WSO2 Makes AI Workspace Fully Self-Managed for Sovereign AI Governance

Mumbai: WSO2 has made its AI Workspace available as part of its fully self-managed API Platform offering, allowing enterprises and government organisations to run their AI governance infrastructure within their own environments, including fully air-gapped deployments.

The release enables organisations to operate a governance layer over their artificial intelligence infrastructure, covering large language models (LLMs), AI agents, Model Context Protocol (MCP) servers and AI-related costs without requiring a vendor-hosted component.

Governance Within the Customer Environment

With the self-managed version of AI Workspace, organisations can control employee access to external and sovereign LLMs and MCP servers. Internal resources can also be exposed as MCP servers under the same policy controls.

The platform supports cost controls, quotas and chargeback across internal and external AI resources, while applying governance measures across AI traffic routed through the company’s AI gateways.

AI Workspace also provides developers with a central workspace to discover and access organisationally approved AI models and services while keeping the associated infrastructure within the customer’s environment.

Self-Hosted Option for Regulated Organisations

WSO2 said the release addresses governance requirements among regulated enterprises and governments that need to maintain control over AI infrastructure, policy enforcement and audit information.

The company noted that regulatory requirements in Europe, including the EU AI Act, proposed EU Cloud and AI Development Act, DORA and NIS2, are increasing the focus on organisational control over AI systems.

According to WSO2, a vendor-hosted control plane can place policy enforcement, routing decisions and audit trails outside an organisation’s direct infrastructure boundary.

Derric Gilling, Vice President and General Manager, API Platform, WSO2, said sovereignty requirements are increasingly part of discussions with regulated organisations, including financial institutions and government agencies.

He said the self-managed AI Workspace extends the company’s existing SaaS capabilities to a fully self-managed model, giving customers the option of SaaS, hybrid or self-managed deployments.

Deployments Across Different Infrastructure

WSO2 customers could previously deploy the company’s AI gateway within their own environments, while the AI control plane was available through WSO2-operated SaaS.

With the latest release, both the AI gateway and AI control plane can be deployed according to customer requirements, including on-premises environments, national or sovereign clouds and hyperscalers.

The deployment can also operate in a fully air-gapped environment, without an outbound connection to WSO2.

AI Workspace was launched in March 2026 with SaaS and hybrid deployment options. The latest release adds the fully self-managed option, allowing organisations to self-host the AI capabilities available through the API Platform.

WSO2 currently serves 42 national government customers and more than 3,800 local government agencies, in addition to regulated organisations in financial services, healthcare and telecommunications.

Unbundled Platform Approach

The release also continues WSO2’s unbundled platform strategy, allowing customers to adopt individual capabilities such as the AI gateway, API management, identity and integration according to their requirements.

Customers can deploy individual components or combine capabilities as needed. The company also continues to offer an option to deploy the complete platform, including API Manager, as a single binary.