IBM came out swinging at its annual THINK 2025 event this week, announcing a wave of fresh AI capabilities, deeper partnerships, and robust new infrastructure that signals the company is going all-in on enterprise AI.
From five-minute AI agents to a hybrid cloud revolution, IBM says it’s time to stop dabbling in artificial intelligence and start delivering results.
“The era of AI experimentation is over,” said IBM Chairman and CEO Arvind Krishna during the event. “Today’s competitive advantage comes from purpose-built AI integration that drives measurable business outcomes.”
AI agents that work, not just talk
IBM is positioning its new suite of agent capabilities in Watson Orchestrate as the backbone of the next generation of AI in the workplace. Unlike traditional chatbots, these new agents aren’t just conversational – they’re built to act.
Need an agent to find sales prospects in Salesforce, handle HR questions in Slack, or run procurement workflows in Oracle? IBM says you can build that agent in under five minutes using no-code or pro-code tools. The Watson Orchestrate platform connects with more than 80 business applications, including Microsoft, ServiceNow, and Workday.
To simplify deployment, IBM is also launching an Agent Catalog, featuring more than 150 pre-built agents and tools from partners like Box, Mastercard, and Symplistic.ai.
Oracle + IBM: A new AI power duo
In a major partnership expansion, IBM is teaming up with Oracle to bring Watsonx Orchestrate to Oracle Cloud Infrastructure (OCI) starting in July. This integration aims to streamline how companies use AI across Oracle and non-Oracle apps.
“AI delivers the most impactful value when it works seamlessly across an entire business,” said Greg Pavlik, executive vice president of AI and Data Management Services at Oracle Cloud Infrastructure. “Our expanded partnership will provide customers new ways to help transform their businesses with AI.”
Under the deal, IBM’s Granite AI models will also be available in Oracle’s AI tools, and watsonx.ai is now certified to run on OCI, allowing for AI deployments closer to enterprise data and applications.
Hybrid integration for the AI age
Integration remains one of the most challenging hurdles for enterprise AI. To address this, IBM is introducing webMethods Hybrid Integration, a unified platform that connects APIs, apps, events, files, and even mainframe data across hybrid cloud environments.
SEE: New IBM z17 Mainframe Will ‘Redefine AI at Scale’
IBM’s hybrid integration approach leverages AI to automate the process, from generating integrations with natural language to managing APIs and B2B data flows.
According to a Forrester Consulting study, companies using IBM’s webMethods tools achieved a 176% return on investment (ROI) over three years, cut downtime by 40%, and saw a 67% time savings on simple projects.
Turning unstructured data into AI fuel
IBM’s new watsonx.data updates aim at a persistent enterprise problem: unstructured data. Think contracts, spreadsheets, presentations — all the stuff hiding in folders that traditional systems ignore.
The updated platform introduces data lakehouse architecture, AI-powered insights, and easier connections to AI agents. IBM claims the enhancements can boost AI accuracy by up to 40% compared to conventional retrieval-augmented generation (RAG) methods.
To further this mission, IBM is acquiring DataStax, known for its unstructured data and vector search capabilities, and integrating with Meta’s Llama Stack for open AI deployment.
A new engine for enterprise AI: LinuxONE 5
Behind all these AI tools lies serious computing power. IBM launched LinuxONE 5, its next-generation Linux server built for massive AI workloads, capable of running up to 450 billion inference operations per day.
This powerhouse combines security, energy efficiency, and scalability. IBM says switching from x86 to LinuxONE 5 could cut total cost of ownership by up to 44% over five years. It also includes built-in AI accelerators and quantum-safe encryption, which aims to “address quantum-enabled cybersecurity attacks.”
Agentic AI across the enterprise
From consulting services to integration platforms, IBM is wrapping all of its moves into an “agentic AI” strategy. IBM Consulting is helping customers adopt this model, including migrating legacy workloads to Red Hat OpenShift Virtualization on OCI.
“Agents are becoming an important strategic advantage for organizations to optimize operations and deliver better customer and employee experiences,” said Ritu Jyoti, group VP at IDC, as quoted in the IBM/Oracle press release. “The Agentic AI approach IBM is taking with Oracle is a leading example of how the ability to orchestrate agentic workflows across systems can drive significant advantages.”
