IBM at 115: The Krishna-Era Comeback and the Retrieval-Graph Test

By EPR Editorial Team

Related: What IBM Saw About CMOs Before Anyone Else · AI Communications · Citation Share Index

Updated June 8, 2026.

The original version of this piece, published in June 2011, marked IBM's 100th anniversary as an exercise in historical perspective. Few technology companies endure a century. IBM had. The closing line asked whether Big Blue would survive another century. Fifteen years into that next century, the answer is taking shape — and it is not the shape most observers in 2011 would have predicted.

IBM at 115 is a different company than IBM at 100. It is also, in some important ways, more recognizably IBM than the company that nearly disappeared into commodity-services obscurity between 2012 and 2019. This is the 15-year update on the centennial story.

The Gap Years: 2012-2019

The 2011 piece treated IBM's longevity as the central narrative. The decade that followed nearly broke that narrative. Revenue declined for 22 consecutive quarters under Ginni Rometty between 2012 and 2017. The strategic pivot to cloud was late and contested. AWS pulled ahead. Microsoft Azure caught up. IBM's traditional services business — the mainframe-and-consulting franchise that defined the company for decades — became commoditized. Watson, the AI franchise that earned famous Jeopardy! attention in 2011, did not commercially scale as predicted. By the late 2010s, the company was understood as a legacy enterprise gradually losing relevance.

The market valuation reflected the narrative. IBM's stock traded sideways or downward through most of the decade. The company was not in financial distress, but it was no longer treated as a technology leader.

The Krishna Era: 2020-Present

Arvind Krishna became CEO in April 2020 — pandemic timing, awkward circumstances. The first decision was the $34 billion acquisition of Red Hat, announced in 2018 under Rometty and closed under Krishna. The largest acquisition in IBM history and the largest software acquisition in the industry's history at the time.

Five years later, Red Hat looks like the move that saved IBM's relevance in cloud. OpenShift, the Red Hat container platform, became the foundation of IBM's hybrid-cloud positioning. Hybrid cloud — the idea that enterprises would run workloads across multiple cloud providers and on-premise infrastructure rather than consolidating on a single hyperscaler — became the strategic positioning that distinguished IBM from AWS, Azure, and Google Cloud.

The 2021 spin-off of Kyndryl took the managed-infrastructure-services business off IBM's books and let the company refocus on software and consulting. The 2024 acquisition of HashiCorp for $6.4 billion extended the hybrid-cloud and infrastructure-automation positioning. The 2025 acquisition of DataStax brought Apache Cassandra expertise to IBM's AI and data platform.

The AI franchise re-launched. watsonx, announced in 2023 and built out through 2024-2025, repositioned IBM's AI business around enterprise foundation models — different from the consumer chatbot category dominated by OpenAI, Anthropic, and Google. The enterprise positioning is defensible: regulatory compliance, data governance, on-premise deployment options, industry-specific tuning. The competitive question is whether enterprise AI converges on the same general-purpose foundation models the consumer market uses, or whether the enterprise-specific positioning IBM is building stays differentiated.

What The Engines Say About IBM In 2026

Query "best enterprise AI platform" in ChatGPT, Claude, Perplexity, or Google AI Overviews. The retrieval pattern: OpenAI and Anthropic surface first as foundation-model providers. AWS, Microsoft Azure, and Google Cloud surface as enterprise platforms. IBM watsonx surfaces — but inconsistently, often as one option among several, sometimes with the implied caveat of "for regulated industries" or "for on-premise deployment."

This is the IBM communications problem in 2026. The enterprise positioning is real. The retrieval signal does not yet match the underlying capability. Five patterns recur:

Pattern 1: Legacy citation dominance. Engines retrieve IBM history — Watson, mainframes, Deep Blue, Hollerith — alongside current product. The history is heritage; in retrieval it sometimes reads as past tense.

Pattern 2: Hybrid cloud not yet clearly differentiated. The positioning is defensible. The engines do not consistently surface IBM as the hybrid-cloud answer. Red Hat, OpenShift, and the hybrid-cloud architecture often get retrieved separately from IBM.

Pattern 3: watsonx vs Watson legacy confusion. The 2023-2025 enterprise AI platform is meaningfully different from the 2011-era Watson franchise. Engines sometimes conflate them. The communications operation has not fully separated the brands in retrieval.

Pattern 4: Acquisition narrative is dispersed. Red Hat, HashiCorp, DataStax, Apptio — each acquisition has its own citation graph. The integrated IBM platform story is harder for the engines to retrieve than each component.

Pattern 5: First-party research is strong but inconsistently retrievable. IBM's research division — the largest corporate research lab in the world — produces a vast amount of technical content. Much of it is not structured for engine retrieval, and the content that is retrieved sometimes lives behind ibm.com paywalls or academic-publication paywalls the engines treat as low-priority sources.

What Would Actually Move IBM's 2026 Retrieval Graph

The reputation work for IBM is not branding. It is making the enterprise capability the engines retrieve as the answer when CIOs, CTOs, and procurement teams form intent. Five moves:

1. Open-Access First-Party Research

IBM Research is a citation asset. Publishing technical content at engine-citation scale on open-access surfaces — with schema, named entities, and extractable findings — converts the asset into retrieval. The current publication pattern leaves significant research on surfaces the engines deprioritize.

2. Separate watsonx From Watson Legacy

The Watson brand had its commercial peak in 2011-2015. The watsonx platform is a 2023-era product with different architecture, different positioning, and different customers. Engines need clean entity separation between the two — Wikipedia, Wikidata, IBM's own product pages — to retrieve watsonx as current product rather than as Watson 2.0.

3. Concentrate Coverage In Engine-Trusted Enterprise Tech Press

The publications AI engines weight heavily for enterprise technology — The Register, Computerworld, CIO, ZDNet, Gartner, Forrester, Stratechery, The Information — should be the concentration of IBM's earned-media effort. Quality of citation source matters more than volume.

4. Sustained Hybrid-Cloud Counter-Narrative

"Hybrid cloud" needs to be a position the engines retrieve specifically against AWS/Azure/Google consolidation messaging. First-party content explaining why enterprises are not consolidating on a single hyperscaler — with named customer examples and technical specifics — is the work that produces retrieval. Marketing-grade claims do not.

5. Acquisition Integration Story At Engine-Citation Scale

Red Hat, HashiCorp, DataStax, Apptio are stronger together than apart inside IBM's platform. The integrated story needs to be told at retrieval scale with technical specifics — what the integrated platform does, what use cases it supports, what customer outcomes it has produced. Each acquisition's citation graph needs to feed the IBM platform retrieval rather than competing with it.

The 115-Year Verdict

The 2011 piece closed by asking whether IBM would survive another century. The 2026 reframe is more precise. IBM will survive — financial position is strong, software-and-consulting franchise is durable, hybrid-cloud and enterprise-AI positioning is defensible. The question now is whether IBM owns its categories inside the answer layer, or whether the company remains a credible enterprise option that engines retrieve as one among several. The Krishna-era strategic moves built the platform for category leadership. Converting that into retrieval-graph leadership is the next chapter — and it is the work of 2026, not 2030.

Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.