@FinanceYF5: Neoclouds that only rent GPUs end up competing purely on price — and that's a race to the bottom on margins. Renting GPUs is no longer scarce. Shoe company Allbirds, telecoms, former bitcoin miners, AI labs, and even video sites are all offering compute now. The Neoclouds that are truly ahead are competing for a new edge: raising the switching cost for customers through software.
Summary
Neoclouds are moving beyond pure GPU rental and building moats by acquiring software companies, because renting compute alone is no longer scarce and inevitably leads to price wars. Nscale, Nebius, IREN, CoreWeave, Crusoe and others are all making acquisitions to boost customer stickiness and product differentiation.
View Cached Full Text
Cached at: 10/20/26, 05:29 AM
Neoclouds that only rent out GPUs will ultimately have to rely on price competition, and this is a race that continuously squeezes profits.
Renting out GPUs is no longer scarce. Footwear company Allbirds, telecom companies, former Bitcoin miners, AI labs, and even video sites are starting to offer computing power.
The truly leading Neoclouds are competing for new advantages: increasing the cost for customers to switch cloud providers through software.
- Nscale acquires Anyscale, allowing large AI tasks to be distributed across multiple servers
- Nebius acquires Eigen AI and Tavily, making models run faster and cheaper, and providing web search for AI Agents
- IREN acquires Mirantis, turning GPU racks into clouds that enterprises can use directly
- CoreWeave acquires Weights & Biases, helping AI teams track model training
- Crusoe acquires Atero, making model response times faster
The best software is hard to develop in-house in a short time, so these Neoclouds choose to buy it directly.
Source:
Search is becoming infrastructure for Agents.
This time, DeepSeek has opened up its self-developed Search & Fetch API, priced at just $0.75 per 1,000 calls, and it is currently still free to use in Claude Code, Codex, and DeepSeek Harness.
For developers building Agents, this is definitely worth trying out.
Similar Articles
@FinanceYF5: All paths ultimately lead to Neocloud. It is no longer just "renting out GPUs"; it also encompasses starting from energ…
The article discusses the evolution of Neocloud from simple GPU renting to encompassing data centers, computing power aggregation, and AI inference services due to rising demands in AI companies.
@FinanceYF5: Nebius hits $3B annualized revenue — is it a bubble? Explained in 8 minutes Just three months ago, Nebius's annualized revenue was $1.9 billion. Now it's $3 billion. Yet just last week, peer CoreWeave plunged 15% in a single day. The GPU cloud isn't the next AWS — it's the landlord business of the AI era: rent is real, but the buildings are built with borrowed money. Those multi-hundred-billion-dollar contracts, backlogs, and secured power deals are mostly leases and land, not revenue. https://t.co/WqM9RV1xya
The article analyzes Nebius's annualized revenue growth from $1.9 billion to $3 billion, contrasting it with CoreWeave's sharp stock decline. It argues that the GPU cloud business is essentially a "landlord" play in the AI era — the rent is real, but the expansion relies on borrowing and long-term leases rather than current income.
@FinanceYF5: The Neocloud boom is, at its core, an Anthropic boom. Of the $157 billion in publicly priced deals signed by Neoclouds this year, Anthropic accounted for $90 billion, or 57%. Anthropic’s revenue grew sevenfold in a year, while…
The neocloud boom is essentially driven by Anthropic: of the $157 billion in orders at publicly listed prices signed this year, Anthropic accounts for $90 billion (57%), mitigating the risk of compute shortages through a multi-chip strategy spanning Trainium, TPU, NVIDIA, and AMD.
@FinanceYF5: Why do GPU rental prices keep rising? One key reason is that Token demand is exploding. Since November last year, OpenRouter's Token usage has grown by about 27x: Approximately 6 trillion per week → 165 trillion. Among these, the Tokens processed by open-weight models...
GPU rental prices keep rising, primarily driven by exploding Token demand: since November last year, OpenRouter's Token usage has grown by approximately 27x (from about 6 trillion per week to 165 trillion), with the share processed by open-weight models rising from around 28% to 72%.
@Saccc_c: Universities, enterprises, and AI professional users can now own dedicated GPU servers at a lower barrier? And you can earn income without having to operate it yourself? There are currently two main ways to purchase AI compute: the first is renting from cloud vendors, with obvious drawbacks—prices are dictated by GPU supply and demand and are very unstable; the second is buying a whole machine, but...
B3 Labs has released B3IQ, allowing universities, enterprises, and AI professional users to purchase and host NVIDIA GPU servers in installments with a 30% down payment. When idle, the computing power can be rented out for income, which can be used to offset the purchase price or as direct earnings. It has already been used by Stanford, NYU, and many other universities.