# What Is a Neocloud? Definition, Examples, and What It Means for Compliance

Source: https://truvocyber.com/blog/what-is-a-neocloud
Published: 2026-10-18
Topic cluster: Cybersecurity

> What is a neocloud? A GPU-focused cloud for AI workloads. Definition, neocloud vs hyperscaler, provider examples, and what a neocloud changes for SOC 2.

A neocloud is a cloud provider built around renting GPU compute for AI training and inference, often on bare metal servers, with far fewer managed services than AWS, Microsoft Azure or Google Cloud. CoreWeave, Lambda, Crusoe and Nebius are well-known examples. For SOC 2, the neocloud's own report covers less of the stack, so more controls and evidence fall to the customer.

At a glance

- **Neocloud meaning**: A cloud provider specializing in GPU-as-a-service for AI and high-performance computing.
- **Market size**: JLL counts about 190 neocloud operators.
- **Differences from a hyperscaler**: GPU focus, bare metal servers, 400G/800G GPU networking, few managed services, and GPU clusters reserved for one to three years.
- **Who buys**: AI model builders, enterprises running AI workloads, and the hyperscalers themselves.
- **What changes for SOC 2**: The neocloud becomes a subservice organization in the customer's SOC 2 report, and the customer performs the complementary user entity controls (CUECs) the neocloud's report assigns to it.

## What Is a Neocloud?

A neocloud sells GPU compute plus the networking and storage that keep the GPUs busy. Cisco's definition: *Neocloud providers offer specialized, high-performance infrastructure designed to power AI workloads. They combine public cloud elasticity with dedicated GPU acceleration.*

The business model is GPU-as-a-service (GPUaaS): customers rent GPU servers by the hour or reserve whole clusters for a fixed term. A hyperscaler sells hundreds of services, and GPUs are one of them. In an August 2025 report, [JLL](https://www.jll.com/en-ca/insights/the-rise-of-neocloud-in-the-ai-landscape) counted approximately 190 neocloud operators offering GPUaaS.

## Neocloud vs Hyperscaler: How Do They Differ?

The table summarizes the differences [Cisco](https://www.cisco.com/site/us/en/learn/topics/computing/what-is-neocloud.html) describes between a neocloud and a hyperscaler, with the compliance consequence of each.
AreaHyperscaler (AWS, Azure, Google Cloud)NeocloudWhat it means for complianceFocusGeneral-purpose cloud; GPUs are one product among manyBuilt for GPU-as-a-serviceThe vendor file centers on one high-impact production vendorServer modelMostly virtual machines and managed servicesBare metal accessOn bare metal servers, the customer often patches and hardens the operating systemNetworkingGeneral-purpose virtual networks400G/800G networking with RDMA for GPU-to-GPU trafficSegmentation and firewall rules are often more manualManaged servicesExtensive: databases, analytics, native logging and security toolsInfrastructure-level AI services, few application servicesThe customer often builds logging, monitoring and key managementBuying modelOn-demand, fixed-term commitments and pay-per-token platformsOn-demand, plus dedicated GPU clusters reserved for a fixed term, typically one to three yearsMulti-year contracts need security terms at signing

[CoreWeave Kubernetes Service](https://docs.coreweave.com/products/cks/clusters/introduction) runs Kubernetes directly on bare metal GPU nodes *without any intermediary software, such as a hypervisor or virtual machine*. That gives the GPUs full performance and puts more of the system inside the customer's control boundary.

Key insight

**A neocloud trades managed services for GPU performance, and with them some of the compliance evidence those services produced.**

On a hyperscaler, native audit logging and API-readable configuration produce much of the evidence a SOC 2 audit samples. On a neocloud, the customer produces it with its own tools.

## Which Companies Are Neocloud Providers?

JLL's 2025 report names CoreWeave, Lambda, Crusoe and Nebius among the major neocloud providers. The table lists the attestations each published on its own site as of October 2026; confirm the current report before relying on it.
Neocloud providerPublished security attestationsWhere to request reportsCoreWeaveSOC 2 Type II, ISO 27001, ISO 27017 and ISO 27018[CoreWeave trust portal](https://trust.coreweave.com/), reports on requestLambdaSOC 2 Type II; ISO 27001, 27017, 27701 and 22301[Lambda trust page](https://lambda.ai/trust)CrusoeSOC 2 Type II, announced July 2024 after a Type I in December 2023[Crusoe announcement](https://www.crusoe.ai/resources/blog/crusoe-cloud-achieves-soc-2-type-ii), trust.crusoe.aiNebiusSOC 2 Type II including HIPAA, ISO 27001, ISO 22301 and others, covering Nebius AI Cloud, AI Studio and TractoAI[Nebius announcement](https://nebius.com/blog/posts/soc-2-type-ii-hipaa-iso-27001-enterprise-security-standards)

Some neoclouds rent space in colocation data centers. CoreWeave's [security page](https://coreweave.com/security) states that it requires its data center colocation providers to adhere to SOC 2 and ISO 27001 for the length of their contracts.

## Who Buys From Neocloud Providers?

Cisco lists neocloud users as *enterprises, model builders, and hyperscalers*, and JLL describes the relationship with hyperscalers as complementary.

CoreWeave's IPO filing showed Microsoft accounted for 62 percent of its 2024 revenue, and two customers made up 77 percent, as [The Register](https://www.theregister.com/2025/03/04/coreweave_ipo/) reported.

We also often see B2B SaaS and AI companies that keep their application and customer data on a hyperscaler and train or serve models on a neocloud. Their enterprise customers then ask about the GPU nodes that hold prompts, fine-tuning data or model weights.

## What Does a Neocloud Mean for SOC 2 and Compliance?

SOC 2 treats a neocloud the same way as a hyperscaler or colocation facility; the responsibility line moves, so the customer operates and evidences more controls. The [guide to SOC 2 on a neocloud](https://truvocyber.com/blog/soc-2-neocloud) covers the control-by-control detail, GRC platform evidence gaps and mid-audit migrations.

### How the Service Model Sets the Shared Responsibility Line on a Neocloud

CoreWeave's [shared responsibility model](https://coreweave.com/security) is one example. It assigns the physical, network, compute (including the operating system) and platform parts of the stack to CoreWeave, and the application, access, data and customer-data parts to the customer. Across neoclouds, the line typically follows the service model, and the provider's SOC 2 report and contract set the exact split:

- **Managed Kubernetes or virtual machines:** the neocloud operates the facility, hardware and host; the customer operates workloads, access, data and logging.
- **Bare metal:** the neocloud operates the facility and hardware; the customer takes on the operating system, patching, hardening, segmentation and logging.
- **Firmware and out-of-band management (BMC, IPMI):** often shared, so settle the split in the contract.

### What a Neocloud's SOC 2 Report Covers

A neocloud's SOC 2 report covers the services and facilities the neocloud operates, for the products, regions and period named in the report. It gives no assurance on the customer's own controls. Three checks decide whether the report helps the customer's audit:

1. **Report type.** A SOC 2 Type II report tests controls over a period; a Type I tests design at a point in time.
2. **Products and regions in scope.** The GPU product and region the customer buys need to appear in the system description.
3. **Period.** The report period should overlap the customer's audit period, with a bridge letter for any gap.

Check the scope statement on ISO 27001 certificates the same way. The [neocloud SOC 2 guide](https://truvocyber.com/blog/soc-2-neocloud) has a full due diligence checklist.

### How a Neocloud Appears in a Customer's SOC 2 Report: Subservice Organization and Carve-Out

In the customer's SOC 2 report, a neocloud is a subservice organization: a separate company whose services are part of the customer's system. Under the [AICPA description criteria](https://www.aicpa-cima.com/resources/download/get-description-criteria-for-your-organizations-soc-2-r-report), the system description names the neocloud, what it does for the customer, and how its controls are treated.

Customers of large neoclouds typically use the carve-out method. The customer's auditor leaves the neocloud's controls out of testing, and the report lists the complementary subservice organization controls the customer relies on the neocloud to perform, such as physical security and hardware disposal. The customer's annual review of the neocloud's report is the evidence those controls operate. The inclusive method has the customer's auditor test the neocloud's controls as well, which needs the neocloud to take part in the customer's audit; we rarely see it with large providers.

A neocloud that rents colocation space typically carves out that provider in its own report, adding a third report to the chain. The [guide to SOC 2 vendor management for data centers](https://truvocyber.com/blog/soc2-vendor-management-datacenter) covers how to review each report in that chain.

### Which CUECs a Neocloud Leaves to the Customer

Complementary user entity controls (CUECs) are the controls a neocloud's SOC 2 report assumes its customers perform. The list tends to be longer than on a hyperscaler. The customer side of CoreWeave's shared responsibility model shows the ground CUECs cover:

- Identity and access management policies, network policies and firewall rules
- Data classification, encryption, backups and disaster recovery plans
- Application code, reliability and security
- Collection, protection and use of customer data, including privacy requirements

The binding list is in each provider's report. The [guide to complementary user entity controls in SOC 2](https://truvocyber.com/blog/soc-2-complementary-user-entity-controls) covers mapping each CUEC to a control and its evidence.

### How SOC 2 Scoping Works With a Hyperscaler and a Neocloud Together

A split setup is one system with two subservice organizations, and the boundary follows customer data. If customer prompts, datasets or model weights trained on customer data reach the GPU nodes, those nodes are in scope. The [SOC 2 scoping guide for hybrid and on-prem infrastructure](https://truvocyber.com/blog/soc-2-hybrid-on-prem-scoping) covers drawing one boundary across providers.

## A Neocloud Is a GPU Cloud With a Shorter Compliance Handoff

Companies that read the neocloud's SOC 2 report, map its CUECs and plan evidence for the controls that move to them keep their own SOC 2 report, and the enterprise deals behind it, on track. The [SOC 2 for bare metal and on-prem infrastructure](https://truvocyber.com/services/soc-2-bare-metal) page lists the controls that change when the customer operates the host.

The [guide to SOC 2 consultants for GPU clouds and on-prem AI infrastructure](/blog/soc-2-consultants-gpu-cloud-on-prem) compares four types of consultant and the skills to check before hiring.

## Related Guides

- [SOC 2 for Chip Startups: Design IP, EDA Environments, Foundries and Compute Services](https://truvocyber.com/blog/soc-2-semiconductor-startups)
- [SOC 2, ISO 27001 and HIPAA for Infrastructure Providers: One Program, Three Outputs](https://truvocyber.com/blog/soc-2-iso-27001-hipaa-infrastructure-providers)
- [On-Prem SOC 2 Cost and Timeline: What Changes When You Run Your Own Hardware](https://truvocyber.com/blog/soc-2-cost-timeline-on-prem)

### Moving workloads to a neocloud?

Truvo scopes SOC 2 across hyperscalers, neoclouds and your own hardware, then collects the evidence platforms can't reach.[Book Your Strategy Call](/contact-us)

## Frequently Asked Questions

### What is a neocloud in simple terms?

A neocloud is a cloud provider, such as CoreWeave, Lambda, Crusoe or Nebius, that rents GPU computing power for AI training and inference, with far fewer managed services than AWS, Azure or Google Cloud.

### What is the difference between a neocloud and a hyperscaler?

A hyperscaler sells hundreds of general-purpose services, with GPUs as one product. A neocloud is built for GPU-as-a-service, often on bare metal servers, and sells dedicated GPU clusters on terms of typically one to three years. Neocloud customers usually operate more of their own stack.

### Are neocloud providers SOC 2 compliant?

CoreWeave, Lambda, Crusoe and Nebius each publish a SOC 2 Type II attestation. Before relying on one in the company's own audit, check that the report names the product, region and period in use.

### Does a neocloud's SOC 2 report cover my company?

A neocloud's SOC 2 report covers only the services the neocloud operates. The company needs a SOC 2 report on its own system, in which the neocloud appears as a subservice organization, usually under the carve-out method, and the company performs the CUECs the neocloud's report lists.

### Can I use a neocloud and stay SOC 2 compliant?

Yes. Add the neocloud to the system description and vendor register, review its SOC 2 report each year, map its CUECs to the company's controls, and produce evidence for the host, network and logging controls the provider assigns to the customer.

## Sources

- [Cisco, What is a neocloud?](https://www.cisco.com/site/us/en/learn/topics/computing/what-is-neocloud.html)
- [JLL, The rise of neocloud in the AI landscape (August 2025)](https://www.jll.com/en-ca/insights/the-rise-of-neocloud-in-the-ai-landscape)
- [The Register, CoreWeave IPO filing (March 2025)](https://www.theregister.com/2025/03/04/coreweave_ipo/)
- [CoreWeave, CoreWeave Kubernetes Service clusters introduction](https://docs.coreweave.com/products/cks/clusters/introduction)
- [CoreWeave, Security and shared responsibility model](https://coreweave.com/security)
- [CoreWeave, Trust portal](https://trust.coreweave.com/)
- [Lambda, Trust page](https://lambda.ai/trust)
- [Crusoe, Crusoe Cloud achieves SOC 2 Type II (July 2024)](https://www.crusoe.ai/resources/blog/crusoe-cloud-achieves-soc-2-type-ii)
- [Nebius, SOC 2 Type II, HIPAA and ISO 27001 announcement (October 2025)](https://nebius.com/blog/posts/soc-2-type-ii-hipaa-iso-27001-enterprise-security-standards)
- [AICPA, Description Criteria for a SOC 2 report (DC 200)](https://www.aicpa-cima.com/resources/download/get-description-criteria-for-your-organizations-soc-2-r-report)
- [AICPA, 2017 Trust Services Criteria (With Revised Points of Focus, 2022)](https://www.aicpa-cima.com/resources/download/2017-trust-services-criteria-with-revised-points-of-focus-2022)

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is a neocloud in simple terms?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "A neocloud is a cloud provider, such as CoreWeave, Lambda, Crusoe or Nebius, that rents GPU computing power for AI training and inference, with far fewer managed services than AWS, Azure or Google Cloud."
      }
    },
    {
      "@type": "Question",
      "name": "What is the difference between a neocloud and a hyperscaler?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "A hyperscaler sells hundreds of general-purpose services, with GPUs as one product. A neocloud is built for GPU-as-a-service, often on bare metal servers, and sells dedicated GPU clusters on terms of typically one to three years. Neocloud customers usually operate more of their own stack."
      }
    },
    {
      "@type": "Question",
      "name": "Are neocloud providers SOC 2 compliant?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "CoreWeave, Lambda, Crusoe and Nebius each publish a SOC 2 Type II attestation. Before relying on one in the company's own audit, check that the report names the product, region and period in use."
      }
    },
    {
      "@type": "Question",
      "name": "Does a neocloud's SOC 2 report cover my company?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "A neocloud's SOC 2 report covers only the services the neocloud operates. The company needs a SOC 2 report on its own system, in which the neocloud appears as a subservice organization, usually under the carve-out method, and the company performs the CUECs the neocloud's report lists."
      }
    },
    {
      "@type": "Question",
      "name": "Can I use a neocloud and stay SOC 2 compliant?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. Add the neocloud to the system description and vendor register, review its SOC 2 report each year, map its CUECs to the company's controls, and produce evidence for the host, network and logging controls the provider assigns to the customer."
      }
    }
  ]
}
