Start with the object it controls. Then ask what remains outside.
Compute & factory infrastructure NVIDIA Run:ai / DSX Commercial infrastructure control plane + open scheduler Turns scarce GPU clusters into schedulable, quota-governed AI capacity; DSX broadens the lens to the data-center AI-factory software stack. Established product Compute & factory infrastructure HPE AI Factory / Private Cloud AI Integrated hardware, software, services and partner stack A packaged route to private/sovereign GPU infrastructure with HPE hardware, NVIDIA software and deployment/support services. Established product ML/data lifecycle Iguazio + MLRun / Nuclio Commercial platform plus distinct open-source projects MLRun composes data preparation, pipelines, training, deployment and monitoring on Kubernetes; Nuclio supplies event/serverless execution for real-time paths. Open components + paid platform ML/data lifecycle ClearML Open-source MLOps stack + commercial platform/services A broad, inspectable AI-workload stack: experiment/data tracking, pipelines, agents/workers, orchestration, serving and self-hosting options. Open components + paid platform Agent build & run DataRobot Agent Workforce Platform Commercial enterprise AI platform + extensive public developer ecosystem A particularly broad commercial control plane joining predictive ML, GenAI and agents with registry, deployment, monitoring and governance across managed, private and self-managed environments. Open components + paid platform ML/data lifecycle Dataiku Commercial enterprise platform with public SDK/plugin ecosystem Collaborative data science and analytics across visual and code workflows, with governed projects, model deployment, monitoring, stress testing and external MLOps integrations. Established product ML/data lifecycle Domino Data Lab Commercial enterprise AI platform Governed, collaborative data science and model lifecycle across hybrid/multicloud/on-prem compute, model registries, approvals and reproducible environments. Established product ML/data lifecycle H2O AI Cloud / MLOps Commercial platform around an extensive OSS model ecosystem Connects model development/AutoML, notebooks, pipelines, registry, deployment and monitoring into an enterprise AI app platform. Established product ML/data lifecycle Valohai Commercial MLOps orchestration with open client utilities Makes reproducible model/data workflows portable across cloud, on-prem, Kubernetes, Slurm and existing tools using versioned executions and configuration. Established product Compute & factory infrastructure VESSL AI Cloud Managed AI cloud with usage-based commercial service Provides GPU-backed workspaces, batch jobs, storage and inference APIs for teams that need AI compute without first building the substrate. Established product Agent build & run Cake AI Platform Commercial managed integration/operations around OSS components Offers a managed AI/ML stack story spanning infrastructure, extraction, observability, FinOps and AIOps, while letting customers compose open components. First-party claim Agent build & run Katonic AI Factory / AI Cloud Commercial sovereign AI platform; separate enterprise/MSP products The broadest product claim in this cohort: employee AI Workroom, developer Studio and IT Control Room, with model/tool gateway, agents, governance, GPU controls and a separate multi-org AI Cloud/MSP surface. First-party claim Agent build & run LangSmith Engineering Platform Commercial platform with separate open-source frameworks A framework-agnostic agent development loop joining traces, evaluation, production failure clustering, deploy/runtime and agent fleet management; supports SaaS, BYOC and self-hosting claims. Established product Agent build & run Oracle AI Database Private Agent Factory Commercial/private deployment, Oracle data-centered A deployable package for building, testing, governing and operating private agents close to Oracle Database and enterprise documents/APIs, with OCI/on-prem/multicloud paths claimed. Established product Enterprise AI application platform C3 AI Agentic AI Platform Commercial integrated enterprise data/AI application platform Builds domain AI applications and agents over a shared enterprise data/ontology model, with ingestion, models, workflows, APIs, apps and operational monitoring under one vendor platform. Established product Agent build & run deepset / Haystack Open framework plus commercial enterprise platform A composable open-source framework for RAG/agent pipelines, with a commercial platform positioned to build, run and govern AI applications in controlled environments. Open components + paid platform AI governance & observation ModelOp Enterprise AI Command Center Commercial cross-vendor AI governance/system of record Tracks internal and vendor AI assets through intake, risk reviews, approval, lifecycle state, costs and retirement across other development/execution tools. First-party claim AI governance & observation IBM watsonx.governance Commercial AI governance / model-risk platform AI asset visibility, model cards/factsheets, risk and regulatory workflow, monitoring and governance over IBM and some external model platforms. Established product AI governance & observation Arthur AI Commercial AI/agent observability, evaluation and governance Collects inference/agent behavior, runs policy/evaluation checks, surfaces anomalies and connects AI inventory/monitoring to human governance workflows. First-party claim Services-led foundry Infosys Agentic Foundry / Topaz Systems-integrator delivery and reusable platform components Offers industry blueprints, reusable agents, evaluation frameworks, AgentOps and MCP management as a cross-platform delivery package. First-party claim Services-led foundry Accenture AI Refinery Partner-powered consulting and industry-agent delivery Packages industry-specific agents and implementation expertise on NVIDIA AI Enterprise/Blueprints and partner/hyperscaler stacks. First-party claim Enterprise work-execution platform OpenAI Frontier Commercial enterprise platform for AI coworkers and cross-system business processes Combines shared business context, agent execution, evaluation/improvement loops, enterprise identity, permissions and auditable actions to deploy agents into business workflows. First-party claim Services-led foundry OpenAI Frontier Alliances Frontier deployment and systems-integration partner network Adds enterprise architecture, workflow redesign, data/system integration, implementation capacity and change management around the Frontier product. First-party claim Enterprise work-execution platform Salesforce Agentforce 360 + MuleSoft Agent Fabric Commercial business-agent platform with cross-system orchestration and governance claims Builds agents on Salesforce business data/metadata and can orchestrate workflows, APIs and third-party agents through MuleSoft and open protocols. Established product Agent fleet governance ServiceNow AI Control Tower Commercial discovery, governance, security and measurement layer for enterprise AI Aims to discover, observe, govern, secure and measure AI systems, agents and workflows across ServiceNow and external platforms, with integrations and an AI Gateway. First-party claim Enterprise work-execution platform ServiceNow Now Assist agentic workflows Commercial agents embedded in IT, customer, employee and creator workflows Uses ServiceNow's workflow/service-management substrate for incident context, ITSM agent assistance, and bounded agentic workflows that can investigate, remediate or analyze routine but variable procedures. First-party claim Agent fleet governance Workday Agent System of Record + Agent Gateway Commercial inventory, analytics, lifecycle and governance for human-and-agent workforce Treats AI agents as managed workforce members: register, configure, activate, deactivate and measure them alongside people; Agent Gateway claims to govern and meter first- and third-party agent interactions involving Workday data. First-party claim Agent build & run SAP Joule Studio and Joule Agents Commercial low-code/pro-code agents embedded in SAP business processes Builds or customizes agents within SAP's business application and process context, where identity, business objects and workflow actions are already modeled by SAP products. First-party claim Agent build & run Oracle Fusion AI Agent Studio Commercial agent builder integrated into Oracle Fusion Cloud Applications Creates, extends, validates and deploys agents with direct access to Fusion application knowledge stores, tools and APIs inside business workflows. Established product Agent build & run Microsoft Copilot Studio Commercial low-code/pro-code agent builder in the Power Platform / Microsoft 365 ecosystem Builds and publishes agents with Microsoft data, connectors, actions, environment ALM and admin controls; the developer surface is tightly integrated with Microsoft identity and Power Platform. Established product Agent fleet governance IBM watsonx Orchestrate agent control plane Commercial enterprise agent orchestration and workforce management platform Provides a centralized layer to observe, govern, evaluate, optimize and orchestrate IBM and external agents across different build/run environments, with catalog, identity, audit and policy claims. First-party claim Agent fleet governance Microsoft Agent 365 Commercial agent registry, identity and tenant-administration control plane Extends Microsoft tenant management to agents: registry, Entra Agent ID, lifecycle policies, access controls, security integrations and organization-wide inventory claims across native and registered third-party agents. Established product Agent fleet governance Microsoft Foundry Control Plane Commercial developer/AI-engineering observation and runtime control plane Adds end-to-end agent traces, pre/continuous evaluations, runtime guardrails, routing and fleet operations across Foundry and registered external agents; OTel traces and AI Gateway are explicit interoperability paths. Established product Enterprise work-execution platform Glean Enterprise AI / Agent Development Lifecycle Commercial enterprise context, search and cross-app agent work platform Combines permission-aware enterprise search/context with employee-facing agents, scheduled/event triggers, app actions, builder traces, sandboxed execution and a seven-stage Opportunity→Monitor lifecycle. This is a context-plus-work execution system, not merely a coding factory. Established product Enterprise work-execution platform Ema AI Employees / Builder Commercial configurable workflow/agent platform sold as AI employees A unusually inspectable documented workflow shape: an AI Employee is a versioned DAG of typed nodes; reasoning agents can pause for human approval; workflows use knowledge bases, integrations, tools, schedules, APIs and multi-turn sessions. Current docs expose tests, monitoring, admin, RBAC, tenant isolation and audit-log concepts. Established product Enterprise work-execution platform UiPath Agentic Automation / Maestro Commercial deterministic, agentic and human work orchestration with RPA Uniquely bridges existing robotic automation, APIs, AI agents and human decisions in durable, stateful business workflows (Maestro claims minutes-to-months), with exception routing, Action Center/HITL, process mining and transaction-level audit concepts. Established product Agent security & access control Entro Agent & Non-Human Identity Security Commercial NHI/secret discovery, posture, behavior and lifecycle security Connects agent and NHI discovery to ownership attribution, secret lineage, privilege/usage posture, anomaly detection, remediation and decommissioning—an important security organ a factory/workforce platform often does not own. First-party claim Agent security & access control Microsoft Agent Governance Toolkit (OSS) Open-source multi-language policy, identity, gateway, runtime and reliability toolkit A rare OSS attempt to span policy enforcement, agent identity/trust, MCP gateway checks, audit, sandbox/runtime, discovery and agent-SRE hooks with Python/TypeScript/.NET/Go/Rust packages and adapters for many agent frameworks. Open components + paid platform Agent security & access control Cisco AI Defense OSS security stack Open scanners, policy/evidence tooling, AI supply-chain and selected enterprise integrations A practical OSS stack for scanning agent skills/plugins/MCP/A2A surfaces, producing AI BOM/model provenance evidence and pairing selected findings with a runtime enforcement gateway (DefenseClaw) and Cisco's commercial AI Defense service. Open components + paid platform Agent security & access control Palo Alto Prisma AIRS agent security Commercial AI runtime/security platform with public SDKs and deployment examples Bridges AI-specific traffic inspection (prompt/response/data flow) with AI asset/agent risk and security control; its public perimeter/firewall and Python API surfaces offer some inspectable integration edges. Established product Agent security & access control Assury Runtime / Lifecycle Management Early commercial agent runtime-governance and inventory claims Worth tracking as a quiet provider specifically claiming in-path deterministic MCP policy, agent discovery/lifecycle, session tracing/replay and behavioral drift visibility—areas underserved by generic coding factories. Discovery lead New discovery lead Jentic Sovereign agent operating system / collaboration platform claim Worth screening because its positioning explicitly targets enterprise-owned agents, models and cloud/on-prem execution rather than a single model vendor. Discovery lead New discovery lead Sovereign AI / SovAI Enterprise AI work-platform and governance claims A relevant search hit because it explicitly pitches an in-boundary AI work platform, model gateway, spend metering and audit receipts. Discovery lead New discovery lead GoSovern Sovereign agent platform claim Potentially adjacent to Katonic/Cake: claims to connect enterprise data, build agents and operate/gov them inside a customer's network. Discovery lead New discovery lead ATAILA Sovereign cloud + AI-built software delivery claim A high-value adjacency because it claims the full chain from AI-assisted app building through governed release and continued operation on European private compute. Discovery lead ML/data lifecycle Anaconda + Outerbounds / Metaflow Commercial enterprise platform + open-source workflow framework Production-oriented Python data/ML workflows with versioned steps/artifacts, orchestration and compute portability; acquisition may connect this to Anaconda's package/security/environment platform. Open components + paid platform ML/data lifecycle JFrog ML / Qwak lineage Commercial MLOps within software supply-chain platform Connects model build/train/test/deploy/serve/monitor with software artifact registries, security scanning and DevSecOps controls—a useful bridge between AI lifecycle and software supply chain. Established product ML/data lifecycle MLflow Open-source AI/ML lifecycle project + managed commercial offerings Common open foundation for experiments, models, registry, evaluation and increasingly LLM/agent tracing and monitoring; often the substrate beneath enterprise platforms. Open components + paid platform ML/data lifecycle Flyte / Union AI Open-source durable workflow platform + commercial control plane Typed, versioned, reproducible workflow execution for data and ML; Flyte 2 emphasizes durable execution while Union sells hosted/enterprise operation and developer productivity. Open components + paid platform ML/data lifecycle Red Hat OpenShift AI / Open Data Hub Supported enterprise Kubernetes AI distribution with public component repos An enterprise-supported combination of notebooks, pipelines, training, model serving, model registry, distributed workloads and evaluation on Kubernetes. Open components + paid platform Agent build & run Microsoft Foundry Commercial AI app/agent factory in Azure Explicit build/optimize/govern lifecycle for AI apps and agents, including hosted runtimes, templates, agent service, knowledge, observability/evaluation, and an agent control plane. Established product Agent build & run Google Gemini Enterprise Agent Platform Commercial cloud AI/agent platform A wide service surface spanning model access, notebooks, Agent Development Kit, managed sandboxes, sessions/memory, registry, identity/gateway, evals, tracing and optimization. Established product ML/data lifecycle Amazon SageMaker Unified Studio Commercial AWS data, analytics and ML workbench Unified AWS workbench for data, analytics, machine learning, model development and governance across AWS data services. Established product Agent build & run Amazon Bedrock AgentCore Commercial managed agent runtime and operations services Provides managed agent runtime/session isolation, identity, gateway, memory, evaluations and observability as separable services for agents built with chosen frameworks. Established product Enterprise AI application platform Palantir Foundry / AIP / Apollo Commercial operational application and AI platform An unusually broad operational-application environment joining ontology, governed data, workflows, agent/model integration, app building and deployment/release management. Established product ML/data lifecycle Databricks Mosaic AI / Unity AI Governance Commercial data/AI platform with open-source project roots Builds AI systems alongside governed lakehouse data, model serving, experiments, vector search, agent tooling and model/agent traffic controls. Established product Enterprise AI application platform Snowflake Cortex AI Commercial warehouse-native AI/agent services Lets teams build and operate AI features close to governed Snowflake data, with model functions, agents, search and a Cortex AI Gateway for routing/controls. Established product ML/data lifecycle Cloudera AI Commercial hybrid data and AI platform Hybrid data platform with workbench, ML pipelines, model serving and governance close to private/on-prem and cloud enterprise data. Established product Compute & factory infrastructure Rafay AI / Kubernetes Control Plane Commercial multitenant Kubernetes/AI infrastructure control plane Turns Kubernetes, VMs, Slurm, bare metal and GPUs into managed multi-tenant self-service environments with catalog, provisioning and lifecycle controls. Established product Compute & factory infrastructure VMware Cloud Foundation AI Factory Commercial private-AI infrastructure and runtime stack Automates private AI infrastructure from hardware/VCF Kubernetes through GPU services, model serving and tenancy; Tanzu Agent Foundations is a related agent app runtime/security layer. Established product Compute & factory infrastructure CoreWeave AI Cloud Commercial GPU cloud and AI infrastructure Supplies GPU cloud capacity and Kubernetes/HPC infrastructure for training and inference without building a private GPU plant. Established product Compute & factory infrastructure Nebius AI Cloud Commercial AI cloud/compute provider with open SDKs and operators Provides AI compute and services, with public Kubernetes/Slurm operators, SDKs and solution examples that improve the inspectable edges. Established product New discovery lead Galene Private AI Sovereign/private AI stack claim A relevant sovereign-AI lead reportedly layering proprietary platform software over open infrastructure for private/on-prem deployment. Discovery lead New discovery lead Nscale AI Cloud AI infrastructure / GPU cloud provider Worth classifying because AI neoclouds sell the capacity layer for model training, inference and enterprise AI factories. Discovery lead Agent fleet governance SAP LeanIX AI Agent Hub Commercial agent inventory and enterprise-architecture context Makes agents, models and MCP servers visible as architecture-managed assets, with discovery inboxes, ownership/context links, APIs and a LeanIX MCP surface for querying and managing the estate. Established product Agent connectivity & action gateway Boomi Agent Control Plane Commercial integration-platform control layer for agents, APIs and MCP Leverages Boomi's iPaaS/API estate to connect agents to enterprise systems, publish governed MCP tools, provide enterprise context, track usage and coordinate agent lifecycle controls. Established product Agent connectivity & action gateway Gravitee AI Agent Management / AI Gateway Open API-management core with separately licensed AI/agent capabilities Turns existing APIs into governed MCP tools and mediates LLM, MCP and A2A traffic through a shared gateway; the agent catalog, policy, OAuth mediation, rate limits and telemetry extend the API-management boundary to agent interactions. Open components + paid platform Agent connectivity & action gateway Runlayer AI Control Plane Commercial governance, MCP gateway and hosted agent runtime A practical control surface for employee-facing AI adoption: approved clients and skills, shadow-use discovery, MCP routing, scoped credentials, inline checks, usage/cost visibility and hosted recurring agents. Established product Agent security & access control 1Password Unified Access Commercial secrets, SaaS visibility, credential brokering and privileged access Closes the identity/credential seam around agents: discover endpoint AI and exposed secrets, vault credentials, broker scoped access to agent and machine workloads, and layer in just-in-time privileged access and audit. Established product Agent security & access control Broadcom AgentMinder Newly announced commercial agent identity and runtime-enforcement gateway Targets per-action authorization at the boundary: validate agent identity and declared mission/context/risk, mediate tool calls, and emit OpenTelemetry-based session/action evidence. First-party claim Durable process & agent orchestration Camunda Agentic Process Orchestration Commercial process engine with source-visible, non-OSI production licensing Combines deterministic BPMN/DMN, human tasks, connectors, and agents in one long-running process, with retries, timers, compensation, versioning, escalation and instance-level operations. Established product Durable process & agent orchestration Temporal Durable Execution Open-source workflow service with commercial Temporal Cloud Persists workflow progress through crashes, outages, timers, retries and human waits; LLM/model and tool calls can be isolated as activities while deterministic code controls the outer lifecycle. Open components + paid platform Durable process & agent orchestration Restate Durable Execution Source-visible durable execution platform under BSL 1.1 Code-first durable services, workflows, stateful actors, timers and reliable messaging; its official examples cover AI agents, LLM calls, MCP and A2A, with SDKs for TypeScript, Java/Kotlin, Python, Go and Rust. Open components + paid platform New discovery lead Opstral Operations Intelligence Commercial cross-domain AIOps / autonomous operations platform The most directly relevant newly surfaced Day-2 claim: federates telemetry, topology, ITSM and operational knowledge; uses operator-authored RCA trees; executes governed procedures through action tickets; and claims independent post-checks, rollback and recurrence feedback across NOC/SRE/IT domains. Discovery lead