What Is a Neocloud: The AI Cloud Model Explained

Discover our guide to the neocloud sector, how big the AI cloud market is, how the contract model works, and where Aethir fits.

Community
  |  
What Is a Neocloud: The AI Cloud Model Explained

Key Takeaways

  1. Neocloud meaning: A neocloud is a cloud built only for AI, selling GPU capacity rather than a catalog of general services. Gartner defines it as a provider built specifically for AI and high-performance workloads.

  2. A fast-growing AI cloud market: Neocloud revenue passed $25 billion in 2025, and SRG expects it to approach $400 billion by 2031.

  3. Contracts carry the neocloud model: Most neocloud capacity is sold on multi-year contracts that get signed before the GPUs arrive. Axe Compute signed $3.45 billion of contracts in seven months.

  4. Aethir runs both supply lines: Aethir pools more than 430,000 GPU containers, run by independent operators in 94 countries, into one decentralized GPU cloud. It also plans to build contracted clusters through Aethir ACCELERATE, partnering with companies like Axe Compute.

Neocloud Meaning: A Cloud Built Only for AI

The neocloud meaning analysts settled on in 2026 is narrow and useful. In a June 2026 forecast, Gartner defines neocloud providers as clouds built specifically for AI and high-performance workloads and expects them to capture 20% of a $267 billion AI cloud market by 2030. A hyperscaler sells hundreds of services. A neocloud sells GPU time, networking, and the storage that keeps those GPUs busy. The category grew out of the 2023 GPU shortage, and Aethir covered that rise of the first neoclouds as it happened.

Why Bare Metal GPU Access Defines the Category

Bare metal GPU servers: Most neoclouds hand customers whole machines with little or no virtualization layer in between. Cisco describes the typical neocloud build as bare metal or lightly virtualized nodes with 4 to 8 GPUs each. That gives training and inference jobs direct access to the hardware.

A second network for GPUs: A neocloud runs a lossless back-end fabric at 400G or 800G so GPUs can talk to each other without the general network in the way. Large training runs depend on it.

Several ways to buy: Capacity comes reserved, on demand, or as a serverless endpoint. The reserved route drives most revenue, and that shapes how every neocloud model is financed.

How Big the AI Cloud Market for Neoclouds Is

Neocloud revenue passed $25 billion in 2025, with $9 billion booked in the fourth quarter alone, up 223% on the year, according to SRG (Synergy Research Group). SRG expects a 58% compound growth rate, bringing the segment close to $400 billion by 2031. 

Gartner lands near $53 billion for 2030. These estimates differ so much because the neocloud sector is relatively new, and analytics firms measure the neocloud boundary differently.

fdgdfg.png

ABI Research expects inference to make up 80% of the neocloud market by 2030, with GPU-as-a-service revenue reaching $250 billion. Training built the first neoclouds. Serving models will pay for the next ones.

Neocloud providers win on GPU availability and price per hour. Hyperscalers keep the managed services, compliance tooling, and procurement contracts most enterprises already run on.

The Neocloud Model Runs on Signed Contracts

The neocloud model sells capacity before it exists. CoreWeave, the largest listed neocloud reported a revenue backlog of about $104 billion at the end of June 2026, against 1.5 GW of active power and roughly 3.7 GW contracted. 

What Axe Compute Shows About the Neocloud Model

Axe Compute, listed on Nasdaq as AGPU, reported $3.45 billion in contracts signed over seven months, with terms of up to 60 months. It has received more than $184 million in customer prepayments since its Q2 call.

In October 2026, Axe Compute took full ownership of its Georgia AI cluster of 2,304 NVIDIA B300 GPUs. The customer contract now runs through 2031, and Axe expects about $364.6 million in revenue over its life.

Axe Compute has created a repeatable deployment model that includes a committed customer, covered power, and owned GPUs. That’s the neocloud model in one cluster, and Axe intends to repeat the structure at future sites.

Where Aethir’s Decentralized GPU Cloud Fits Among Neocloud Providers

Aethir’s decentralized GPU cloud pools GPUs that independent operators already run. According to Messari data, the wider DePIN sector reached about $10 billion in size in 2025, and compute is one of its working categories.

33243.png

The network: Aethir aggregates enterprise GPU capacity from independent operators, called Cloud Hosts, into one orchestrated pool. It spans more than 430,000 GPU containers across 94 countries and 200+ locations.

The contracted line: Aethir ACCELERATE plans to bring dedicated NVIDIA B300 and GB300 clusters of 64 to 256 nodes online across the US and Europe. It targets up to $700 million in contracts by the end of 2026.

Partners who build: Axe Compute is an infrastructure partner in the program through Axe Build. That pairs a working neocloud operator with a GPU cloud that already serves customers in 94 countries.

A neocloud is a GPU-first cloud built for AI demand, and its business runs on contracts signed before the hardware lands. The model works, as the Axe Compute ramp shows. Aethir brings a decentralized GPU cloud to the same market and adds contracted clusters on top, so buyers get flexible capacity and dedicated capacity from one provider.

Frequently Asked Questions

What is a neocloud in simple terms?

A neocloud is a cloud provider built only for AI workloads, selling GPU capacity, fast networking, and storage instead of general cloud services. 

How do neocloud providers differ from hyperscalers?

Neocloud providers specialize in GPUs, usually offer bare metal GPU servers, and often price below hyperscalers. Hyperscalers offer wider managed services and compliance tooling. Many enterprises use both.

How big is the AI cloud market for neoclouds?

SRG puts 2025 neocloud revenue above $25 billion and forecasts nearly $400 billion by 2031. Gartner expects neoclouds to hold 20% of a $267 billion AI cloud market in 2030.

Is Aethir a neocloud or a decentralized GPU cloud?

Both labels fit. Aethir runs a decentralized GPU cloud of 430,000+ GPU containers and, through Aethir ACCELERATE, plans to add contracted DePIN data center capacity built with partners like Axe Compute.

Disclosure

Aethir Foundation is an Axe Compute shareholder, through its 2025 treasury transaction. We cover Axe as an interested holder. This article reflects Aethir's views and is not investment advice. For official company information, see Axe Compute's SEC filings (CIK 0001446159) and investors.axecompute.com.

Nothing in this article should be relied upon as a guarantee of future performance or results. 

Keep Reading