What XPerf Inc does

XPerf Inc is an early-stage AI infrastructure software company founded in 2025 in Round Rock, Texas by ex-Intel engineers, building a vendor-agnostic, AI-native control plane that targets the 60-70% of GPU capacity routinely wasted in private cloud, on-premise, and air-gapped AI data centers. The platform is delivered as on-premise software and is composed of three integrated products: ClusterReady (pre-production GPU rack and cluster validation that compresses multi-week L11/L12 testing into hours by running real AI inference and training workloads with tokens/sec baselines), OpsPilot (autonomous failure prevention and self-healing driven by patent-pending per-node safety models, 10-30 minute failure prediction, AI-agent root cause diagnosis, and an incident knowledge base), and LoadSteer (an intelligent inference and post-training scheduler built on NVIDIA KAI Scheduler and slated for October 2026 release). The stack supports NVIDIA, AMD Instinct, and ASIC accelerators across InfiniBand, NVLink, and RoCE fabrics and is integrated with Kubernetes, GitOps, Helm, Prometheus, Grafana, and Loki.

The company operates an enterprise field-sales motion targeting large AI data center operators and cloud service providers, with proof-of-concept agreements used as entry points, a 'Request a demo' lead capture on xperf.ai, and conference presence at SC25 and NVIDIA GTC. Revenue is generated through enterprise software licensing and subscriptions across ClusterReady and OpsPilot, with a performance-based pricing model introduced for LoadSteer. XPerf emerged from stealth in November 2025 with over $2.5 million in pre-seed funding from Investwide Capital, made its public debut at SC25 in December 2025 with two PoC agreements and a long-term cloud provider collaboration, launched ClusterReady in March 2026, and is a member of NVIDIA Inception and the Cloud Native Computing Foundation.

Name

XPerf Inc

Legal name

XPerf Inc.

Website

https://xperf.ai

Company type

Private

Founded year

2025

Operating status

Operating

Headcount range

11–50 employees

Short description

XPerf Inc is an early-stage AI infrastructure software company building a vendor-agnostic, AI-native control plane that validates, operates, and schedules GPU and ASIC clusters across NVIDIA, AMD, and custom accelerators for private cloud and on-premise AI data center operators.

Ownership category

akta.pro rank

Product category

AI Infrastructure Software

NAICS

Software Publishers (5132), Computer Systems Design Services (541512)

SIC

Services-Prepackaged Software (7372), Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA)

akta.pro secondary industries

AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG), Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC)

Keywords

• GPU utilization optimization
• AI infrastructure software
• Cluster validation platform
• Workload scheduling software
• Data center observability

Headquarters

HQ city

Round Rock

HQ country

United States

HQ region

North America

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Marketing or Sales, Operations, Infrastructure

Revenue model

1. Software Licenses and Subscriptions: Enterprise software licensing model for ClusterReady, OpsPilot, and LoadSteer products. Performance-based pricing mentioned for LoadSteer. On-premise deployment model suggests perpetual license or subscription terms.

Go-to-market motion1 record

Distribution channels3 records

Marketing channels7 records

Core offering

XPerf develops an AI-native control plane software platform for heterogeneous AI data centers. The platform combines three integrated products—ClusterReady (pre-production GPU rack/cluster validation), OpsPilot (autonomous failure prevention and self-healing operations), and LoadSteer (intelligent inference and post-training workload scheduling)—across NVIDIA, AMD Instinct, and ASIC accelerators. It targets private cloud, on-premise, and air-gapped AI datacenter environments with full on-premise deployment.

Product overview

XPerf Inc is an AI infrastructure software company offering a unified AI-native control plane for heterogeneous data centers, consisting of three integrated products: ClusterReady (pre-production rack/cluster validation), OpsPilot (autonomous failure prevention and self-healing), and LoadSteer (intelligent workload scheduling). The platform targets private cloud, on-premise, and air-gapped datacenter environments, supporting NVIDIA, AMD Instinct, and other accelerators across InfiniBand, NVLink, and RoCE fabrics. The same validation engine powering ClusterReady serves as the calibration source for OpsPilot's autonomous cluster operations.

Differentiator

Problem solved

Functional benefit

Brands

• ClusterReady: Automated AI rack and cluster validation platform that validates GPU racks using real AI inference and training workloads with tokens/sec output. Delivers 10x faster validation than conventional methods.
• OpsPilot
• LoadSteer

Products and services

• ClusterReady Automated AI rack and cluster validation platform that validates GPU hardware using real AI inference and training workloads. Generates tokens/sec performance baselines and provides root cause diagnosis with suggested remedies, reducing validation from weeks to hours. For AI data center operators, OEMs, ODMs, and rack integrators.
• OpsPilot AI-native operations system for autonomous cluster reliability and failure prevention. Continuously monitors AI clusters, predicts failures 10-30 minutes ahead using per-node safety models, performs AI-driven diagnosis, and applies self-healing with human-in-the-loop controls. Fully on-premise with data never leaving the facility. For AI data center operators.
• LoadSteer Intelligent inference and post-training workload scheduler for GPU and ASIC clusters. Built on NVIDIA KAI Scheduler, it routes workloads across NVIDIA, AMD, and ASIC clusters with sub-second response times and performance-based pricing. For AI data center operators managing heterogeneous accelerator fleets.

Quantifiable outcome

• GPU utilization improvement of 20%+
• +6 more outcomes

Named customers2 records

Segments3 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability6 records

Feature4 records

Partnerships

Seven partnerships are on record, tiered minor, core and strategic.

• InfraeominorGTM or Marketing Partner · 1 December 2025Infraeo shared XPerf's booth at SC25 in St. Louis, collaborating on industry event presence. Described as professional team with high industry standards.
• Globally Operating Cloud Service ProvidercoreStrategic or Co-development Partner · 1 December 2025Long-term collaboration established at SC25 with a globally operating cloud service provider for deployment and product development. Represents significant enterprise customer relationship.
• Cloud Service Providers and Data Center Owners (PoC Partners)strategicStrategic or Co-development Partner · 1 December 2025Two proof-of-concept agreements secured at SC25 with cloud service providers and data center owners for deployment and product testing. Early-stage customer development partnerships.
• tidalTAMU / Texas A&M UniversityminorGTM or Marketing Partner · 1 October 2025XPerf sponsored the tidalTAMU Hackathon 2025 at Texas A&M University, a one-day innovation sprint bringing students, creators, and tech enthusiasts together. Supports talent pipeline development and community engagement.
• MiTAC ComputingcoreStrategic or Co-development Partner · 1 March 2025Strategic partnership for ClusterReady product launch and validation demonstration. MiTAC Computing collaborated with XPerf to demonstrate strong cluster-level performance on AMD Instinct MI325X racks, validating AI workloads in hours rather than weeks. Working relationship showcases product capabilities with real hardware partners.
• gBETA / gener8tor / Round Rock ChamberstrategicGTM or Marketing Partner · 1 January 2025XPerf selected for gBETA Round Rock Fall 2025 accelerator cohort (one of six from sixty applicants). Program provides mentorship, investor connections, and startup ecosystem access.
• NVIDIA Inception ProgramminorGTM or Marketing PartnerStartup program membership providing access to NVIDIA resources, technology, and ecosystem for AI and deep learning companies.

Scale indicators10 records

Recent moves6 records

Expansion highlights5 records

Direct peers

• Run:ai: GPU orchestration and workload scheduling platform acquired by NVIDIA. Directly competes with XPerf's LoadSteer for inference and training workload scheduling across GPU clusters, with the deepest overlap in AI infrastructure virtualization.
• Anyscale: Ray-based AI compute platform offering scalable compute orchestration and managed Ray clusters. Competes with XPerf's LoadSteer scheduler and targets the same AI data center operators with end-to-end workload management.
• Weights & Biases: MLOps platform with experiment tracking, model registry, and infrastructure observability tools. Competes with XPerf's OpsPilot for AI workload monitoring, observability, and failure diagnostics across enterprise ML deployments.
• Determined AI: Open-source ML training platform providing cluster orchestration, hyperparameter tuning, and resource scheduling for deep learning. Directly comparable to XPerf's LoadSteer for GPU training workload management.
• Domino Data Lab: Enterprise MLOps platform providing model development, deployment, and cluster governance. Comparable to XPerf's full stack targeting enterprise AI data center operators with platform-level orchestration.
• Lightning AI: AI development platform built on PyTorch Lightning offering training infrastructure and multi-cloud orchestration. Competes with XPerf's training workload management and targets similar AI engineering teams.

Broad incumbents

• NVIDIA AI Enterprise: NVIDIA's enterprise AI software suite including Base Command, NGC, and AI Enterprise frameworks. Broad incumbent offering overlapping orchestration, validation, and scheduling capabilities as part of a wider portfolio tied to NVIDIA hardware.
• Hugging Face: AI platform with model serving, inference endpoints, and Spaces orchestration. Broad incumbent with overlapping inference deployment and optimization capabilities comparable to LoadSteer's inference scheduling.
• CoreWeave: GPU cloud infrastructure provider offering Kubernetes-native orchestration and managed GPU compute. Broad incumbent that competes with XPerf's on-prem orchestration thesis by offering cloud-native alternatives for enterprise AI workloads.

Emerging players

• Modal Labs: Serverless compute platform for AI/ML workloads with sub-second cold starts and GPU orchestration. Emerging alternative to traditional GPU schedulers with comparable inference optimization focus to XPerf's LoadSteer.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• LinkedInlinkedin.com/company/xperf-inc

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds1 record

Investors1 record

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about XPerf Inc