IBM

Overview

IBM is a long-standing enterprise technology company with a significant presence in AI, cloud, open source, and enterprise infrastructure. In the newsletter context, IBM appears primarily through two AI-relevant threads: the Granite family of models and its role in agent interoperability work, including the donation of its ACP protocol that was later merged into A2A. IBM also surfaces through thought leadership, including commentary from IBM researcher David Cox.

For AI Product Managers, IBM matters because it sits at the intersection of enterprise adoption, open model strategy, and ecosystem standards. Its Granite models represent one path for teams evaluating smaller, practical foundation models, while its protocol and open ecosystem activity signal how large vendors are shaping the future of multi-agent systems, governance, and enterprise AI integration.

Key Developments

• 2026-01-03 — DeepLearningAI featured IBM’s David Cox in coverage related to open source wins, placing IBM in broader discussion around AI openness, research, and industry direction.
• 2026-02-12 — IBM’s ACP agent communication protocol was noted as having merged into A2A shortly after donation, highlighting IBM’s influence on emerging agent interoperability standards.
• 2026-05-05 — Simon Willison published prompting experiments with different quantized variants of IBM Granite 4.1 3B, asking the model to generate an SVG of a pelican riding a bicycle. His results suggested no clear relationship between model size and output quality, with most outputs performing poorly.

Relevance to AI PMs

• Evaluate practical open-model options. IBM’s Granite family is relevant for PMs comparing smaller or enterprise-friendly models for internal copilots, structured generation, or cost-sensitive deployments. The Simon Willison experiment is a useful reminder to benchmark actual task performance instead of relying on parameter count or vendor positioning.
• Track agent interoperability standards. IBM’s ACP-to-A2A connection matters for PMs designing agentic products that need to communicate across tools, platforms, or vendors. Protocol convergence can reduce integration risk and shape architectural decisions early.
• Watch enterprise AI signal from major incumbents. IBM often reflects what large regulated organizations care about: governance, open ecosystems, deployment practicality, and standards. PMs selling into enterprises can use IBM-related developments as signals for buyer expectations.

Related

• granite-41-3b — IBM’s Granite 4.1 3B model, referenced in Simon Willison’s SVG prompting experiments.
• simon-willison — Tested Granite 4.1 3B variants and shared qualitative results relevant to model evaluation.
• a2a-agent2agent-protocol — Connected via the merger of IBM’s donated ACP protocol into A2A.
• david-cox — IBM-affiliated figure mentioned in DeepLearningAI coverage related to open source and AI industry discussion.
• deeplearningai — Featured IBM’s David Cox, helping surface IBM’s perspective to the AI practitioner audience.
• google-cloud — Related as a major AI/cloud ecosystem player that intersects with enterprise platform decisions IBM also competes and collaborates within.
• linux-foundation — Relevant through standards, open governance, and ecosystem coordination themes adjacent to IBM’s interoperability work.