Community Trust ScoreLikely Real
Neocloud stocks got hammered this month. CoreWeave’s unsecured debt was yielding around 15% by Oct. 21, and its shares had already shed 29% in October alone — a brutal stretch that’s put the whole sector under a microscope.
Why It Matters
The move highlights how quickly the market is reassessing the risk profile of GPU-rental businesses as investors rotate from growth expectations to balance-sheet and financing concerns. When debt and equity in the same subsector sell off together, it can signal that higher funding costs and tighter sentiment may start to matter as much as demand for AI infrastructure itself.
Nebius dropped 27% over the same period. IREN fell 32%. All three trade on the Nasdaq, and all three are in the same basic business: renting high-end graphics processing units — GPUs — to companies running artificial intelligence workloads. The selloff hasn’t been random. Investors are staring hard at balance sheets, debt loads, and whether these companies can actually earn enough from expensive hardware to cover what they owe. The cost of insuring CoreWeave’s bonds against default rose 30% in a single month. That’s not a subtle signal.
The math is pretty unforgiving.
What Neoclouds Actually Sell
Strip away the jargon and the business is straightforward: neoclouds own GPU clusters and rent them out. Customers — AI labs, big tech companies, startups — can lock in capacity through multi-year deals or just pay by the hour. The provider collects revenue from that capacity while carrying the cost of buying the chips, powering the facilities, and financing everything.
Unlike hyperscalers such as Amazon Web Services, Microsoft Azure, and Google Cloud, neoclouds don’t spread their resources across dozens of different services. They focus almost entirely on GPU-heavy AI work — training and inference. Training builds the model. Inference is what happens when a model answers a question or generates an image. Hyperscalers do all of this too, but they’re also running e-commerce platforms, enterprise software, streaming services. Neoclouds aren’t. That specialization is the pitch. It’s also the risk.
Three things make or break the model: chips, power, and capital. GPUs are expensive and hard to get. Power is scarce and increasingly fought over. And capital — debt, equity, whatever — has to be raised continuously to fund infrastructure that depreciates fast.
CoreWeave’s numbers drive that home. The company reported nearly $1.4 billion in depreciation and amortization in the second quarter of 2026, against revenue of about $2.6 billion. Nebius recorded roughly $260 million in depreciation and amortization on $582 million in revenue over the same quarter. That’s a lot of value walking out the door every quarter, regardless of how full the data centers are.
CoreWeave’s Crypto Origins and the Miner Pivot
CoreWeave didn’t start in AI. The company launched in 2017 mining Ethereum with GPUs. When Ethereum moved to proof of stake in 2022, GPU mining on the network ended overnight. CoreWeave had hardware, it had infrastructure, and it needed a new use for both. AI and cloud computing was the answer.
Bitcoin miners are now making a nearly identical move, just a few years later. In the two years through July 2026, miners contracted about 7 gigawatts of power capacity across 19 deals worth more than $135 billion with hyperscalers, neoclouds, and chipmakers. That’s less than a quarter of the 30 gigawatts sitting in miners’ planned power pipeline — meaning there’s probably more of this coming.
Power is the prized asset.
Mining companies already hold land, grid connections, and permits. Those things can take years to secure. Converting a mining site into an AI data center is hard, but it’s a lot faster than building from scratch. TeraWulf signed a 20-year lease with Anthropic for its Kentucky campus. The deal is expected to generate about $19 billion — more than TeraWulf’s entire market value at the time it was signed.
It’s worth being clear about what neoclouds are not. They’re not crypto projects. They don’t issue tokens. Decentralized physical infrastructure networks — Render, Akash, io.net — also offer GPU computing, but those run on tokens to pay providers and reward participants. Render’s RENDER token is one way to get exposure to GPU computing demand, though the model is fundamentally different from a company leasing out data center capacity it owns outright.
Gartner’s Warning and What Comes Next
On Oct. 14, Gartner dropped a forecast that didn’t help sentiment. The research firm said cost pressure and a lack of differentiated business models would push more than 80% of neocloud providers to stop operating independently by 2030. Gartner also put a number on the downside: more than $3.5 trillion in AI infrastructure spending could face impairment by the end of 2030 if providers kept spending at their current pace.
That’s a big number. Unclear yet whether the market had fully priced it in before October’s slide, but the timing wasn’t great.
Demand is still real, though. CoreWeave co-founder Brannin McBee said nearly all of the company’s 2027 capacity had already been sold, with contracts running into 2028 and beyond. That’s not nothing. But bookings don’t automatically mean profits. The question — still wide open — is whether those commitments generate enough cash to cover infrastructure costs and service the debt.
CoreWeave’s value sat at roughly $34 billion after the October declines, down from a peak of around $90 billion in 2025. TeraWulf, Cipher, Hut 8, and Core Scientific are among the listed Bitcoin miners that have moved into AI hosting alongside the pure-play neoclouds.
CoreWeave’s unsecured debt: 15% yield, Oct. 21, 2026.
Hub: RENDER price, news, and analysis
Frequently Asked Questions
Why are neocloud stocks falling so hard in October 2026?
CoreWeave shares dropped 29% in October, Nebius fell 27%, and IREN declined 32%, as investors focused on high infrastructure costs and debt loads. CoreWeave’s unsecured debt was yielding about 15% by Oct. 21, and the cost of insuring its bonds against default rose 30% in a month.
What did Gartner forecast for neocloud providers?
On Oct. 14, 2026, Gartner forecast that more than 80% of neocloud providers would stop operating independently by 2030, and that more than $3.5 trillion in AI infrastructure spending could face impairment by end of 2030.
