Arvind Krishna: IBM's Reinvention, AI Bets and Quantum | Podcast | In Good Company

Revenue mix shifted to software majority

IBM now generates roughly 50% of revenue from hybrid cloud and AI software, approximately 33% from consulting, and only 20% from hardware, reversing the company's historic identity as a hardware giant.

Red Hat acquisition enabled partnership strategy

Krishna concluded IBM could not win the public cloud wars against hyperscalers spending $5-10 billion annually, so acquired Red Hat to become a hybrid cloud partner for all providers rather than a distant fifth-place competitor.

Spun off declining IT services business

The company divested its infrastructure services unit (Kyndryl) because the declining, low-margin business was fundamentally deflationary and dragged down growth targets for the remaining innovation-focused company.

Confluent acquisition targets real-time data

The purchase adds critical data streaming infrastructure that helps enterprises expose and move data in real-time, which is essential infrastructure for unlocking AI value across organizations.

Infrastructure buildout is ahead of demand

With over 100 gigawatts of AI data center capacity committed—requiring $6-8 trillion in semiconductor investment—the industry would need $1-2 trillion in annual revenue to justify the spend, a figure Krishna believes is unsustainable.

Large models will commoditize rapidly

Krishna predicts large language models will become commodities with low switching costs, meaning only two or three providers will survive the capital intensity while the rest struggle to generate returns.

Enterprise AI remains wide open

Unlike the consumer AI space where distribution advantages predetermine winners, the enterprise AI market has no established leaders yet, creating opportunity for IBM's B2B focus.

Adoption is only in the second inning

Comparing AI to previous technology cycles, Krishna places current adoption in the second inning of nine, predicting three to four years of cautious enterprise uptake before full embrace becomes standard.

Instilled risk-taking culture

Krishna identifies his greatest achievement as transforming IBM's risk-averse, inward-focused decline mentality into one where leaders present 50% confidence proposals rather than 90% certainty, unlocking innovation.

Integration philosophy balances freedom with scale

Acquired engineering teams retain autonomy to build capabilities while go-to-market functions integrate immediately to leverage IBM's global footprint, except for Red Hat which maintains separate sales and engineering due to its open-source scale.

Admits slow mid-market expansion

Krishna cites his biggest mistake as failing to expand beyond large enterprise clients quickly enough, recognizing that mid-market customers purchase discrete capabilities rather than comprehensive transformation partnerships.