manta token has also crashed 98%.
Manta is the perfect example of a “phantom TVL” project: all activity was farmed. No genuine user demand exists.
Contrarian Angle: The Real Failure Is Not Technology
Read most post-mortems on these projects, and you will hear about “tech not ready” or “competition too fierce.” That is surface-level.
The root cause is structural: VC incentives are misaligned with long-term protocol health.
Consider Berachain’s Brevan Howard clause: a one-year unconditional refund right. That clause exists because sophisticated money knew the project was high-risk, retail was the exit liquidity. The VC could claim a $100 million paper return, then pull out when reality hit. The team, meanwhile, had millions in liquid token allocations, sold gradually, and walked away rich. Who loses? The retail buyer who accumulated tokens after listing.
In my seventeen years analyzing financial systems, I have seen this pattern repeat: early money structures terms that guarantee profit or capital preservation, while late money bears all the downside. Crypto accelerated it with airdrops and farm-to-dump cycles.
Security-first skepticism should have flagged this earlier. A chain that has no daily fee revenue but a multi-billion token market cap is a liability, not an asset. The open-source code may be audited, but the economic model is a Ponzi.
And the tech? Berachain’s PoL requires validators to lock liquidity tokens. That introduces systemic risk: a depegging event in one of the bonded assets could cascade into a chain halt. Eclipse’s SVM on Ethereum sounds neat until you realize the execution environment is different from both ecosystems—too unfamiliar for Solana developers, too alien for Ethereum users.
Standardization advocacy is important, but when no one uses the standard, it becomes a museum piece.
Blind spots everyone missed:
- Airdrop as admission of failure. If the only way to attract liquidity is to give away tokens, the project has no product-market fit. The data shows that airdrop farming accounts for >90% of TVL at peak. Once the incentive ends, TVL drops 80-99%. No organic growth.
- Team velocity matters more than tech. Eclipse stopped blogging. Cronje left Sonic. Scroll’s key developers are silent. A dead team means a dead protocol, regardless of code quality.
- Investor legitimacy does not equal success. Brevan Howard’s presence gave Berachain credibility. It also ensured that sophisticated money could exit first. The “halo effect” of top-tier VCs blinded retail.
- Data availability is a race to zero. Celestia’s thesis assumes a world of many rollups all needing DA. But Ethereum itself is lowering blob costs with EIP-4844. Competitors are giving DA away for free. The unit economics cannot support a $5 billion valuation.
Takeaway: The Infrastructure Narrative Is Buried
Each of these projects had time, money, and talent. They built working mainnets. They onboarded airdrop farmers. And then they died.
The crypto market has moved on. The current narrative is AI agents executing on-chain, real-world asset tokenization, and user-owned social networks. None of these ghost chains will participate in the next cycle.
What should a rational investor do? Monitor daily fee revenue on platforms like DeFiLlama. If a chain with a $500 million valuation generates less than $1,000 per day in fees, it is a zombie. Do not touch it.
And watch for the next wave of VC-funded infrastructure. It will come—perhaps AI-focused L1s or DePIN-specific chains. The pattern will repeat. The only way to avoid being the exit liquidity is to demand proof of product-market fit before investment.
Execution is final. Intention is merely metadata. These six projects executed poorly. Their intention to build “the next internet” was always a sales pitch.
The graveyard is full. Future builders should study these tombstones before writing their own white papers.
