Terminal 3 and Affinidi have launched TENON, a public trust network designed to verify, govern and secure AI agents operating across organizations. Unveiled at the Singapore FinTech Festival on Nov. 19, 2026, the network combines confidential computing, agent identity verification, programmable policy controls and privacy-preserving execution to enable autonomous AI systems to handle sensitive enterprise workflows, initially targeting financial services.
Terminal 3 and Affinidi Build Public Trust Infrastructure for AI Agents
Enterprise AI adoption is expanding beyond chatbots and copilots toward autonomous agents capable of executing transactions, accessing enterprise systems and coordinating tasks across organizations. However, granting these systems access to sensitive data, credentials and financial workflows remains a security and governance challenge.
Terminal 3 and Affinidi are addressing that challenge with TENON, a public network designed to verify agent identities, enforce permissions, secure execution and create auditable records of AI-driven actions across organizational boundaries.
The companies introduced TENON at the Singapore FinTech Festival, with financial services as its initial target market. Terminal 3 and Affinidi will operate the network’s first nodes, with plans to expand participation to other organizations and sectors.
A trust layer for autonomous AI operations
TENON is designed to address a gap in enterprise AI infrastructure: organizations often lack a consistent way to establish which AI agent is operating, what authority it possesses and whether its actions comply with organizational policies when interacting with external systems.
The network applies governance and security controls to agentic calls. It verifies agent identity and authority against trusted registries based on open standards, then uses a programmable policy engine to enforce permissions and organizational requirements.
TENON also routes requests to appropriate AI models, applies guardrails and cost controls, and replaces sensitive information in prompts where required. Workflows involving protected data or credentials can be executed inside cryptographic enclaves designed to protect data during processing.
The architecture combines Affinidi’s open-source agent governance technologies and Verifiable Trust Infrastructure with Terminal 3’s confidential computing and hardware-secured agent security platform. Each agent call and action is recorded in a private, immutable audit trail, while participating organizations’ data is isolated from other network operators.
The design aims to separate an AI agent’s ability to reason from its access to protected systems, allowing organizations to authorize specific operations without exposing sensitive databases, credentials or information directly to the agent.
Financial services is the initial use case
The launch reflects growing concerns among financial institutions about deploying autonomous systems across regulated environments. The announcement references United Overseas Bank’s October 2026 Agentic Finance report, which highlights the lack of a broadly accepted framework for verifying autonomous software operating across institutions, platforms and jurisdictions.
Financial services provides a demanding use case for agentic AI because workflows may involve payments, customer information, fraud investigations, compliance and financial transactions. Agent identity, authorization, auditability and data protection become essential when AI systems interact with these processes.
TENON is designed to help enterprises and governments deploy agents for sensitive data analysis, regulated information handling and high-value transactions, while supporting governance requirements. The companies also position the network as relevant to supervisory expectations, including the Monetary Authority of Singapore’s October 2026 Guidelines on AI Risk Management referenced in the announcement.
At the Singapore FinTech Festival, the companies demonstrated an AI banking assistant handling overseas transfers and disputed credit card charges without exposing customers’ account details to the underlying agent.
Agent Command will be the first application
Terminal 3’s Agent Command platform will be the first application built on TENON. The platform is designed for enterprise and government agent governance, security and compliance, routing inference, security and orchestration through the network.
Protected data, credentials and confidential execution will remain within each organization’s cloud or on-premises environment, according to the companies. This approach aims to combine network-level verification with localized control over sensitive information.
TENON will also contribute improvements to Affinidi’s open-source Agent Gateway and Agent Stream projects, which are part of Affinidi Trust Fabric and already integrated into Terminal 3’s enterprise platform.
The companies intend to expand TENON’s node operator ecosystem across finance, insurance, healthcare, logistics, technology and government.
Agentic AI infrastructure moves toward interoperability
TENON’s approach reflects an emerging direction in enterprise AI infrastructure, where agent identity, execution security, observability and policy enforcement are becoming distinct layers in the technology stack.
AI agent frameworks and orchestration platforms increasingly enable autonomous systems to interact with APIs, tools and enterprise data. However, cross-company deployments introduce additional challenges involving trust, permissions, data exposure, compliance and accountability.
Existing enterprise security systems typically enforce access controls within individual organizations. TENON’s public network model seeks to extend verification and governance across organizational boundaries, potentially enabling agents from different enterprises to interact through shared trust mechanisms.
The initiative also intersects with broader AI infrastructure developments involving Microsoft, Google, Amazon and NVIDIA, whose ecosystems support enterprise AI deployment, model infrastructure, cloud computing and AI development. TENON, however, focuses on the trust, governance and confidential execution layer rather than competing directly as a general-purpose AI model platform.
The network’s long-term value will depend on adoption, interoperability, security validation, governance standards and the willingness of organizations to operate nodes and trust shared infrastructure. A public trust network could reduce the complexity of establishing agent-to-agent interactions, but regulated enterprises will still need to assess deployment risks, operational controls and jurisdiction-specific requirements.
For Terminal 3 and Affinidi, TENON represents an effort to make verifiable AI agent operations a shared infrastructure capability rather than a security function confined to individual enterprise platforms.
Market Landscape
The enterprise AI agent market is evolving from experimentation with autonomous assistants toward agentic systems capable of executing business workflows, accessing enterprise resources and coordinating tasks across multiple platforms.
This shift is increasing demand for infrastructure that can manage agent identity, authentication, authorization, execution boundaries, audit trails and data privacy.
TENON enters an emerging market segment that intersects with:
- AI agent governance: Enforcing policies, permissions and accountability for autonomous systems.
- Confidential computing: Protecting sensitive data and workloads during execution.
- AI security: Controlling agent access to enterprise systems, credentials and APIs.
- Verifiable identity: Establishing trusted identities and authority for autonomous agents.
- Enterprise AI orchestration: Connecting agents, models, tools and workflows across organizational environments.
The competitive landscape includes enterprise AI governance platforms, cloud security providers, confidential computing vendors, identity infrastructure providers and AI orchestration platforms.
TENON’s differentiation is its public network model, which combines open-source agent governance technologies with confidential computing and cross-organizational trust mechanisms. Adoption will depend on interoperability, trust in node operators, governance standards and enterprise security validation.
Top Insights
- TENON combines confidential computing, agent identity verification and programmable governance to secure autonomous AI workflows across organizational boundaries.
- Financial services is the network’s initial target, with applications spanning transactions, data analysis and regulated enterprise workflows.
- Terminal 3’s Agent Command will become the first application on TENON, integrating agent governance, security and orchestration.
- The network combines Affinidi’s open-source Agent Gateway and Agent Stream with Terminal 3’s confidential computing infrastructure.
- TENON’s expansion will depend on enterprise adoption, interoperability, node operator participation and trust in cross-organizational AI governance.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
