IBM took advantage of its Think 2026 conference to introduce an expansion of its Artificial Intelligence and hybrid cloud portfolio, with a message directed at large enterprises: deploying more models won’t be enough if there’s no organized way to create, coordinate, govern, and secure AI agents in production.
The company speaks of a new “operating model” for the agentic enterprise, based on four interconnected pillars: agents, data, automation, and hybrid infrastructure. Behind this announcement is a clear market insight. Many organizations have invested in AI, but still face difficulties translating those projects into sustainable results, especially as pilots begin to handle sensitive data, critical systems, and real business processes.
Arvind Krishna, IBM’s Chairman and CEO, summarized the approach by stating that the companies making progress are not necessarily those deploying the most AI, but those rethinking how their business operates. With this idea, IBM aims to position itself in a layer of management and governance, rather than merely providing models.
Watsonx Orchestrate as a control plane for agents
One of the key announcements is the new generation of IBM watsonx Orchestrate, now available in private preview. IBM aims to turn this tool into a control plane for multi-agent environments, where organizations can deploy agents created from various platforms, enforce consistent policies, and maintain traceability over their activities.
The challenge is significant. As companies move from testing a few assistants to managing hundreds or thousands of agents, tough questions arise: who created them, what permissions do they have, which data do they operate on, how are their decisions audited, and what happens if they make mistakes? IBM seeks to address this need with a unified orchestration layer.
Alongside watsonx Orchestrate, the company highlights IBM Bob, now generally available. This is an agentic development assistant designed for enterprise environments, with integrated security controls and cost management. The solution targets development teams wanting to create agents without losing control over risks related to code, data, and resource consumption.
Real-time data for AI to act
The second focus of the announcement revolves around data. IBM asserts that agentic Artificial Intelligence requires up-to-date, connected, and governed information to act reliably. To support this, the company integrates capabilities from Confluent, based on Kafka and Flink technologies, with watsonx.data.
Among the new features is Context in watsonx.data, now in private preview, which adds a federated context layer allowing AI systems to reason over business data with semantic meaning, runtime governance, and more explainable decisions. IBM also announced integrations of Confluent and Tableflow with watsonx.data to unify streaming data and batch loads in hybrid environments.
IBM has also demonstrated progress in watsonx.data with GPU-accelerated Presto, still in private preview. According to the company, a proof of concept with Nestlé achieved cost savings of 83% and a 30x improvement in price-performance ratio over a global data mart spanning 186 countries. As usual, IBM cautions that actual results depend on individual configurations and environments.
IBM Concert and intelligent operations management
The third aspect of the launch is IBM Concert, now in public preview. This platform aims to help organizations transition from passive monitoring to coordinated operational responses in IT. Rather than replacing existing tools, IBM proposes a layer that correlates signals from applications, infrastructure, and networks to provide a shared view of risks, dependencies, and recommended actions.
This approach aligns with a common scenario in large companies: hybrid infrastructures are typically managed with separate teams and tools, forcing staff to act as intermediaries between systems. When AI is integrated into critical processes, this complexity increases. Concert seeks to reduce this noise and improve information flow into execution.
Within this same line, IBM introduced IBM Concert Secure Coder, also in public preview. This feature brings security management into developers’ workflows, available in IBM Bob and Visual Studio Code. It aims to identify and prioritize risks during code development, while automatically generating fixes for vulnerabilities in code, operating systems, middleware, packages, and images.
IBM completes its automation suite with Vault 2.0, now generally available, and IBM zSecure Secret Manager, expected in June 2026. Both address a growing challenge: managing secrets, credentials, and certificates in increasingly distributed environments.
Operational sovereignty for regulated environments
The fourth pillar of Think 2026 is IBM Sovereign Core, now generally available. The platform is designed for organizations that need to operate AI and critical workloads under strict sovereignty, control, and compliance requirements.
IBM Sovereign Core integrates policies at the infrastructure runtime level and leverages open technologies like Red Hat OpenShift and Red Hat AI. The platform includes an expandable catalog that organizations can customize with their own applications, IBM software, third-party solutions, and open source components. Partners mentioned by IBM include AMD, ATOS, Cegeka, Cloudera, Dell, Elastic, HCL, Intel, Mistral, MongoDB, and Palo Alto Networks.
The solution is particularly targeted at regulated sectors, government agencies, banking, healthcare, telecommunications, and critical infrastructures, where AI can’t rely solely on flexible configurations. In these contexts, data location, workload portability, auditing, and operational independence are just as critical as model performance.
Think 2026 thus conveys a fundamental message: IBM wants to compete in the layer where enterprise AI becomes complex. Not just in creating agents or accessing models, but in managing data, security, operations, compliance, and hybrid cloud. The promise is appealing for large organizations, but success will depend on how well these components can be integrated into real-world environments, with reduced complexity and without turning governance into an additional burden for technical teams.
Frequently Asked Questions
What did IBM announce at Think 2026?
IBM introduced new capabilities for agent orchestration, real-time data, intelligent operations, development security, and operational sovereignty in hybrid environments.
What role does watsonx Orchestrate play in IBM’s strategy?
Watsonx Orchestrate is designed as a control plane to manage AI agents from various sources, ensuring policies, traceability, and governance are consistent across the board.
Why does IBM emphasize hybrid cloud and sovereignty?
Because many companies handle sensitive data, critical systems, and regulatory requirements that demand strict control over where AI runs, how it is governed, and how decisions are audited.
via: newsroom.ibm