The numbers are intoxicating: Nebius Q2 revenue surged 454% to $582.3 million, adjusted EBITDA turned positive at $236.2 million, and the stock jumped 34% in a single day. IREN rose over 5%, Coreweave over 3%. The Neocloud sector—a cluster of AI cloud infrastructure providers, many born from Bitcoin mining—is the market’s darling. But as someone who spent years auditing smart contract logic for integer overflows in Lagos, I see a familiar pattern: the rush to build cathedrals on sand.
Context: The Architecture of Control
Neocloud companies like IREN (Iris Energy), Nebius, and Coreweave are not decentralized protocols. They are centralized corporations listed on the Nasdaq, selling GPU compute power to AI startups and enterprises. Their technical edge lies in repurposing Bitcoin mining infrastructure—low-cost power, land, and fiber—for high-performance computing. IREN, for instance, runs both ASIC miners for Bitcoin and NVIDIA GPUs for AI. Nebius, spun off from Yandex’s European assets, markets itself as an “AI-native cloud.” Coreweave is a pure-play GPU cloud, deeply tied to NVIDIA.
This is the blockchain industry’s favorite narrative: miners saving themselves by pivoting to AI. The market applauds. But the governance of these entities is a black box of traditional corporate control—boards, insider stakes, and opaque capital allocation. Trust is not a protocol; it is a promise from a CEO. And promises, as I learned during the 2017 ICO audits, are the first thing to break.
Core: The Code of Capital
Let’s read the financials as code. Nebius’s revenue growth is impressive, but its adjusted net loss of $33.2 million (though narrowing 64%) tells me the unit economics are still under stress. The EBITDA margin of 40.6% is a positive signal, but it masks the capital intensity: to sustain this growth, Nebius must spend billions on new GPUs. That capital comes from debt or equity issuance, which dilutes existing shareholders—like an inflationary token model. The “supply” of GPU compute is expanding, but demand must keep pace or the lease price per GPU crashes. This is the same risk I flagged in early DeFi protocols: the velocity of capital can outrun the value it creates.
IREN’s dual business—Bitcoin mining and AI cloud—is a double-edged sword. In a bull market for both, it gains leveraged upside. But the management team must allocate resources between ASICs and GPUs. Based on my experience auditing smart contract vesting schedules, I see a similar principal-agent problem: the incentives of the leadership (short-term stock price) may not align with the long-term health of the Bitcoin network. IREN is a miner, yes, but it is also a centralized cloud provider. The culture of decentralization does not compile here.
Contrarian: The Silence in the Chain
The market sees Neocloud as a natural extension of crypto’s infrastructure thesis. I see it as a retreat. Bitcoin miners were once the backbone of a permissionless, decentralized network. Now they are becoming aristocrats of compute, serving the same AI giants that are centralizing power. The contrarian angle: the pivot to AI is not a diversification; it is a surrender. By chasing GPU revenue, miners are underinvesting in ASIC hashrate, making the Bitcoin network less secure. Silence in the chain speaks louder than noise in the stock market.
Moreover, the governance of these companies is shockingly opaque. Nebius’s founder, Arkady Volozh, was previously sanctioned by the EU. Coreweave’s secretive debt financing (backed by Magnetar Capital) raises questions about covenant controls. IREN’s early history includes a shell mining company. Vision without verification is just hallucination. The market is pricing in a future where these companies become the AWS of AI, but they are competing against trillion-dollar giants with supply chain advantages. Culture compiles where logic fails—the culture of these companies is not decentralized; it is survivalist.
Takeaway: Building Cathedrals in the Bear Market
The Neocloud boom is a reminder that the blockchain industry’s infrastructure is still built on centralized foundations. The real test will come when the bull market fades and the capital dries up. Will these companies have the governance resilience to survive? Or will they collapse under the weight of their own leverage? I am building governance models for DAOs that can weather storms—but these Neoclouds are not DAOs. They are cathedrals built in a bull market, and the bear will test every stone. The question is not whether they can rise, but whether they can fall without taking the whole ecosystem with them.