On any given day in mid-2026, the combined on-chain fees of Berachain, Celestia, Scroll, Eclipse, Sonic, and Manta amount to roughly $360. That’s less than the cost of a single hour of a senior developer’s time. These six projects collectively raised over $500 million from top-tier venture capital firms. The math is brutal: $500M in, $360 per day out. That’s a payback period of over 3,800 years. This is not a market correction. This is a structural autopsy of a funding model that mistook capital for traction.
Between 2021 and 2024, the crypto narrative shifted from “Ethereum killer” to “modular blockchain” to “ZK-rollup” to “SVM on Ethereum.” VCs poured billions into teams promising scalability, data availability, and novel consensus mechanisms. Berachain introduced Proof of Liquidity. Celestia pioneered a standalone data availability layer. Scroll, Eclipse, and Manta built zkEVM and SVM rollups. Sonic (formerly Fantom) upgraded its EVM-compatible L1. Each raised nine-figure rounds at valuations soaring into the billions. The thesis: build superior infrastructure, and users will come. But the execution revealed a fatal disconnect between technical delivery and product-market fit. All six have mainnets live. All have working code. None have meaningful demand.
The core of this failure lies across four dimensions: technical overdelivery without adoption, tokenomic design that incentivizes extraction, team exodus, and ecosystem emptiness.
Technical Overdelivery, Underadoption
The technical achievements are real. Celestia’s data availability sampling works. Scroll’s zk-prover generates valid proofs. Berachain’s PoL mechanism incentivizes liquidity without MEV extraction. Yet these innovations solve problems that don’t yet exist at scale. The total daily transactions across all six chains is negligible. Scroll sees $24 in daily fees. Manta’s TVL collapsed from $650 million to $4 million after its gamified airdrop—a 99% loss of capital. Eclipse has a mere $1.15 million TVL. Sonic, once a DeFi darling under Fantom, now sits at $16 million TVL—a fraction of its peak. These chains are ghost towns.
"NFTs are art until you inspect the metadata hash." The same applies to these chains: their technical specs look beautiful on paper, but the metadata of on-chain activity reveals empty blocks. During my forensic audit of the Terra Luna collapse, I saw a similar pattern—fragile peg mechanisms masked by high-yield promises. Here, the fragility is even simpler: no users. No amount of modular architecture or zero-knowledge proofs can compensate for zero demand.
Tokenomic Death Spiral
Each project launched a governance/utility token as the primary value-capture vehicle. But without fees, tokens have no fundamental backing. Inflationary supply models—most have no hard cap—dilute holders continuously. Linear unlocks from investors add constant sell pressure. The result: every token has dropped approximately 98% from all-time highs. Berachain’s BERA, Celestia’s TIA, Scroll’s SCR, Eclipse’s ES, Sonic’s S, Manta’s MANTA—all suffered the same fate. The math is simple: if a chain does $60 in daily fees and has a fully diluted valuation of $1 billion, it takes over 45,000 years to earn back that valuation. This is not an investment; it’s a donation.
The airdrop model made it worse. Manta and Scroll artificially inflated TVL through point farming, attracting mercenary capital. Once rewards ended, liquidity vanished. "Your whitepaper is fiction; the contract is fact." The contract here is the on-chain reality of zero fees. I’ve seen this before—during my analysis of the bZx flash loan exploit, I traced how oracle manipulation turned “decentralized” protocols into centralized honeypots. Here, the honeypot is the token itself: a one-way ticket to zero for retail buyers.
Team Exodus and Failed Execution
Execution is not just about code; it’s about continuous development, community building, and adapting to market needs. Eclipse’s core team pivoted to build an “AI agent marketplace,” effectively abandoning the original L2 vision. Sonic’s Andre Cronje, its driving force for years, departed—leaving a chain that relied on his aura. Berachain exposed trust vulnerabilities when it paused its network after a Balancer-based hack, undermining the very principle of immutability. Teams that raised hundreds of millions now maintain skeletal operations. Based on my audit of similar projects during the Terra collapse, I can attest that a team that loses conviction is the fastest path to chain death.
Ecosystem Emptiness
A blockchain without applications is a protocol without purpose. These chains were built for DeFi, NFTs, gaming. Yet their total TVL across all six is likely under $100 million—and that’s mostly from liquidity incentives paying themselves. Real users are absent. Daily active addresses probably number in the dozens. External developers see empty ecosystems and choose to deploy on Ethereum, Solana, or Arbitrum instead. The network effect never materialized. "Flash loans don’t crash markets; flaws do." The flaw here is the assumption that building tech automatically attracts users. It doesn’t.
Contrarian Angle
To be fair, the bulls got some things right. Celestia’s modular DA thesis is sound in theory; if rollups proliferate, demand for cheap DA will grow. Berachain’s PoL mechanism could genuinely reduce MEV and align incentives. Scroll’s zkEVM is technically robust. The problem is timing. These projects launched into a bear market where user activity contracted across the board. But that’s exactly when fundamentals matter most. A chain that can’t demonstrate organic growth during a bull cycle didn’t deserve its valuation in the first place. Moreover, some of the raised capital went to legitimate engineering: Scroll’s provers, Celestia’s light clients, Berachain’s native DEX. But engineering without product-market fit is academic research, not a business.
The contrarian view might argue that residual value exists: Celestia could be acquired by a larger ecosystem; Berachain’s community might reboot. But the data says otherwise. $360 in daily fees is not a base for revival. It’s a signal that the next cycle will ignore these chains entirely. The blockchain industry moves in waves, and these projects were the foam—noticeable, but gone with the tide.
Takeaway
The $500 million raised by these six projects was not an investment in sustainable infrastructure. It was a liquidity event for founders and early backers. The technical architectures are now museum pieces—interesting, but empty. The industry must ask itself: what is the point of a blockchain that no one uses? The answer is that it exists only to transfer wealth from late-stage VCs and retail to early insiders. "Code eats hype for breakfast." But hype ate code for dinner, and the bill is $500 million.
The next generation of L1s and L2s will be judged not by their valuations or fundraising prowess, but by a single metric: daily fees. If you can’t generate a dollar in revenue, you don’t deserve a billion-dollar valuation. The lesson from these six projects is brutally clear—infrastructure is not a product. Users are. And until the industry prioritizes real demand over speculative supply, these ghost chains will remain cautionary tales for those who confuse capital with creation.
