The Rise and Reckoning of the Neoclouds: How Crypto Miners and GPU Specialists Built the Engine of the AI Boom
Main Facts
The explosive growth of artificial intelligence has rewritten the rules of global infrastructure, creating an insatiable appetite for computing power. For the better part of the past decade, renting server capacity meant dealing with the "hyperscalers"—tech behemoths like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. However, the generative AI boom touched off by applications like ChatGPT completely overwhelmed traditional cloud capacity.
Enter the "neoclouds": specialized, hyper-focused infrastructure-as-a-service (IaaS) providers that do essentially one thing—rent out massive clusters of high-end graphics processing units (GPUs) for AI training and inference. Unlike hyperscalers, which balance thousands of services ranging from email hosting to complex database management, neoclouds concentrate their resources entirely on high-performance compute (HPC), advanced networking, and dedicated storage.
Curiously, the roots of this multi-billion-dollar industry run deep through the cryptocurrency sector. CoreWeave, one of the pioneer neoclouds, initially began as an Ethereum mining operation. As the crypto landscape shifted, a broader wave of Bitcoin miners found themselves sitting on a goldmine of a different sort: grid-connected power, land, and permits. These miners have rapidly pivoted their energy-hungry data centers toward hosting AI infrastructure.
Despite securing billions of dollars in multi-year contracts with top-tier AI labs, big tech firms, and startups, the neocloud sector faces a brutal reality check. Driven by heavy capital expenditures, rapid equipment depreciation, and a sudden wave of market skepticism in late 2026, these specialized providers are fighting to prove that their high-stakes business model can survive over the long term.
Chronology: From Blockchain to Silicon Valleys
The evolution of the neocloud ecosystem is tied directly to the cycles of digital assets and the sudden acceleration of artificial intelligence.
- Pre-2023 (The Crypto Foundation): Companies like CoreWeave and various Bitcoin mining enterprises build massive data center footprints in remote areas, acquiring valuable power purchase agreements (PPAs), land, and electrical grid connections. Ethereum and Bitcoin mining consume vast amounts of electricity, forcing these firms to become experts in cheap, high-density energy procurement and thermal management.
- Late 2022 – 2023 (The ChatGPT Catalyst): OpenAI releases ChatGPT, setting off an unprecedented global scramble for Nvidia GPUs. Hyperscalers struggle to keep pace with demand. Neoclouds emerge from obscurity, quickly acquiring hard-to-find silicon and pivoting their crypto-mining infrastructure to host AI workloads.
- 2024 – 2025 (The Golden Era of Expansion): Neoclouds transform from unknown startups into publicly traded giants. Companies like CoreWeave, Nebius, and IREN secure multi-billion-dollar debt and equity financing. Valuations soar, with CoreWeave briefly touching a peak market capitalization of roughly $90 billion in 2025. Major AI labs and even hyperscalers—whose own data centers are overburdened—sign massive contracts with neoclouds to secure guaranteed GPU availability.
- Mid-2026 (The Infrastructure Pivot): According to industry data from Bernstein, Bitcoin miners contract roughly 7 gigawatts of power capacity across 19 major deals worth upwards of $135 billion, partnering with hyperscalers, neoclouds, and chipmakers. Yet, this represents less than a quarter of the massive 30 GW power pipeline miners have under development.
- October 2026 (The Market Reckoning): Pressures mount on the sector. On October 14, 2026, research firm Gartner warns that over $3.5 trillion of AI infrastructure spending could be exposed to impairment by 2030 due to a lack of differentiated business models and severe cost pressures. Credit markets react swiftly: CoreWeave’s unsecured debt yields spike to roughly 15%, bond default insurance costs jump 30% in a month, and stock prices across the sector tumble by 27% to 32% within weeks.
Supporting Data and Financial Realities
The neocloud model operates on the razor’s edge of three scarce resources: cutting-edge silicon, reliable high-density power, and massive amounts of capital.
The Economics of Compute
Neoclouds acquire advanced GPUs—predominantly from Nvidia—often before their competitors can secure allocations. They bundle these chips with high-speed networking and storage fabrics, offering them to clients via multi-year reserved capacity agreements or on-demand hourly rates. Their top-line revenue is driven by the sheer scale of these deployments, but their financial statements reveal intense cost pressures.
The primary vulnerability of the neocloud model is rapid asset depreciation. Because AI hardware evolves at a breakneck pace, expensive GPU clusters age out quickly in terms of efficiency. Financial reports from mid-2026 highlight this heavy burden:
- CoreWeave: Reported second-quarter 2026 revenue of approximately $2.6 billion, but absorbed a staggering $1.4 billion in depreciation and amortization expenses over the same period.
- Nebius: Booked about $260 million in depreciation against $582 million in revenue for the second quarter of 2026.
The Power Pipeline
Power has become the ultimate currency in the artificial intelligence race. Acquiring permits, building sub-stations, and negotiating grid connections can take years. By repurposing legacy Bitcoin mining facilities, neoclouds and converted crypto hosts bypass the most time-consuming bottleneck in data center deployment.
Between 2024 and mid-2026, Bitcoin miners contracted roughly 7 gigawatts of power capacity across 19 separate deals valued at over $135 billion. However, this output accounts for a fraction of the 30 gigawatts sitting in the miners’ broader development pipeline.
Market Valuation Strains
The public markets have expressed growing anxiety regarding the sustainability of neocloud debt loads and capital expenditure requirements. By late October 2026, CoreWeave’s market valuation had dropped from its 2025 peak of $90 billion down to approximately $34 billion, dragged down by a 29% monthly stock decline. Peers experienced similar pullbacks: Nebius fell 27%, and IREN saw its shares drop by 32%.
Official Responses and Industry Perspectives
The rapid rise and subsequent market jitters surrounding neoclouds have sparked intense debate among industry leaders, analysts, and financial institutions.
The Bull Case: Guaranteed Demand and Long-Term Contracts
Defenders of the neocloud model emphasize that demand for raw compute shows no signs of slowing down. Major enterprise clients, sovereign AI initiatives, and foundational model developers require guaranteed access to compute clusters that hyperscalers simply cannot always provide on demand.
Addressing investor concerns, CoreWeave co-founder Brannin McBee noted that nearly all of the company’s projected 2027 capacity is already sold out, with firm contracts extending deep into 2028 and beyond. Proponents argue that even if independent neoclouds face consolidation, the physical data centers, power purchase agreements, and GPU clusters they control represent essential, irreplaceable infrastructure for the future of the global economy.
The Bear Case: Overbuilding and Impairment Risks
Conversely, institutional analysts have raised red flags regarding the long-term viability of the sector. The most damning assessment came in October 2026, when research and advisory firm Gartner issued a sweeping warning. Gartner predicted that "cost pressure and a lack of a differentiating business model" will force more than 80% of independent neocloud providers to cease operating as standalone entities by 2030.
Furthermore, Gartner estimated that over $3.5 trillion in global AI infrastructure spending could be exposed to impairment by the end of the decade if firms continue their current capital expenditure trajectories without establishing sustainable, diversified revenue streams.
Credit markets have echoed these warnings. As bond yields spiked and credit default swap costs climbed, lenders demonstrated that they are no longer willing to finance aggressive expansion without clear paths to profitability independent of speculative AI hype cycles.
Implications: The Future of AI Infrastructure
The trajectory of the neocloud sector carries profound implications for the broader technology and financial landscapes.
1. Consolidation and M&A Activity
With independent neocloud stocks trading well off their all-time highs and debt servicing costs rising, many market watchers anticipate a wave of consolidation. Cash-rich hyperscalers—who already occasionally rent neocloud capacity to smooth out their own supply chain crunches—may find it cheaper to acquire distressed neoclouds outright rather than build out greenfield data centers from scratch.
2. The Divergence Between Centralized Neoclouds and DePIN
While neoclouds operate as conventional, centralized corporations owning physical data centers, an alternative model has emerged in the crypto space: Decentralized Physical Infrastructure Networks (DePIN). Protocols like Render, Akash, and io.net utilize blockchain tokens to coordinate peer-to-peer GPU sharing, rewarding independent hardware owners for contributing idle compute capacity. While neoclouds cater to enterprise-grade, multi-billion-dollar AI labs requiring secure, massive dedicated clusters, DePIN networks offer a flexible, crowdsourced alternative for smaller-scale workloads. Both models are racing to capture market share, though they answer to fundamentally different financial governance structures.
3. Investment Risk and Volatility
For retail and institutional investors, public neocloud stocks—such as CoreWeave (CRWV), Nebius (NBIS), and IREN (IREN)—alongside converted Bitcoin mining hosts like TeraWulf, Cipher, Hut 8, and Core Scientific, represent high-risk, high-beta exposure to the AI hardware trade. As the market digests the tension between multi-year revenue backlogs and heavy depreciation schedules, the coming years will determine whether neoclouds establish themselves as permanent pillars of modern computing or serve as the high-cost scaffolding of the early AI gold rush.
