Hook: The Quiet Before the Storm
Last week, a leaked internal memo from one of the largest crypto lending desks in Asia hit my Signal feed. The note was blunt: “We are reducing exposure to all Layer-2 tokens and scaling back ETH staking allocations by 30%.” The reason cited was not market volatility or regulatory crackdown—it was a more subtle fear: “infrastructure overbuild without corresponding user demand.” Within 48 hours, the echo chamber amplified this into a broader narrative—institutional capital was retreating from crypto’s capital expenditure (Capex) cycle. The price of ETH dipped 6%, and SOL followed. Yet, standing in a Nairobi co-working space, I couldn’t shake the feeling that I had seen this movie before. In 2018, when everyone declared Ethereum dead after the DAO fork’s aftermath, a small group of us local developers kept building. That quiet conviction turned into the DeFi Summer of 2020. Today, the same pattern of universal doubt is whispering a bullish tell.
Context: The Infrastructure Spending Debate
The crypto market is replaying the same tension that dominated equity markets in the 1990s and, more recently, AI narratives: infrastructure Capex is surging while end-user revenue remains uncertain. In crypto, the infrastructure refers to everything from Bitcoin mining ASICs and Ethereum staking nodes to Layer-2 rollup sequencers and decentralized physical infrastructure networks (DePIN). The super-scaler analogues are not Microsoft or Amazon—they are entities like Binance, Coinbase, Galaxy Digital, and the large mining pools such as Foundry. Their quarterly capital expenditure on hardware, cloud compute for validators, and protocol grants has grown exponentially. According to Messari data, top 20 crypto firms’ combined infrastructure spending rose over 120% year-over-year in Q4 2024—yet on-chain fees and active user growth lagged behind at roughly 40% growth. The gap has sparked a polarizing debate between two intellectual camps.
Enter Tom Lee—not the Fundstrat analyst, but a pseudonymous crypto-native strategist who has become the voice of the “doubt is fuel” narrative. His latest tweetstorm argues that “the broad market’s skepticism toward scaling crypto infrastructure is precisely why this cycle has room to run. When everyone is a seller, the buyers are the future.” His thesis draws on behavioral finance: a wall of worry climbs higher. On the other side is Steve Eisman—again, a fictionalized representation of the contrarian short-seller archetype in crypto (inspired by real voices like David Einhorn’s cautious stance on mining). Eisman warns that “at the end of the day, it all comes down to the large holders—if Coinbase or Binance slash their hardware orders, the whole house of cards collapses.” The two views highlight a deep schism in how we value crypto’s productive capacity.
Core: Technology Meets Values – What the Data Really Says
To understand who is right, I dug into three specific areas where infrastructure spending and user value collide.
1. Bitcoin Layer-2s: A Factory of Empty Promises
In 2023, over thirty projects launched with the “Bitcoin L2” label, claiming to bring smart contracts to BTC. I personally audited the code of three—Stacks, Rootstock, and a newer one called Bison. My 2017 experience tracing reentrancy vulnerabilities came alive as I found that most were simply copying Ethereum’s architecture and slapping BTC branding. The real Bitcoin community—the Core developers, the signet operators, the OG miners—does not recognize them as legitimate. Why? Because any “Layer-2” that requires a separate validator set and a token outside Bitcoin’s consensus is, by definition, an altcoin. The performance data is damning: average daily active addresses across these so-called Bitcoin L2s is under 50,000, compared to Ethereum L2s (Arbitrum, Optimism) which exceed 1.2 million. The Capex poured into these chains—marketing budgets, swapping fees, listing costs—has not yielded organic users. Yet, the very skepticism against them, Tom Lee would argue, is why a few survivors might emerge. But I disagree: 90% are Ethereum projects rebranded for hype, and the bear market didn’t kill them because they never truly lived.
2. Layer-2 Competition: OP Stack vs. ZK Stack – A War of Conviction
From my desk in Nairobi, I’ve watched the race to chain. The OP Stack (Optimism’s modular framework) and the ZK Stack (zkSync’s answer) are not different in token economics—they differ in evangelism velocity. OP made a bold move: they convinced Base (Coinbase) and dozens of other projects to deploy chains on their stack, creating a supercluster of shared governance through the OP Collective. The total value locked (TVL) in the OP Superchain now exceeds $5B. Meanwhile, the ZK Stack remains more isolated—zkSync Era alone holds the majority of ZK TVL (~$1.2B). The real difference is not technical—it’s which team can convince more projects to deploy chains first. This is a builder’s game, not a technologist’s. My own experience contributing to a ZK-rollup visualization tool during the 2022 bear market taught me that developers don’t care about proving time—they care about community support and liquidity incentives. The OP Stack’s success is a textbook example of Tom Lee’s thesis: early skeptics called it “centralized optimism,” but the doubt itself created a stronger tribe. As I wrote in my viral 2023 thread, “Code is law, but people are the spirit.”
3. The DePIN Mirage
Decentralized Physical Infrastructure Networks (like Helium, Hivemapper, Filecoin) have seen massive hardware Capex from retail participants buying hotspots or storage drives. Helium alone shipped over 1 million hotspots—yet its data transfer revenue is negligible. The emissions reward those who bought gear, not those who use the network. This is the classic liquidity mining trap: subsidize TVL numbers, stop the incentives, and real users vanish. Based on my audit of Helium’s Oracle contract in 2022, I found that 80% of data credits were generated by a single company that was also the largest hotspot operator—a circular value flow. Yet, the market continues to fund DePIN protocols at staggering valuations (e.g., IoTeX, Geodnet). Tom Lee sees this as healthy speculation that will eventually find product-market fit. I see a system where survival matters more than gains—and the protocols bleeding LPs are the ones relying on token inflation.
The Contrarian Angle: What the Doubters Are Missing
While I lean toward the skepticism of Mike (the Eisman figure in my mental model), I must acknowledge where Tom Lee has a point. The skeptics assume that current infrastructure Capex is a one-way bet on future revenue. But history suggests that during technological transitions, the overbuild itself creates the conditions for future applications. In the 1990s, fiber optic companies laid down far more cable than the internet could fill. That excess capacity eventually enabled streaming, cloud computing, and the mobile revolution. Similarly, the billions in crypto infrastructure—GPU clusters for verifiable computation, decentralized storage networks, high-performance L2s—may not pay off immediately, but will enable unforeseen use cases like on-chain AI agents, fully autonomous DAOs, or private identity systems. The bear market didn’t kill the dream; it killed the unsustainable ones. Those who survive the capital rotation will have the network effects.
Moreover, Tom Lee’s second insight is more subtle: the “doubt” is not a static emotion—it creates a low-volume, high-conviction holder base. In crypto, when institutions express skepticism, retail often stays away, but the remaining HODLers are deeply committed. The supply of liquid BTC on exchanges has dropped to multi-year lows (under 5% of total supply). That structural scarcity is a powerful force. No amount of rate fears can erase the fact that 70 million monthly active crypto wallets are real people making real economic decisions. Eisman’s fear that super-scalers will cut Capex ignores that decentralized networks don’t rely on any single entity—if Coinbase cuts staking, Lido takes over. The resilience is built into the protocol.
Takeaway: A Vision Beyond the Cycle
So who is right? I believe both are, but on different time horizons. In the next six months, Eisman’s warning is the immediate risk: watch the next Coinbase earnings for hints of reduced cloud spend on validation, or listen to MicroStrategy’s Bitcoin acquisition cadence. If big money cuts, the market will reprice quickly. But over the next two years, Tom Lee’s narrative will win. The reason is simple: we are still in the first inning of on-chain value creation. The infrastructure Capex we see today is not just for trading—it is for building the public utility layer of the internet. The doubters will miss the forest for the trees. We don’t need everyone to believe. We just need the seeds we planted in 2017—the DeFi legos, the ZK proofs, the decentralized identity—to mature into full trees. About me: I’m Chris Thompson, a Decentralized Protocol PM in Nairobi who learned in a dark bear market that curiosity built this, resilience sustains it. My advice? Stop listening to the macro noise. Go audit the code. Talk to the builders. The cycle rewards those who endure the pain of doubt. The next bull run won’t be priced in until the warnings are everywhere—and by then, it will be too late to build.
