IBM TechXchange 2025: The Real Headliner is Data, Not AI

Analyst(s): Brad Shimmin
Publication Date: October 10, 2025

IBM TechXchange 2025 focused on the critical role of data in operationalizing enterprise AI, shifting the narrative from AI models to the underlying data infrastructure. The company is leveraging partnerships and new tools to connect powerful AI models safely and effectively with proprietary enterprise data.

What is Covered in this Article:

• IBM strengthened its pivot from AI model experimentation to operationalizing the underlying data infrastructure required for enterprise AI.
• IBM is advancing “AgentOps” as a new data governance paradigm for the metadata and logs generated and consumed by AI agents.
• Key announcements for watsonx.data and Db2, including the new Developer Edition and agentic data capabilities.
• The introduction of Project Infograph has implications for treating infrastructure metadata as a first-class, queryable data asset.

The Event: A Pivot from Experimentation to Operationalization – At TechXchange 2025 in Orlando, IBM marshaled its product and strategy teams to deliver a singular message: the era of AI experimentation is over. While headlines focused on “agentic AI,” a strategic partnership with Anthropic, and a new developer IDE named Project Bob, the more substantive story for the enterprise lies one layer deeper. Beneath the surface glamour of AI agents, the undisputed hero of this event was data.

IBM’s announcements collectively target the most significant barrier to AI adoption, namely the chaotic state of enterprise data. This aligns with recent Futurum Group findings that data quality, trust, and governance remain elusive for most organizations.

Figure 1: Data Quality, Trust, and Governance Sit Atop Enterprise Practitioner Complaints

IBM is making a decisive bet that the winners in the AI race won’t be defined by who has the best model, but instead by who commands the most well-architected, governed, and accessible data ecosystem.

This data-first strategy permeates the portfolio. Enhancements to the watsonx platform are less about the AI models themselves and more about making data AI-ready. Agentic Data Integration aims to use LLMs to simplify data pipeline creation, while Agentic Data Intelligence promises to automate data discovery and lineage. Likewise, the OEM integration with Unstructured.io directly tackles the messy reality of enterprise documents. And crucially, the launch of a free, local watsonx.data Developer Edition is a clear play to embed the core data platform with the developers building the next wave of data-intensive applications. These are not siloed features, but a concerted effort to create a cohesive data foundation for AI.

IBM TechXchange 2025: The Real Headliner is Data, Not AI

Analyst Take: Mastering the “Picks and Shovels” – IBM executing a classic playbook well-aligned with its strengths. The company is leveraging the excitement around shiny new objects like AI agents to sell the picks and shovels (e.g., data platforms and governance) required to make them work at scale. The TechXchange announcements underscore a mature understanding that AI value is not generated by a model in isolation, but by its deep and reliable connection to proprietary data.

Data Readiness, Not Model Choice, is the New Competitive Moat

The Anthropic partnership is instructive. While access to the Claude model family is significant, the enterprise value lies in the frameworks IBM is building around (and beneath) it. IBM offerings like the “Architecting Secure Enterprise AI Agents with MCP” guide and the Project Bob integration are designed to do one thing: safely connect a powerful model to an enterprise’s unique data. This is a massive bet on Retrieval-Augmented Generation (RAG). The value isn’t a generic model that knows COBOL. It’s a model grounded in an internal corpus of 40 years of COBOL documentation, support tickets, and codebase annotations — all managed internally by IBM. The new OpenRAG stack and Unstructured.io integration are the technical underpinnings of this strategy, creating robust pipelines from internal knowledge chaos to AI-ready context.

“AgentOps” is Data Governance for AI’s Unstructured Exhaust

The introduction of “AgentOps” represents a necessary market evolution. Every action an AI agent takes generates a torrent of log entries, plans, and tool execution metadata. This “AI exhaust” is a new, high-velocity data source that is currently ungoverned and underutilized. With AgentOps integrated into watsonx Orchestrate, IBM is creating a governance and observability layer specifically for this new class of data. In effect, IBM is building a flight data recorder for every agent, which can turn unpredictable behavior into a structured, auditable asset that can be monitored and controlled. This transforms the conversation from merely building agents to managing them — a far more complex and valuable enterprise problem.

Project Infograph: Infrastructure Metadata as a First-Class Asset

Perhaps the most foundational data play announced was Project Infograph, born from the HashiCorp acquisition. This initiative transforms a hybrid cloud estate’s sprawling, fragmented metadata into a unified, queryable knowledge graph. This isn’t about business data. Rather, it’s about the operational data describing the intricate relationships between servers, applications, networks, and security policies. By creating this foundational asset, IBM enables a new level of intelligent automation. An AI agent can now query this graph to understand a vulnerability’s “blast radius” or identify the root cause of a performance issue, for example. This move aims to turn infrastructure management from a discipline of tribal knowledge into one of data science, providing the trusted context agents need to operate safely.

What to Watch:

• Adoption of Agentic Data Tools: Will data engineering teams trust LLMs to build mission-critical pipelines? Watch whether these tools transcend ad-hoc developer convenience to become a trusted enterprise pattern.
• The Infograph Integration Roadmap: Project Infograph’s success hinges on its ability to become a central control plane. Monitor how quickly and deeply IBM integrates it with watsonx.governance and Guardium to create a single pane of glass for both infrastructure and data governance.
• The Competitive Response: Databricks and Snowflake are also investing heavily in AI-powered data management. Watch how they respond to IBM’s push for a formalized AgentOps governance layer and its free, local developer experience for lakehouse platforms.

You can read the full press release at IBM’s website.

Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.

Other insights from Futurum:

Neo4j GraphSummit: The OG (Original Graph) Database Player Makes a Bold Play for the Future of AI

AWS Summit New York City: AWS Forges the Enterprise-grade Pipeline for Agentic AI

Cloudera EVOLVE25: While the Market Chased Cloud-Native Deployments, Cloudera Built the Hybrid Endgame