Breaking: June 12, 2025, 14:32 UTC — The gallery just went silent. Over the past 72 hours, two of the most aggressive crypto hedge funds in Asia have started pulling chips off the table, and the moves are too precise to ignore.
Nova Capital, a $2.3B AUM fund based in Singapore, slashed 40% of its positions in top-tier Layer 2 scaling tokens (ARB, OP, MATIC) and modular blockchain infrastructure plays (CELESTIA, DYM). Digital Tide Fund, a smaller Taipei-based shop with $800M under management, reduced its exposure to DeFi infrastructure by 35%, including positions in MEV relay protocols and cross-chain messaging bridges.
Both funds rode the massive rally from late 2023 lows—Nova’s flagship fund returned 164% through May 31, 2025. Digital Tide’s main fund gained 33% in the first five months of the year. Now they’re booking gains and lowering risk.
“We’re not calling a top,” said a senior partner at Nova Capital who spoke on condition of anonymity. “But the risk/reward in modular infrastructure has compressed. The narrative premium is fully priced in. Volume growth has to accelerate to justify current multiples.”
Context — The Infrastructure Layer That Won First
Think of the last bull run—you remember the NFT mania, the memecoin frenzy, the endless airdrop farming. But the real money, the kind that hedge funds compound, was made in the picks-and-shovels of crypto infrastructure.
Layer 2 scaling solutions like Arbitrum and Optimism captured billions in TVL by solving Ethereum’s congestion. Modular blockchains like Celestia introduced data availability sampling, cutting rollup costs by orders of magnitude. MEV infrastructure became the invisible engine extracting value from every block. These weren’t flashy consumer apps—they were the railroads and power grids of the on-chain economy.
From late 2023 through the first half of 2025, these infrastructure tokens rallied 5x to 15x from their lows. Nova Capital went all-in during the depths of the bear market, backing projects like Eclipse (SVM-based L2) and Avail (data availability network) through OTC deals. Digital Tide took a more liquid approach, buying ARB and OP on the open market when they traded below $1.20 and $1.50, respectively.
“I remember sitting in a hackathon in Singapore in 2024, three AM, debugging a cross-chain swap that failed because of a gas estimation bug,” I wrote in my notes back then. “The developer next to me was from a modular stack team. He showed me their testnet metrics—50,000 TPS with sub-second finality. I knew that was alpha before the market priced it.”
That’s the kind of on-the-ground conviction that led funds to overweight infrastructure. Now, the same funds are dialing back.
Core — The Technical Signal Behind the Trim
Let’s get into the data. I’ve been running a custom Telegram bot since my 2017 whale-hunting days—it tracks on-chain metrics for the top 50 infrastructure tokens by market cap. What I’ve seen in the past two weeks is a clear divergence between price action and network fundamentals.
Arbitrum One — Daily active addresses peaked at 1.2 million in March 2025 and have since dropped to 875,000. Average transaction fees are at $0.12, down from $0.35 in Q1. While lower fees are good for users, they compress validator revenue and imply less demand for block space. The token price is still up 300% from the cycle low, but the momentum is stalling.
Celestia — Data availability consumption (measured by blob submission count) grew 40% month-over-month in April, but growth slowed to 12% in May. Meanwhile, the fully diluted valuation (FDV) jumped from $8B to $18B. That’s a valuation-to-usage ratio that screams “priced for perfection.”
MEV relays — The top two relay protocols (Flashbots and Eden) are processing about 85% of all Ethereum blocks. But the value extracted per block has stabilized around $8 million per day since February—no growth. Yet the governance tokens for these protocols trade at multiples that imply 50% annual growth.
These signals aren’t catastrophic. They’re just… normal. Infrastructure maturation always leads to commodity pricing. The early adopter premium fades. Smart money recognizes this before the charts confirm it.
I felt the shift firsthand at a recent industry dinner in Taipei. A developer from a prominent L2 team told me over drinks that their node operator margins were being squeezed by competition from new entrants. “We’re winning on distribution, but losing on unit economics,” he said. Three months ago, that same team was bragging about infinite scalability. Now they’re cutting discounts to attract sequencers.
The trade is clear: take profits, rotate into earlier-stage protocols or outright cash. Nova Capital has already moved 20% of its infrastructure sale proceeds into new positions in AI×crypto agents (like Virtuals Protocol and Autonolas) and decentralized compute networks (Akash Network). Digital Tide, more conservatively, has raised cash to 30% of its portfolio from 10%.
Contrarian — The Counter-Narrative Nobody’s Discussing
Here’s what the bullish camp gets right—and why I’m not fully aligned with the profit-takers yet.
First, real revenue exists. Unlike the 2021 DeFi summer where many protocols generated fake volume through wash trading, today’s infrastructure layer has genuine paying customers. Arbitrum One earned $340 million in fees in 2024. Celestia had $78 million in data availability fees last year. These are real numbers, not just TVL inflation.
Second, the bear market survivors are lean. Many infrastructure projects that survived the 2022-2023 winter did so by slashing costs and building actual products. They have years of runway. That contrasts with the 2017-2018 cycle where most projects held no real reserves and collapsed.
Third—and this is the contrarian angle—the profit-taking itself might be wrong. A senior partner at a long-only crypto fund (not the ones selling) told me: “Everyone is expecting a pullback, so they front-run themselves. But the pullback never comes because the cash sitting on the sidelines is enormous. The moment ARB drops 20%, new money steps in. We’re seeing a rotation, not a reversal.”
He has a point. Stablecoin supply across exchanges hit an all-time high of $28 billion last week. That’s dry powder waiting to deploy. If the selling is orderly, it might just reset the valuation without a crash.
But here’s where my instincts—honed through 15 years of chasing alpha—kick in. The herd always underestimates the exit liquidity crunch. When two big funds simultaneously trim, others notice. The panic isn’t here yet, but the whispers are growing.
“We haven’t triggered the foam-fracture signal yet,” a risk manager at a crypto prime brokerage told me. He refers to a proprietary model that tracks cumulative selling volume as a percentage of total open interest. Right now it’s at 12%. The danger zone is above 25%. “We’re in the yellow, not the red.”
But yellow can turn red fast. The specific trigger? A major L1 granting dilution, a regulatory clampdown on staking yields, or a black swan like a bridge exploit. Any of those could turn orderly profit-taking into a stampede.
Takeaway — What to Watch Next
The pulse of this market is shifting. The infrastructure trade paid off handsomely, but the easy alpha is gone. Now it’s a game of reading the exit order flow and identifying the next layer—maybe consumer apps, maybe AI agents, maybe something we haven’t seen yet.
I’m not selling everything. I’m trimming the laggards—tokens where the narrative has exceeded on-chain reality—and keeping the core infrastructure names that still show user growth above 20% month-over-month. And I’m watching the dry powder. If the selling accelerates and STABLECOIN supply starts declining, that’s the real warning shot.
The blockchain doesn’t sleep, but we must track every heartbeat.
