Executive Summary: Nutanix Cloud Platform with Enterprise AI

Nutanix occupies the same structural position as VMware (an HCI / private-cloud platform, an abstraction layer running on commodity x86 it does not manufacture) and against the 4+1 model it produces a strikingly similar profile: strongest at infrastructure orchestration (Layer 2A), competent across the data and runtime layers, and gapped at data pipelines (Layer 1C) and the reasoning plane (Layer 2C). Its deepest IP and authority sit in the orchestration control plane: Prism, AOS, and AHV.

Two divergences lift Nutanix above a simple VMware echo. It owns its storage (Nutanix Unified Storage at 0C) where VMware is bring-your-own-storage, and Data Lens gives 1A compliance governance, though without the classification a catalog would add. And Nutanix Enterprise AI (NAI) is a genuinely good, less-NVIDIA-locked model-serving product: multi-source models (NVIDIA NIM, Hugging Face, or custom upload), a vLLM default engine, a CPU-only inference path, and portability across any CNCF-certified Kubernetes, arguably a cleaner standalone serving product than VMware's Model Runtime.

The capture mechanism here is decoupled and invisible, the reassuring kind. The edges are open: commodity x86 hardware is Retained (swap Dell for HPE or Lenovo without rebuilding), NKP is unforked upstream Kubernetes (manifests lift to any conformant cluster), NAI runs on any cloud's Kubernetes and is not NIM-locked, and the models are open. The buyer feels unconstrained. Yet the value that actually accumulates (AOS/Prism orchestration opinions, Acropolis Dynamic Scheduling placement logic, NAI's serving and governance configuration, Data Lens policy) is captive to Nutanix. Eight of nineteen scored components are Ceded, concentrated exactly where operational opinions accumulate. Open hardware and open models do not make the control plane portable.

Where the stack thins is identical to VMware. Layer 1C (data pipelines) is a gap: Time Machine and cross-cloud snapshots are storage mobility, not AI pipelines: there is no lineage and no cost-aware movement. Layer 2C (the reasoning plane) is a gap: Agent Gateway is a real, generally available agentic-governance control point (RBAC, rate-limiting, audit including MCP-call audit, health-based failover, capacity load-balancing), but governance and operational routing are not placement reasoning. There is no per-request, content/cost/quality model selection, and the data-side feeders a reasoning plane would query (an AI-metadata catalog at Layer 1A, lineage at Layer 1C) are absent, so data-relative placement is structurally out of reach. The forward 'Nutanix Agentic AI' full stack is Early Access (GA targeted second half of 2026) and does not move the cell.

One 0A caveat bounds the on-platform AI story: AHV virtualizes only vGPU-class accelerators (L40S, H100, L4, and a single Blackwell SKU: the RTX PRO 6000 Server Edition), not HGX training systems (B200/GB200), which remain NVIDIA bare-metal. Nutanix's on-platform AI is inference and fine-tune scale, not frontier training. The June 2026 'NVIDIA Certification' headline is a storage certification: Nutanix Unified Storage feeding external GPU servers over Spectrum-X and GPUDirect, not GPUs running on Nutanix.

The buyer's trade is the VMware trade with a Nutanix accent: the lowest-friction on-ramp to private AI for the Nutanix (and VMware-refugee) installed base: same console, same operational model, no new vendor, plus real owned storage and a serving plane that is not captive to NVIDIA, in exchange for ceding the orchestration and serving control plane to Nutanix, and accepting that data pipelines and the reasoning plane remain the enterprise's own responsibility. Nutanix runs the private-AI platform. It does not yet govern the agents on top of it.

Layer-by-layer status: Layer 0A (Hardware-Agnostic Substrate: AHV and Prism on Any OEM's x86, NC2 on Cloud Bare Metal), Layer 0B (Flow Virtual Networking and Flow Network Security Over Any Fabric), Layer 0C (AOS Distributed Storage and Nutanix Unified Storage; General-Purpose, Not Built for AI Throughput), Layer 1A (Data Lens Compliance and Audit Governance; No Content Classification, No Lineage), Layer 1B (Foundational RAG via Managed Postgres), Layer 1C (No Pipeline Layer), Layer 2A (Nutanix Heritage Strength), Layer 2B (Platform-Native Serving, Agents Pre-GA), Layer 2C (Gateway Governance, No Placement), Layer 3 (+1) (No Applications on Nutanix Paper; the AI Partner Program Is Validation).

Assessment framework: 4+1 Layer AI Infrastructure Model. Scoring model: Decision Authority Placement Model (DAPM) — Retained, Delegated, or Ceded. Published by The CTO Advisor LLC (DBA The Advisor Bench). Author: Keith Townsend. Date assessed: October 27, 2026. Version: v1.28 - 4+1 v3: Layer 0 Sub-Layers.

Nutanix Cloud Platform with Enterprise AI

Mapped to the 4+1 Layer AI Infrastructure Model

v1.28 - 4+1 v3: Layer 0 Sub-LayersAssessed October 27, 2026Sources & revision history
ACTIVE ASSESSMENT

Summary Finding

Nutanix occupies the same structural position as VMware (an HCI / private-cloud platform, an abstraction layer running on commodity x86 it does not manufacture) and against the 4+1 model it produces a strikingly similar profile: strongest at infrastructure orchestration (Layer 2A), competent across the data and runtime layers, and gapped at data pipelines (Layer 1C) and the reasoning plane (Layer 2C). Its deepest IP and authority sit in the orchestration control plane: Prism, AOS, and AHV.

Two divergences lift Nutanix above a simple VMware echo. It owns its storage (Nutanix Unified Storage at 0C) where VMware is bring-your-own-storage, and Data Lens gives 1A compliance governance, though without the classification a catalog would add. And Nutanix Enterprise AI (NAI) is a genuinely good, less-NVIDIA-locked model-serving product: multi-source models (NVIDIA NIM, Hugging Face, or custom upload), a vLLM default engine, a CPU-only inference path, and portability across any CNCF-certified Kubernetes, arguably a cleaner standalone serving product than VMware's Model Runtime.

The capture mechanism here is decoupled and invisible, the reassuring kind. The edges are open: commodity x86 hardware is Retained (swap Dell for HPE or Lenovo without rebuilding), NKP is unforked upstream Kubernetes (manifests lift to any conformant cluster), NAI runs on any cloud's Kubernetes and is not NIM-locked, and the models are open. The buyer feels unconstrained. Yet the value that actually accumulates (AOS/Prism orchestration opinions, Acropolis Dynamic Scheduling placement logic, NAI's serving and governance configuration, Data Lens policy) is captive to Nutanix. Eight of nineteen scored components are Ceded, concentrated exactly where operational opinions accumulate. Open hardware and open models do not make the control plane portable.

Where the stack thins is identical to VMware. Layer 1C (data pipelines) is a gap: Time Machine and cross-cloud snapshots are storage mobility, not AI pipelines: there is no lineage and no cost-aware movement. Layer 2C (the reasoning plane) is a gap: Agent Gateway is a real, generally available agentic-governance control point (RBAC, rate-limiting, audit including MCP-call audit, health-based failover, capacity load-balancing), but governance and operational routing are not placement reasoning. There is no per-request, content/cost/quality model selection, and the data-side feeders a reasoning plane would query (an AI-metadata catalog at Layer 1A, lineage at Layer 1C) are absent, so data-relative placement is structurally out of reach. The forward 'Nutanix Agentic AI' full stack is Early Access (GA targeted second half of 2026) and does not move the cell.

One 0A caveat bounds the on-platform AI story: AHV virtualizes only vGPU-class accelerators (L40S, H100, L4, and a single Blackwell SKU: the RTX PRO 6000 Server Edition), not HGX training systems (B200/GB200), which remain NVIDIA bare-metal. Nutanix's on-platform AI is inference and fine-tune scale, not frontier training. The June 2026 'NVIDIA Certification' headline is a storage certification: Nutanix Unified Storage feeding external GPU servers over Spectrum-X and GPUDirect, not GPUs running on Nutanix.

The buyer's trade is the VMware trade with a Nutanix accent: the lowest-friction on-ramp to private AI for the Nutanix (and VMware-refugee) installed base: same console, same operational model, no new vendor, plus real owned storage and a serving plane that is not captive to NVIDIA, in exchange for ceding the orchestration and serving control plane to Nutanix, and accepting that data pipelines and the reasoning plane remain the enterprise's own responsibility. Nutanix runs the private-AI platform. It does not yet govern the agents on top of it.

Strength
Moderate
Gap
Partner
Layer 0A · ComputeComputeHardware-Agnostic Substrate: AHV and Prism on Any OEM's x86, NC2 on Cloud Bare Metaldecides: vendor · Delegated▼

CPUs, accelerators, servers, integrated systems, OS and hypervisor, physical plant

Vendor-Provided

AHV (Hypervisor), with PrismCeded

AHV (KVM-derived) plus the proprietary AOS distributed storage OS, managed in Prism. GPU passthrough and GA NVIDIA vGPU with live migration. AHV/AOS storage policies, VM configuration, and Prism opinions are captive; they cannot be lifted to ESXi or Hyper-V without rebuilding. Proprietary Nutanix platform; opinions captive, no open exit.

Multi-Vendor Hardware SupportRetained

Runs on NX appliances and OEM platforms (Dell, HPE, Lenovo, Cisco UCS, Fujitsu) plus software-only on Supermicro/ProLiant/PowerEdge. Commodity x86 substrate: the enterprise swaps OEM without rebuilding workloads. The same hardware-agnostic position VMware occupies.

NVIDIA GPU Integration (vGPU on AHV)Delegated

GA NVIDIA vGPU with live migration and GPU passthrough; L40S/H100/L4/RTX PRO 6000 Server Edition. Delegated: the enterprise can replace Nutanix's GPU integration with another hypervisor's without rebuilding its GPU opinions. Pressure-tested October 5, 2026 against the Red Hat standard: the vGPU profiles, MIG layout, scheduler policy, and license server are NVIDIA's definitions and lift intact to any of the five hypervisors NVIDIA supports (vSphere, AHV, RHEL KVM, Ubuntu KVM, Citrix). They don't lift off NVIDIA; that capture is NVIDIA's and sits in the NVIDIA column under the owner's-row ruling.

Nutanix Cloud Clusters (NC2) on AWS and Azure Bare MetalCeded

The Nutanix stack (AOS, AHV, Prism) run on AWS and Azure bare-metal instances in the customer's own cloud account and VPC or VNet. Customers pay Nutanix for the software and the cloud provider for the instances, and Nutanix says it's not a managed service. The bare metal is read on the AWS and Azure rows. The AHV and AOS opinions are the same captive ones as on premises.

NVIDIA-Provided

NVIDIA vGPU on AHV

Generally available NVIDIA vGPU with live migration on AHV. Supported accelerators are vGPU-class: L40S, H100, L4 and one Blackwell SKU, the RTX PRO 6000 Server Edition (via vGPU 19.0). The runtime and drivers are NVIDIA's; Nutanix integrates and schedules them.

HGX Training GPUs Not GA on AHV

HGX B200-class systems aren't generally available on AHV: topology-aware AHV for eight-GPU HGX profiles is in early access (March 16, 2026), and Nutanix's NCI 7.6 post on eight-B200 profiles "reflects an experiment in a test environment." Until it ships, that caps on-platform AI at inference and fine-tune scale, not frontier training.

◆ Gap Analysis

Like VMware, Nutanix is an abstraction layer, not physical infrastructure. It manufactures no silicon and no networking. The buyer runs AHV on the commodity x86 it already buys: NX appliances, OEM platforms from Dell, HPE, Lenovo, Cisco and Fujitsu, or software-only on Supermicro, ProLiant and PowerEdge. Nutanix Cloud Clusters (NC2) runs the same stack on AWS and Azure bare metal in the customer's own account. GPU passthrough and generally available NVIDIA vGPU with live migration give accelerated VMs under the Prism console. The architect's concern is the ceiling. Generally available AHV tops out at vGPU-class accelerators; topology-aware AHV for HGX is in early access, while VCF 9.1 supports HGX Blackwell with NVLink Switch. Platform software at 0A reads moderate. Calibration: Red Hat, SUSE and VMware read moderate on the same shape; Dell and HPE strong on hardware they sell.

◆ Borrowed Judgment

Delegated, on a proprietary platform. AHV's scheduler (ADS) places VMs, but a VM-host affinity policy is enforced mandatorily, so neither the scheduler nor high availability can power on or migrate a pinned VM outside its host list: vendor decides, visible, overridable, Delegated. Portability: AHV, AOS and Prism opinions can't be lifted to ESXi or Hyper-V without rebuilding, Ceded; the choice of OEM stays the enterprise's, Retained.

◆ Working Notes

Named, not scored: AMD Instinct support (strategic partnership February 2026, first platform targeted late 2026, no GA date). Confidence flag: AHV MIG-backed vGPU couldn't be confirmed from open sources (login-gated compatibility matrix); it doesn't affect the cell. NC2's bare metal is read on the AWS and Azure rows. Watch-list, notes only (October 23, 2026): topology-aware AHV for HGX B200-class servers, early access since March 16, 2026 and "expected to be available soon," with no date.

Layer 0B · FabricNetwork FabricFlow Virtual Networking and Flow Network Security Over Any Fabricdecides: vendor · Ceded▼

Interconnects, switching and routing, network virtualization

Vendor-Provided

Flow Virtual Networking / Flow Network SecurityCeded

AHV software-defined networking: virtual routers and overlays (Flow Virtual Networking) plus distributed stateful microsegmentation (Flow Network Security, now with a Next-Gen policy model). Its gateways also connect sites over IPsec VPN, Layer 2 VXLAN or BGP peering. Proprietary SDN: the network policy opinions do not port. Proprietary Nutanix platform: opinions captive, no open exit.

NVIDIA-Provided

Assessment pending

◆ Gap Analysis

Flow Virtual Networking is AHV's software-defined networking: virtual routers, overlays and virtual private clouds (VPCs), each with a gateway host. Those gateways also connect sites over IPsec VPN, Layer 2 VXLAN or BGP peering. Flow Network Security adds distributed stateful microsegmentation, now with a Next-Gen policy model. The buyer gets routing and segmentation under the same Prism console, on whatever switches it owns. The architect's concern is that it isn't an AI fabric. GPU-to-GPU bandwidth and congestion control come from the switch vendor. The network policy opinions don't port off Nutanix, either. Calibration: Red Hat, SUSE and VMware read moderate on network virtualization; NVIDIA and HPE strong on fabrics they sell.

◆ Borrowed Judgment

Ceded where the grade rests. The rules the enterprise writes execute as written, but the Flow network controller randomly selects the AHV host that acts as each VPC's gateway, and scale-out sets how many gateways run, which is a bound, not a pin: vendor decides, visible, not overridable, Ceded. That's the Red Hat reading; VMware reads Delegated because NSX lets the enterprise pin gateways. Portability: Flow is Nutanix's proprietary networking, Ceded.

◆ Working Notes

Named, not scored (October 22, 2026): GPUDirect RDMA on AHV. Nutanix's NCI 7.6 post (July 28, 2026) demonstrates it but says the content "reflects an experiment in a test environment," and no product documentation was found. Even when documented, it would be enablement of NVIDIA's NICs, a note rather than a chip.

Layer 0C · StorageStorage SubstrateAOS Distributed Storage and Nutanix Unified Storage; General-Purpose, Not Built for AI Throughputdecides: vendor · Delegated▼

Arrays, storage servers and software, front-end protocols: where the data sits

Vendor-Provided

AOS Distributed Storage OSCeded

The AOS half of "AOS + AHV": the distributed storage OS that places, replicates and tiers data across the cluster nodes, with captive storage policies managed in Prism.

NUS: Files & VolumesCeded

Unified file (SMB/NFS) and block (iSCSI, NVMe-oF/TCP) storage on AOS, managed in Prism. File and block platform opinions (tiering, performance policy, AOS integration) are captive and cannot be lifted to PowerScale or VAST without rebuilding. Proprietary Nutanix platform: opinions captive, no open exit.

NUS: Objects (S3 interface)Ceded

S3-compatible object store on AOS. The S3 API is its own chip. The object store behind it, its placement, replication and lifecycle on AOS and the policies managed in Prism, is Nutanix's and doesn't lift to another S3 store: Ceded.

Front-End Protocols on NUS Files & Volumes (SMB, NFS, iSCSI)Delegated

SMB, NFS and iSCSI, the protocols Nutanix's Unified Storage datasheet lists, which independent implementations share: hosts and applications re-point to another server or target speaking the same protocols, and the data and provider-specific policies migrate with it. Delegated. AOS and Prism's policies stay in the Files & Volumes chip, Ceded.

S3 Front End on NUS ObjectsDelegated

The S3 API, which independent object stores implement. Workloads confined to the supported common S3 operations move by changing endpoints, credentials and provider-specific configuration; the data, policies and unsupported service features migrate or adapt. Delegated. The object store on AOS is the NUS Objects chip, Ceded.

NVIDIA-Provided

NUS NVIDIA Certification (Storage Path)

The June 2026 NVIDIA certification is a storage certification: NUS feeding external HGX and x86 GPU servers over Spectrum-X, Spectrum-4 and BlueField-3, with GPUDirect over RDMA. It isn't GPUs running on Nutanix, and it doesn't extend the AHV GPU support matrix.

◆ Gap Analysis

AOS is the distributed storage OS under AHV, placing, replicating and tiering data across the cluster nodes. Nutanix Unified Storage (NUS 5.3) runs on it: file over SMB and NFS, block over iSCSI and NVMe-oF/TCP, and S3-compatible objects, all managed in Prism. Unlike VMware, Nutanix owns its storage instead of leaning on external arrays. The buyer gets one storage platform for VMs, files and objects on the same nodes. The architect's concern is that NUS is general-purpose. Its NVIDIA certification describes GPUDirect Storage for external GPU servers, but no product document shows it on Files yet, and S3 over RDMA is still roadmap. Calibration: Red Hat, SUSE and VMware read moderate on platform storage; WEKA and VAST strong on storage built for the job.

◆ Borrowed Judgment

Delegated where the grade rests. Each storage container carries its own replication factor, and Prism Central storage policies set two or three copies for the VMs in their categories, overriding the container, a per-object handle the engine must honor: vendor decides, visible, overridable, Delegated. Portability: AOS and the file and block opinions can't be lifted to PowerScale or VAST without rebuilding, Ceded; NUS Objects sits behind the S3 standard, Delegated.

◆ Working Notes

Named, not scored (October 16, 2026): the NVIDIA enterprise storage certification announced May 31, 2026, pNFS and GPUDirect Storage over NFS on RDMA on an all-NVMe cluster, sourced to a press release and a blog. Public evidence that moves the cell: Nutanix product documentation of pNFS and GPUDirect Storage on Files. Watch-list, notes only: S3 over remote direct memory access (RDMA), a roadmap item for later 2026, and NetApp ONTAP integration, targeted for the second half of 2026. Neither is scored.

Layer 1A · DataData Foundation & GovernanceData Lens Compliance and Audit Governance; No Content Classification, No Lineagedecides: code · Retained▼

The data itself, organized and governed: the Governance Catalog that Layer 2C queries

Vendor-Provided

Data Lens (Governance)Ceded

Compliance and audit governance: access auditing, GDPR/HIPAA/SOX reporting, permission/risk analytics, ransomware detect-and-block. Data Lens 2.0 adds on-prem/air-gapped operation. Governs on metadata and access patterns, not AI/ML semantic content. Proprietary Nutanix platform: opinions captive, no open exit.

NVIDIA-Provided

No Direct NVIDIA Layer 1A Governance Dependency

NVIDIA provides nothing in the governance layer. Its only touch is the NUS storage data path (Spectrum-X / GPUDirect over RDMA) feeding external GPU servers: data-plane performance plumbing, not a Layer 1A governance catalog.

◆ Gap Analysis

With NUS at 0C, Nutanix's 1A rests on Data Lens. It's genuine compliance governance over the data estate: access auditing, permission and risk analytics, and ransomware detect-and-block. It reports against the General Data Protection Regulation (GDPR), the Health Insurance Portability and Accountability Act (HIPAA) and Sarbanes-Oxley (SOX). Data Lens 2.0 runs on premises and air-gapped. However, it governs on metadata, permissions and access patterns. There's no semantic content classification, no AI metadata enrichment and no data lineage of the kind a reasoning plane would query. 1A is graded on a catalog that classifies the data, and that's governance over data without classification. It answers who touched the data, not what the data is. Calibration: Dell and WEKA read moderate on metadata-only indexes; Red Hat, SUSE and VMware gap with no data governance at all. Moderate.

◆ Borrowed Judgment

The grade rests on access auditing, permission analytics and compliance reporting, which execute without inference; anomaly detection runs inside thresholds and exclusions the administrator sets, and the grade holds without it: code decides, visible, overridable, Retained. Portability: Data Lens is Nutanix’s proprietary platform, opinions captive, Ceded.

◆ Working Notes

Public evidence that moves the cell: content classification or lineage in Data Lens.

Layer 1B · RetrievalContext Management & RetrievalFoundational RAG via Managed Postgresdecides: vendor · Delegated▼

Low-latency retrieval for RAG — vector/hybrid search, context windows

Vendor-Provided

Vector Search via NDB-Managed PostgreSQL (pgvector)Ceded

GA. NDB v2.10 automates the lifecycle of PostgreSQL with the pgvector extension, framed by Nutanix as a vector DB for RAG. The PostgreSQL with pgvector surface is its own chip. The service behind it, its operations, management and configuration, is Nutanix's and doesn't lift: Ceded.

PostgreSQL With pgvector on Nutanix Database ServiceDelegated

PostgreSQL with pgvector has independent implementations: code and definitions written to it move to another compatible implementation within the features both support, the data moves by export, replication or rewrite, and the provider's control-plane configuration and operational state don't lift. Delegated. The service behind it is its own chip, Ceded.

NVIDIA-Provided

NVIDIA AI Enterprise + NeMo Retriever (Accelerated Path)

The accelerated-retrieval path layers NVIDIA AI Enterprise and NeMo Retriever over NUS. The same non-differentiating stack is available on Dell, HPE, and VMware: NVIDIA is the accelerator, not the retrieval owner.

◆ Gap Analysis

A Nutanix buyer who wants RAG has one genuinely shipping, owned path: NDB-managed PostgreSQL with pgvector (GA since NDB v2.7, Jan 2025; current v2.10), which Nutanix markets explicitly as turning Postgres into a vector database for RAG, with lifecycle automation (provision, patch, clone, HA). For moderate-scale enterprise RAG on data already kept in Postgres, that is a real, low-friction capability. The retrieval intelligence, however, is not Nutanix's. pgvector is OSS; the heavier AI-native retrieval path Nutanix markets, third-party Milvus plus the NVIDIA AI Data Platform on NUS, has no GA statement on Nutanix's pages and isn't scored. Critically, NUS itself has no native vector database, no vector search, and no RAG retrieval in the storage layer: the AI-ready-data messaging is data-plane plumbing, not retrieval. There is no retrieval-quality observability (recall@k, latency percentiles) a Layer 2C could consume: the same universal gap noted in Dell and VMware. This calibrates almost exactly to VMware Layer 1B (moderate): VMware's owned vector path is also pgvector-on-Postgres, the same pragmatic, performance-ceilinged choice, short of Dell (Elasticsearch hybrid search + MetadataIQ) and VAST (native InsightEngine).

◆ Borrowed Judgment

Moderate. The retrieval opinions live in stock OSS PostgreSQL + pgvector; NDB provides lifecycle automation around a portable open core. A pg_dump/restore moves the vector schema, embeddings, index definitions, and queries to any other Postgres (RDS, Azure, self-hosted), so the retrieval layer lifts out. Nutanix's contribution is operation, not retrieval intelligence.

◆ Working Notes

pgvector index types (HNSW/IVFFlat) under NDB management are not confirmed from open sources. Does not affect the cell. The KServe serving substrate is industry-attested but not officially named by Nutanix (vLLM is the documented engine). Removed in the 1B dependency audit (October 24, 2026), to notes: Milvus and the NVIDIA AI Data Platform on NUS. Nutanix's NVIDIA page calls it an "NVIDIA-validated AI Data Platform (AIDP) solution" with no GA wording, and the AI-Q tie-in is a March 2025 blog. NDB 2.10's general availability is stated in Nutanix's April 7, 2026 blog; the NDB release notes are behind a Nutanix portal login, so the badge isn't publicly confirmed. The blog, Nutanix's own GA announcement of a versioned release rather than an availability statement, is accepted as the public evidence (October 24, 2026).

Layer 1C · PipelinesData Movement & PipelinesNo Pipeline Layerdecides: absent · Absent▼

Move/transform data — ETL/ELT, lineage, cost-aware movement, KV cache tiering

Vendor-Provided

NVIDIA-Provided

No NVIDIA Layer 1C Dependency

NVIDIA provides nothing at Nutanix Layer 1C. There is no GA CMX/KV-cache tiering on AHV; kvCache-aware routing in NAI is Tech Preview and is request-scheduling within an endpoint, not data-pipeline tiering. A near-empty NVIDIA column here is itself the finding: same as VMware Layer 1C.

◆ Gap Analysis

Within the Nutanix world, data mobility is genuinely good, but it is storage mobility, not AI pipelines. NDB Time Machine provides copy-data management (point-in-time recovery, thin clones); snapshots replicate across clusters and clouds (NC2); native DB replication (Oracle Data Guard, Postgres pglogical) moves data between instances. None of that is a Layer 1C AI data pipeline. There is no dedicated ETL/ELT product, no CDC pipeline, no feature engineering, no lineage, no cost-aware movement, and no KV-cache tiering. The closest thing to data movement is OSS pglogical, which is not Nutanix IP. The enterprise running Nutanix for AI must bring its own pipeline orchestration (Airflow, Kubeflow, Spark, or a commercial alternative) and run it on NKP. This is the same call as VMware Layer 1C (gap): both lack an equivalent to Dell's Dataloop-powered orchestration, HPE's Ezmeral, or VAST's DataEngine. The gap has a downstream cost. A reasoning plane (Layer 2C) makes data-relative placement decisions (run compute where the data lives, weigh moving data versus moving compute), and to do so it needs the lineage and cost-aware-movement primitives this layer would provide. Their absence is one of the structural reasons a Nutanix Layer 2C cannot do data-relative placement.

◆ Borrowed Judgment

The enterprise retains responsibility for the function by default: no vendor has claimed it. Any data-pipeline judgment is borrowed from whatever tooling the enterprise deploys on NKP (Airflow, Kubeflow, commercial), not from Nutanix. Storage and DB mobility (Time Machine, snapshot replication, pglogical) exist but are adjacent to, not the same as, the AI-pipeline function this layer scores.

◆ Working Notes

Watch-list (dated, none scored): NetApp ONTAP integration targeted 2H 2026; S3-over-RDMA roadmap (later 2026); kvCache-aware routing in NAI 2.7 is Tech Preview (and is request-scheduling within an endpoint, not data-pipeline tiering). Time Machine and cross-cloud snapshots are described here as storage/DB mobility, not promoted to scored components: they do not clear the AI-pipeline bar.

Layer 2A · OrchestrationInfrastructure OrchestrationNutanix Heritage Strengthdecides: vendor · Delegated▼

GPU scheduling, quotas, RBAC, fair-share scheduling, utilization optimization

Vendor-Provided

Prism Central (Unified Fleet Management)Ceded

Multi-cluster management, RBAC, quotas, self-service, and full-stack lifecycle across the Nutanix estate. The operational backbone, comparable to VMware's SDDC Manager and VCF Operations. Prism policies and management opinions do not port. Proprietary Nutanix platform: opinions captive, no open exit.

AOS + Acropolis Dynamic Scheduling (ADS)Ceded

Automated VM placement, load-balancing, and resource management across the cluster. The placement opinions are captive to AOS, a proprietary scheduler. Proprietary Nutanix platform: opinions captive, no open exit.

Nutanix Kubernetes Platform (NKP)Delegated

Turnkey Kubernetes built on pure upstream, CNCF-conformant Kubernetes (unforked, no proprietary API wrapper) via Cluster API. The consumed interface is the standard Kubernetes API, so manifests and workloads lift to EKS, AKS, or any conformant cluster without rebuilding. The Konvoy/Kommander management layer is Nutanix's; the consumed interface keeps the workloads portable.

NVIDIA-Provided

NVIDIA GPU Operator

GPU lifecycle and scheduling on NKP via the NVIDIA GPU Operator (partner OSS) in the Kommander catalog. There is no Nutanix-owned GPU scheduler and no Run:ai equivalent: the GPU-aware layer is borrowed, the same dependency every on-prem vendor shares.

NVIDIA vGPU Manager

vGPU profiles and allocation on AHV. Multi-tenant GPU partitioning is managed through Prism, but the virtualization layer itself is NVIDIA's.

◆ Gap Analysis

Layer 2A is Nutanix's home turf and its deepest IP. Prism Central provides mature, multi-cluster fleet management with RBAC, quotas, and self-service; AOS with Acropolis Dynamic Scheduling (ADS) automates VM placement and rebalancing; and NKP 2.17 delivers Kubernetes. One operational model (the console the buyer already runs) now extended to AI workloads. For a VMware-refugee shop (Nutanix's core go-to-market), this is the lowest-friction orchestration on-ramp to AI available. The calibration matches VMware Layer 2A (strong): both are HCI/private-cloud platforms whose deepest competency is unified infrastructure orchestration, and both carry the same NVIDIA GPU-scheduling dependency, which the layer treats as universal. Two honest qualifiers keep Nutanix a hair behind VMware: VMware's 2A remains the peak of the series on raw operational depth (100M+ cores, two decades, a single vSphere Supervisor managing VMs, containers, and AI from one plane), and Nutanix's unification is marginally less seamless: VMs via Prism/AHV and Kubernetes via a distinct NKP product rather than one supervisor. Still well ahead of Dell and Cisco 2A (moderate), where GPU orchestration is fully NVIDIA-owned. Capability and deployment are separate readings here, as they are on VMware. AOS and AHV are frequently bought as the virtualization substrate underneath another vendor's orchestration, and OpenShift or a customer-run Kubernetes distribution on AHV is a common shape. In that stack the customer consumes Nutanix for Layer 0 and infrastructure lifecycle while Red Hat or the customer's own platform team owns the Kubernetes control plane, so NKP is licensed and unconsumed and the 2A judgment the enterprise inherits belongs to whoever owns the orchestration. The grade reflects what an architect can deploy on Nutanix paper. It does not assert that a given customer runs it. A strong cell is an available option, not an occupied slot. GPU-aware scheduling is the borrowed piece: NKP schedules GPUs via the NVIDIA GPU Operator, and there is no Nutanix-owned GPU scheduler. Policy-driven GPU scheduling (which workload gets which GPU on cost or compliance grounds) is NVIDIA's, not the platform's: the identical gap Dell and VMware carry.

◆ Borrowed Judgment

Low: the lowest of any Nutanix layer. Prism, AOS, and ADS are Nutanix IP. The primary borrowed piece is GPU-aware scheduling (NVIDIA GPU Operator), the same dependency every on-prem peer shares. NKP's consumed interface (unforked upstream Kubernetes) keeps container workloads portable to any conformant cluster. Decision authority under the override rule (September 4, 2026): ADS honors the mandatory VM-host affinity policy, and NKP is upstream Kubernetes, where nodeSelector and required affinity pin pods. Vendor decides, visible, overridable, Delegated, the Red Hat and SUSE reading.

◆ Working Notes

NKP is explicitly pure upstream CNCF Kubernetes with no proprietary API wrapper (D2iQ lineage: Konvoy/Kommander/NKP Insights are Nutanix IP; the runtime is unforked). NKP Metal (bare-metal Kubernetes) is Early Access with GA targeted 2H 2026, logged as watch-list.

Layer 2B · RuntimeApplication Runtime & ExecutionPlatform-Native Serving, Agents Pre-GAdecides: model · Delegated▼

Model serving, agent execution, inference APIs, distributed inference

Vendor-Provided

Nutanix Enterprise AI (NAI): Model Serving & GovernanceCeded

GA. RBAC-governed inference endpoints; 74-model validated catalog; multi-source models (NVIDIA NIM, Hugging Face, custom upload); CPU-only inference via Intel AMX; batch inference; vLLM default engine. Runs on any CNCF Kubernetes. NAI is proprietary Nutanix software: the serving and governance opinions (catalog, endpoint configuration, RBAC policy) do not lift to another serving stack without rebuilding, and running it on EKS does not change that. Proprietary Nutanix platform: opinions captive, no open exit.

Model Access via OpenAI-Compatible Endpoints (Over vLLM, NVIDIA NIM, and TensorRT-LLM)Delegated

The model execution engines NAI orchestrates sit behind OpenAI-compatible endpoints. vLLM is the documented OSS default. Applications built against the endpoint repoint to any compatible provider: Delegated on the inference-interface ruling.

NVIDIA NIM Through NAI (Optional Path, With TensorRT-LLM, Under NVIDIA AI Enterprise)Ceded

NIM is an optional path in NAI's multi-source catalog, not a lock; the vLLM default doesn't need it. Where the enterprise takes it: Proprietary containers on an NVIDIA AI Enterprise license with no source; the TensorRT-LLM engine builds, profiles, and NIM Operator configuration the enterprise accumulates don't lift to another runtime. A surface beyond the interface: Ceded to NVIDIA, the SUSE, Dataiku, and WEKA reading.

Customer-Created and Managed Tools (Function Calling / Client-Side)Retained

GA through NAI's OpenAI-compatible inference endpoints. The enterprise owns and operates the business logic, APIs, commands and deterministic validators invoked through tool calls: the seam where control passes from instructing the model to executing code outside it. The opinions are the enterprise's and run against a multi-vendor interface, so they lift to another provider with an adapter and re-evaluation, not a rebuild. What the enterprise cedes at this seam is the decision to act: the model chooses whether, when and with which arguments to call, and that judgment is inherited under the serving component above. Added instrument-wide on September 2, 2026 (M4): the seam exists on every runtime that hands control to portable customer code, and a real Retained position is scored, not narrated.

NVIDIA-Provided

NVIDIA NIM (One Serving Option, Not a Lock)

NIM is a model-serving option inside NAI (alongside Hugging Face and custom upload), not a requirement. NeMo/Nemotron are enabled and TensorRT-LLM is available via NIM. The finding is the de-emphasis: NAI is explicitly not NIM-locked.

Lighter NVIDIA Dependency Than Peers

A vLLM OSS default engine plus an Intel-AMX CPU-only inference path mean Nutanix's runtime is less NVIDIA-captive than Dell's all-NVIDIA NemoClaw/OpenShell 2B path. A genuine, scoreable differentiator.

◆ Gap Analysis

Nutanix Enterprise AI (NAI 2.7, GA May 2026) is Nutanix's strongest owned AI asset. It deploys LLMs as secure, RBAC-governed inference endpoints from NVIDIA NIM, Hugging Face, or a custom upload; offers a 74-model validated catalog, batch inference, a CPU-only inference path (Intel AMX), and vLLM as the default engine; and runs on any CNCF Kubernetes (EKS/AKS/GKE/bare-metal), not just Nutanix. As a standalone model-serving product it is arguably cleaner and less NVIDIA-locked than VMware's Model Runtime. Two concerns. The inference runtime is borrowed: vLLM (OSS) and TensorRT-LLM via NIM (NVIDIA); NAI is the control and governance plane on top, and that plane's opinions are captive to NAI. And the agent-execution half of this layer is not GA: Layer 2B is model serving, agent execution, inference APIs, and distributed inference, but the native agent framework (NAI Labs – Agent), MCP servers, RAG pipeline, LoRA fine-tuning, and kvCache-aware routing are all Tech Preview in NAI 2.7. Per the GA-gate, none of those count toward the score. The result lands at VMware Layer 2B (moderate), with a different composition: VMware's Model Runtime and Agent Builder are both GA, whereas Nutanix's serving is GA and arguably more open, but its agent execution is preview where VMware's is GA. Nutanix trades VMware's GA agent-builder for a more open, less NVIDIA-captive serving runtime. Both sit well above Dell 2B (entirely NVIDIA NemoClaw/OpenShell-dependent).

◆ Borrowed Judgment

Moderate. Nutanix owns the NAI serving and governance plane (its authority, captive to Nutanix), while the inference runtime is borrowed from vLLM (OSS) and NVIDIA (NIM): the same structural runtime dependency Dell, HPE, and VMware carry, softened here by the OSS default engine and the CPU-only path. Agent execution is not yet GA, so there is no agent-runtime judgment to score.

◆ Working Notes

GA-gate watch-list (dated, none scored): native agent framework / NAI Labs – Agent (Tech Preview), MCP servers (TP), RAG pipeline (TP), LoRA fine-tuning (TP), kvCache-aware routing (TP), all in NAI 2.7; and the 'Nutanix Agentic AI' full stack (Early Access, GA targeted 2H 2026). Palo Alto Prisma AIRS model scanning is GA in NAI 2.7 (scored at Layer 3).

Layer 2C · ReasoningAgentic Infrastructure — The Reasoning PlaneGateway Governance, No Placementdecides: vendor · Delegated▼

Policy-driven placement and resource coordination — the Autonomy Layer

Vendor-Provided

Agent Gateway (NAI 2.7): Governed Agent and Model TrafficCeded

GA. A single governed control point for agentic traffic: a unified API in front of both cloud and self-hosted models, RBAC, per-token rate-limiting, and full audit including MCP-call audit, plus health-based endpoint failover and capacity load-balancing. All agent and model traffic flows through it, which is real enterprise dependence rather than an emerging signal. The gateway is proprietary Nutanix software: the RBAC policies, rate limits, and routing rules configured here do not lift to another agent control plane without rebuilding. Matches the Ceded call on Salesforce MuleSoft Agent Fabric governance and on VMware AgentMinder’s runtime enforcement gateway. Not scored within this chip under the GA evidence table: MCP server traffic and MCP-call audit ("Model Context Protocol (MCP) servers* (in Tech Preview)").

NVIDIA-Provided

No NVIDIA Layer 2C Dependency

Agent Gateway is Nutanix IP. NVIDIA provides no agent governance, model routing, or placement reasoning in the Nutanix stack: the same pattern as Google's and Cisco's 2C, where the governance layer is vendor-owned, not NVIDIA-dependent.

◆ Gap Analysis

With Agent Gateway (GA in NAI 2.7), the buyer gets a single, governed control point for agentic traffic: a unified API in front of both cloud and self-hosted models, RBAC, per-token rate-limiting, full audit including MCP-call audit, plus health-based endpoint failover and capacity load-balancing. For an enterprise worried about ungoverned agent and model sprawl, that is a real, shipping governance surface. Applying the 'Routing Is Not Reasoning' test: Agent Gateway governance (RBAC, rate-limiting, audit, MCP audit) is access control: genuine intelligence-layer governance, but not placement. Health-based failover and capacity load-balancing are single-variable operational routing (is the endpoint up, is it at capacity?), not multi-variable policy-driven placement. Per-request model selection on content, cost, or quality does not exist; kvCache-aware routing is Tech Preview, and even that schedules to GPU workers within a single endpoint, not among models. None of it is the reasoning plane Layer 2C defines. The gap is structural, not merely unbuilt. A reasoning plane makes data-relative placement decisions, and the data-side feeders it would query (an AI-metadata catalog at Layer 1A, lineage and cost-aware movement at Layer 1C) are absent. So even when the forward 'Nutanix Agentic AI' full stack reaches GA (Early Access today, targeted 2H 2026), its placement reasoning will be starved of data-side inputs. The grade follows the shipped surface rather than the missing one. A GA gateway that every agent and model call passes through is real enterprise dependence, which is the line between gap and moderate, and it is the same class of capability that reads moderate on Cisco (Duo Agentic Identity, AI Defense, Splunk), Salesforce (MuleSoft Agent Fabric registry, broker, gateway) and VMware (AgentMinder identity, enforcement, audit). Nutanix sits at the thinner end of that band: a gateway without the agent registry Salesforce ships or the identity-and-intent model VMware ships. What none of them has, Nutanix included, is a reasoning plane. Missing legs, named (September 2, 2026): first-class agent identity, registry, cross-agent orchestration. Gateway governance alone is moderate's floor.

◆ Borrowed Judgment

Low for what ships, and the enterprise owns what does not. Agent Gateway is Nutanix IP with no NVIDIA dependency, so the governance and operational-routing functions carry little borrowed judgment, though the policies written into the gateway are captive to it. Placement, per-request model selection, and outcome validation are not provided here and are not provided by any vendor on the instrument, so the enterprise supplies that judgment itself or operates without it.

◆ Working Notes

Watch-list (dated, none scored): kvCache-aware routing (Tech Preview); the 'Nutanix Agentic AI' full stack (Early Access, GA targeted 2H 2026); native agent framework and MCP servers (Tech Preview). Agent Gateway is scored, not logged: it is GA and carries production agent traffic. The data-side structural point still holds and is the reason this cell does not read higher: an AI-metadata catalog at 1A and lineage at 1C are absent, so any future placement reasoning would be starved of the inputs it would need to query. Instrument note: live Infrastructure-2C (a policy engine deciding at request time where inference runs relative to the data, which model serves the request, and how cost, latency, and compliance are arbitrated) does not ship from any vendor on this instrument. It is recorded here as a universal finding about the market, not as a charge against this vendor. You can't prompt your way to deterministic output: the controls here answer whether an action is legal, never whether the outcome was right. That deterministic outcome-validator is missing instrument-wide, and like live placement it is noted as universal rather than charged to one vendor. Out of the Agent Gateway chip under the GA evidence table (October 24, 2026): MCP server traffic and MCP-call audit ("Model Context Protocol (MCP) servers* (in Tech Preview)").

Layer 3 (+1) · ApplicationsAI Application Layer — The Value PlaneNo Applications on Nutanix Paper; the AI Partner Program Is Validationdecides: absent · Absent▼

AI-powered business capabilities — business logic, workflow automation

Vendor-Provided

NVIDIA-Provided

NVIDIA Model Ecosystem (Nemotron + NIM)

Nemotron models and NIM containers are available through the NAI catalog. NVIDIA provides a slice of the model layer; Nutanix provides serving and governance. The same non-differentiating pattern as VMware Layer 3.

◆ Gap Analysis

Nutanix sells no AI business applications. It sells the platform to build and run them (Nutanix Enterprise AI, scored at 2B and 2C) and a validated on-ramp to partners: the AI Partner Program offers "a set of validated full-stack solutions" carrying "Nutanix AI Ready validation" (DataRobot, Codeium, Instabase, Pryon, Lamini, UbiOps), and the Marketplace is a deployment catalog of apps "developed and certified by Nutanix in collaboration with the preferred partners," with no Nutanix billing or support for the applications. The validated model catalog is model choice, read at 2B, and Palo Alto Prisma AIRS scans models in the serving path, also 2B. Layer 3's grade rests on business applications the enterprise can buy or get supported on the vendor's paper (as ruled October 19, 2026: business applications bought or supported on the vendor's paper). Certification isn't procurement. Calibration: Cisco, Arista and Nokia read moderate on first-party security applications; HPE reads partner on ISV applications customers buy from HPE on the GreenLake Marketplace; Dell, Nutanix, Red Hat, VMware, VAST and Supermicro read gap on the same certification-only shape.

◆ Borrowed Judgment

None on Nutanix paper. Applications are built by the enterprise on Nutanix Enterprise AI or bought from ISVs directly; Nutanix validates the combinations.

◆ Working Notes

Named, not scored: Nutanix AI Partner Program (validation, May 2024), Nutanix Marketplace (deployment catalog), the validated model catalog with Hugging Face (2B), Prisma AIRS (2B).

Sources & revision history · v1.28 - 4+1 v3: Layer 0 Sub-Layers

.NEXT 2026 (Apr 2026), .NEXT 2025, NAI 2.7 (May 2026), AOS 7.5 / AHV 11 / Prism Central pc.7.5, NUS 5.3, NDB v2.10, NKP 2.17, Nutanix Bible, NVIDIA vGPU product-support matrix for AHV, published 4+1 model. v1.1 (instrument reconciliation): 1A NUS Objects Retained→Delegated, resolving the internal split with NKP (2A): proprietary implementation behind a multi-vendor standard is Delegated, not Retained. /reconcile (September 1, 2026): 2C moved gap to moderate. Agent Gateway (GA, NAI 2.7) is real enterprise dependence and the same capability class that reads moderate on Cisco, Salesforce, and VMware; the prior gap rested on a calibration to VMware 2C that no longer holds. Retired status vocabulary removed from borrowedJudgment. /reconcile (September 1, 2026): shared 2C findings stated explicitly (live-placement universality). /reconcile (September 1, 2026): 2A separates available capability from occupied slot; NKP is commonly displaced by OpenShift or customer-run Kubernetes on AHV, matching the VMware 2A treatment. /reconcile (October 5, 2026): Layer 3 moderate to partner (platform enablement earns no Layer 3 grade); the Nutanix Enterprise AI platform-enablement component removed from Layer 3, where it is already scored at 2B and 2C. Summary component count corrected to nineteen (October 5, 2026). /reconcile (October 5, 2026): NIM column (ruled October 4, 2026: the endpoint in front is Delegated, the NIM containers the enterprise operates are Ceded to NVIDIA). The mixed runtime chip split: endpoints Delegated, the optional NIM path Ceded to NVIDIA (logged on the SUSE row, September 5, 2026). Grade unchanged. /reconcile (October 5, 2026): override rule applied to a cell scored before it; layer0, layer1b, layer2a authority Ceded to Delegated on documented pins or per-request overrides. Grades and components unchanged. /reconcile (October 5, 2026): Layer 0 NVIDIA GPU Integration holds Delegated on a corrected reason: portable across hypervisors, captive to NVIDIA (NVIDIA column). Grade unchanged. v1.10 (Layer 0 sub-layers, October 16, 2026): Layer 0 split into 0A Compute, 0B Network Fabric and 0C Storage Substrate by runtime decision (Keith, October 16, 2026); storage left 1A for 0C, and 1A, now Data Foundation & Governance, is graded on the catalog that classifies the data. 0A moderate, 0B moderate, 0C moderate; summary edits (summary worksheet, Keith, October 16, 2026). v1.11 (October 19, 2026, /reconcile partner ruling): Layer 3 partner to gap, authority Absent; the AI ISV program is certification, not procurement on the vendor's paper. (Same pass) Plain-language pass: methodology rule numbers in the prose and audit notes replaced with the reasoning they stood for; no grade, chip or authority reading changed. v1.12 (October 19, 2026, /reconcile voice pass): em-dashes in the prose replaced with periods, colons, commas or parentheses; wording, grades, chips and authority readings unchanged. v1.13 (October 22, 2026, /reconcile): 0B Flow chip records site connectivity over IPsec VPN, Layer 2 VXLAN or BGP peering (Nutanix's Flow Virtual Networking page), and the notes name GPUDirect RDMA on AHV as a vendor test, not scored. Found by the 0B fabric research. Grade and authority unchanged. v1.14 (October 23, 2026, /reconcile): 0A's NVIDIA ceiling reads "not GA" in place of "not virtualizable": topology-aware AHV for HGX B200-class servers is in early access (March 16, 2026), and it goes to a named, unscored note; the unverified GB200 and GH200 claim comes out; a rule code comes out of the gap. Found by the 0A compute research. Grade and authority unchanged. v1.15 (October 23, 2026, storage door pass): NUS Files and Volumes and NUS Objects get front-end doors; the storage on AOS reads Ceded. Grade unchanged. v1.16 (October 23, 2026, Layer 1 managed-services pass): Vector Search via NDB-Managed PostgreSQL splits. Grade unchanged. v1.17 (October 24, 2026, /reconcile): 1B Milvus and NVIDIA AI Data Platform chip to notes: marketing with no GA wording, built on NVIDIA's reference blueprints. Found in the 1B dependency audit under the headend standard. Grade unchanged. v1.18 (October 24, 2026, /reconcile): 1B note: NDB 2.10's general availability rests on Nutanix's own GA announcement, accepted as the public evidence. Found in the 1B dependency audit follow-ups. Grade unchanged. v1.19 (October 24, 2026, /reconcile): GA evidence recorded on 2 of 2 Layer 1B chips under the GA evidence table. Grade unchanged. v1.20 (October 24, 2026, /reconcile): Layer 1B GA evidence migrated to the GA evidence table version 2. Grade unchanged. v1.21 (October 24, 2026, /reconcile): GA evidence recorded on 1 of 1 Layer 1A chips under the GA evidence table version 2. Grade unchanged. v1.22 (October 24, 2026, /reconcile): GA evidence recorded on 1 of 1 Layer 0B chips under the GA evidence table version 2. Grade unchanged. v1.23 (October 24, 2026, /reconcile): GA evidence recorded on 4 of 4 Layer 0A chips under the GA evidence table version 2. Grade unchanged. v1.24 (October 24, 2026, /reconcile): GA evidence recorded on 5 of 5 Layer 0C chips under the GA evidence table version 2. Grade unchanged. v1.25 (October 24, 2026, /reconcile): GA evidence recorded on 3 of 3 Layer 2A chips under the GA evidence table version 2. Grade unchanged. v1.26 (October 24, 2026, /reconcile): GA evidence recorded on 4 of 4 Layer 2B chips under the GA evidence table version 2. Grade unchanged. v1.27 (October 24, 2026, /reconcile): GA evidence table version 2.1 applied: 1 chips recorded or narrowed. Grade unchanged. v1.28 (October 27, 2026): each component now carries a durable offering id; nothing else changed.

4+1 Layer AI Infrastructure Model · Vendor Assessment Series · The CTO Advisor LLC (DBA The Advisor Bench) · thectoadvisor.com