IBM CEO Arvind Krishna said there is no law that enterprise AI has to be expensive, pegged to large language models and experimental.
In a keynote at Think 2025, Krishna outlined IBM's strategy and rollout of tools to build AI agents quickly and then orchestrate them.
"I think that the era of AI experimentation is over. Success is going to be defined by integration and business outcomes," said Krishna. "There is no law of computer science that says that AI must remain expensive and must remain large."
That theme played a big part of the news flow from Think 2025. At a high level:
- IBM's unified platform for building and managing AI agents, watsonx Orchestrate, will offer pre-built agents with use cases for industries with skills, integration and capabilities in HR, sales and procurement. There are also tools to build agents, orchestrate them and monitor and optimize with observability tools. Krishna said Watson Orchestrate Agent Builder can enable business users to build their own agent in 5 minutes.
- IBM launched next-generation Granite models. The company launched Granite 4.0 Mixture-of-Expert models and a new runtime engine.
- Watsonx.data will get new integration features to orchestrate data access and engineering in one interface and data governance software to curate, manage and use data.
- IBM announced GenAI Lakehouse to unify, prepare and govern structured and unstructured data for generative AI.
- The company launched IBM LinuxONE Emperor 5, IBM's next-gen Linux platform.
- A partnership with Lumen Technologies to bring IBM's portfolio of AI products and models to Lumen's Edge Cloud network.
- Expanded partnerships on AI agents with AWS, Oracle and Salesforce.
Krishna featured interviews with customers such as the Ferrari race team, Lumen and PepsiCo on managing data and AI agents.
- IBM launches Flex Plan for quantum computing, aims to expand use cases
- IBM maintains Q2 outlook, reports better-than-expected Q1
- IBM launches z17 mainframe, eyes AI workloads
For IBM, the company is trying to leverage horizontal and business expertise via its consulting unit and integration knowhow. Krishna is also talking business value given that projects are moving out of proof-of-concept mode.
Here’s a look at some of the key points from Krishna's keynote:
- "As I think about last year compared to this year, the one big difference is that AI has moved from experimentation to business value. What is the use case? How do I get my business to scale leveraging AI? We're thinking about ROI. We're thinking about business value," said Krishna.
- Success is defined by integration, intelligence and automation in enterprise workflows. AI is an accelerant, said Krishna.
- Smaller, special purpose models will be critical to drive more cost effective AI, he said.
- "IBM is incredibly focused on enterprise AI. We do not focus at all on consumer AI, and that is why you see our focus on purpose built models, cost, effectiveness, sovereignty, and security. How do we make and decrease the cost of enabling AI inside the enterprise?" said Krishna.
- Speed of inference will depend on hybrid infrastructure that brings data closer to the edge.
- "Quantum is no longer science fiction. It's now in the realm of engineering. And when you're in the realm of engineering, the question becomes, how do you get that next step? And then these systems are going to be at a scale that is going to be truly remarkable and truly surprising. We are building these full stack quantum systems," said Krishna.
