Realtime AI News
IBM launches agent identity and access controls at TechXchange
At TechXchange 2026 in Atlanta, IBM introduced capabilities for discovering, identifying, governing and evaluating AI agents across business systems. HashiCorp Vault now issues short-lived credentials to agents with real-time revocation, while watsonx Orchestrate provides a single view spanning Amazon Bedrock, Microsoft Foundry and Google Gemini.
IBM on Tuesday announced at TechXchange 2026 in Atlanta a set of capabilities aimed at letting developers, IT and security teams put AI agents to work with clear authority and accountability: a verified identity, only the access each agent needs, and one place to discover, secure and evaluate agents wherever they were built.
The company cited Gartner's projection that by 2028 the average Fortune 500 enterprise will run more than 150,000 AI agents, while only 13% of organizations believe they have the right agent governance in place. Persistent credentials, unclear ownership and disconnected controls make agents hard to manage as they invoke tools and act across teams, platforms and clouds.
The centerpiece is an extension of HashiCorp Vault, IBM's secrets manager, to AI agents. Vault already issues and rotates credentials for applications and services; it will now issue secured, short-lived credentials to agents, limiting each one to what it needs and enabling real-time revocation when an agent attempts an unauthorized action.
A companion capability, Agent Identity, establishes the authoritative record for each agent, covering who owns it, its lifecycle status and the authority it has been granted, and shares that context with Vault's credential controls to enforce least-privilege access. IBM said Vault already secures non-human identities at organizations including ABN AMRO, Adobe, BBVA, BlackRock, Canva, the London Stock Exchange, Starbucks, Swiss Federal Railways and Target.
IBM watsonx Orchestrate supplies the cross-platform layer: a single place to discover, contain and evaluate agents built on Amazon Bedrock, Microsoft Foundry and Google Gemini. Working with Vault and Agent Identity, it is designed to define what each agent's job is, govern its authority and audit its actions.
IBM also said Agent Identity and Vault integrate with NVIDIA OpenShell secure runtime, pairing a protected execution environment with verified identity and short-lived credentials so each action can be traced, limited and reviewed. The company additionally introduced IBM Integration Intelligence Hub, built on a Unified Context Engine so agents can reuse existing APIs and integration knowledge, alongside IBM Concert for vulnerability response and Terraform policy, a developer portal and a remote MCP server for governed infrastructure changes.
Availability is mixed. HashiCorp Vault 2.2, Agent Identity, watsonx Orchestrate's evaluation capabilities, IBM Concert and Terraform policy are generally available. Secure Hub, the Terraform Developer Portal, the Terraform remote MCP server and IBM Z Atlas, a z/OS system-intelligence layer, are in public preview, while Integration Intelligence Hub and Orchestrate's discovery and containment features remain in tech preview.
The announcements converge on one premise: enterprises are deploying agents faster than they can verify them. Once agents call APIs, spend credentials and change infrastructure on a company's behalf, identity and permissions stop being peripheral features and become the precondition for scale. The open question is whether governance interfaces that span vendors can hold up in real multi-cloud estates.
Why it matters
IBM is reframing agent governance as a stack of identity, credential, discovery and evaluation controls rather than a single security product, which concedes that trust, not raw capability, is what gates agent scale. If buyers accept these cross-cloud interfaces, procurement criteria will shift from what an agent can do to whether it can be audited and shut off at any moment.
Nearby Updates
All10/27, 10:52
Tencent Hunyuan open-sources Hy-SimulMT, a speech interpreter with about 3 seconds of lag
Tencent Hunyuan has open-sourced Hy-SimulMT, a speech-to-speech interpreter that turns Chinese speech into English and English into Chinese while the speaker is still talking, cloning the speaker's voice in real time. Tencent reports about three seconds of lag on long-form speech using the StreamLAAL metric, and the model is already live in the free Hy Translate app.
10/27, 09:05
OpenAI reports 87% average pass rate across 20 coding languages on new MultiLangBench
OpenAI has published MultiLangBench results showing its GPT-6 Astra model reaching an 87.4% average test pass rate across 20 programming languages, far above the 26.8% scored by GPT-5.6 Sol. A companion evaluation tested COBOL migration into modern languages with byte-for-byte output checks, and OpenAI is releasing the test suite on GitHub.
10/27, 06:55
GammaDX launches agent experience practice as its index finds only 12 of 50 Australian sites agent-ready
Australian digital experience agency GammaDX has launched a dedicated agent experience design practice alongside its first Gamma AXD Index, which scored 50 Australian websites across five sectors on a 100-point scale. Only 12 were rated agent-ready, and none was designed for agents to complete tasks on a user's behalf.
10/27, 06:52
Ollama for Mac adds support for local text-to-speech models
Ollama's Mac app now supports local text-to-speech models, 9to5Mac reports, extending the runtime from generating text to speaking it. Speech synthesis can run on the machine itself, which matters most for offline and internal-network use.