Arvind Krishna

CEO of IBM, focused on repositioning the company around hybrid cloud, AI, and quantum computing.

Claims

As speaker

Mentioned in

Key positions and views

The current AI infrastructure build-out is financially unsustainable at its current pace, as the $1-2 trillion in additional annual revenue required to justify the $8 trillion in capex does not yet exist.

Large, general-purpose AI models will inevitably become commoditized with low switching costs, leading IBM to strategically focus on smaller, cost-effective, enterprise-specific models like its Granite family.

Quantum computing represents a near-term national security threat due to its encryption-breaking capabilities and a long-term, multi-hundred-billion-dollar commercial opportunity in which IBM holds a significant lead.

A hybrid cloud architecture is the dominant and most logical long-term strategy for the majority of enterprises, validating IBM's acquisition of Red Hat and its decision to not compete directly with hyperscale public cloud providers.

Artificial General Intelligence (AGI) is not achievable with current LLM technology alone and will require new breakthroughs, likely involving the fusion of LLMs with technologies that represent hard knowledge, such as neuro-symbolic AI.

Podcast consensus on Krishna

Points of consensus

▶Both sources convey Krishna's view that the current AI infrastructure build-out is massive, with commitments for over 100 gigawatts of data centers representing a potential $6 to $8 trillion in capital expenditures.May 2026

▶Krishna consistently frames IBM's past Watson strategy as a mistake across both podcast appearances, citing its monolithic design and misguided focus on the complex healthcare vertical instead of broader enterprise use cases.May 2026

▶He repeatedly emphasizes the transformative potential of quantum computing, framing it as both a major national security threat due to its ability to break current encryption and a future multi-hundred-billion-dollar business opportunity.May 2026

▶Krishna's strategic vision for IBM, centered on hybrid cloud (enabled by the Red Hat acquisition) and a refusal to compete directly with public cloud hyperscalers, is a consistent theme in both discussions.May 2026

Points of debate

▶Krishna presents a bullish long-term vision for AI, predicting 1 billion new apps and a 1000x drop in compute costs, yet expresses bearish short-term concerns that the current infrastructure build-out is ahead of the market's ability to generate sufficient revenue for a return on investment.May 2026

▶He highlights massive AI-driven productivity gains, such as a 45% increase for an internal IBM developer team, while simultaneously predicting significant job displacement of up to 10% in the US workforce over the next two years.

▶Krishna is highly skeptical about current technologies leading to AGI, assigning it a near-zero probability, yet he compares the AI shift to the internet in scale, suggesting a fundamental societal transformation that seems to border on AGI-like impact.Feb 2026

▶He argues that the largest AI models will become a low-margin commodity, yet acknowledges that only two or three companies will be able to afford to build and operate them, suggesting a highly concentrated, oligopolistic market rather than a truly commoditized one.May 2026

Key themes

▶Pragmatic AI EconomicsMay 2026

Krishna expresses strong optimism about the long-term technological and application potential of AI, predicting a 1000x reduction in compute cost and a billion new applications. However, he tempers this with a sober analysis of current market dynamics, arguing the massive $8 trillion infrastructure build-out is ahead of the revenue required to support it, and that some capital will be lost.

Investors should note Krishna's distinction between technological potential and short-to-medium term economic reality, suggesting a potential correction or consolidation in the AI infrastructure and hardware sectors.

▶IBM's Strategic RepositioningMay 2026

Krishna details a deliberate pivot for IBM away from declining legacy businesses and direct competition with cloud hyperscalers. Key moves include spinning off the IT services division, acquiring Red Hat to dominate the hybrid cloud middleware layer, and focusing on high-margin software and consulting.

This theme reveals a 'picks and shovels' and 'enabling layer' strategy for IBM, positioning it as a partner to enterprises navigating a multi-cloud world rather than a primary infrastructure provider.

▶The Quantum Computing ImperativeMay 2026

Krishna frames quantum computing as a dual-sided imperative: a near-term national security threat that can break modern encryption and a long-term, multi-hundred-billion-dollar commercial opportunity. He positions IBM as being several years ahead of competitors in this field, with tangible assets like post-quantum cryptography in its mainframes.

Krishna's focus on quantum suggests that IBM is playing a long game, investing heavily in a technology that could redefine computing and security, potentially creating a durable competitive moat.

▶Commoditization of Frontier AI

A core tenet of Krishna's AI outlook is that the largest, most powerful AI models will inevitably become commodities with low switching costs and unsustainable margins. This belief directly informs IBM's strategy to not compete in building frontier models, but to instead focus on smaller, cost-effective, and specialized models like its Granite family for enterprise use.

This contrarian view suggests that the long-term value in the AI stack may not be in the base models themselves, but in the data, applications, and specialized models built on top of them.

Source episodes

Sentiment over time

Not enough data for timeline

Changes over time

2017

Krishna states that IBM strategically decided against competing in the public cloud market, deeming the $5-10 billion annual investment required to catch up an unattractive proposition.

2018

IBM announced its acquisition of Red Hat, a move Krishna frames as pivotal to establishing IBM's dominance in the hybrid cloud ecosystem and partnering with, rather than competing against, major cloud providers.

Post-2018

Under Krishna's leadership, IBM spun off its large IT services business (Kyndryl), which he described as declining 5% annually, to sharpen the company's focus on high-growth areas like AI and hybrid cloud.

Last 6 Years

Krishna notes a significant turnaround in IBM's mainframe business, which has seen consistent annual growth for six years following a decade of decline, now incorporating features like AI inferencing and post-quantum cryptography.

Present

Krishna articulates a clear vision for IBM's role in the AI era, focusing on smaller, efficient models (e.g., Granite), avoiding the frontier model race, and asserting a multi-year lead in the development of quantum computing.

Suggested prompts

How does Krishna's view on the commoditization of large AI models justify IBM's strategy of not building frontier models? &nearr;What are the potential contradictions between Krishna's prediction of a 1000x decrease in AI compute cost and his concern that the current $8 trillion infrastructure build-out may not see a return? &nearr;How does Krishna's assessment of quantum computing as both a national security threat and a multi-hundred-billion-dollar business opportunity shape IBM's R&D and go-to-market strategy? &nearr;Given Krishna's analysis of the Watson strategy's failures, what specific measures is IBM taking to ensure its current AI products avoid similar pitfalls of being too monolithic or vertically-focused? &nearr;

Key concepts

AI Infrastructure & Investment 2 ep Quantum Computing 2 ep IBM Business Strategy 2 ep AI Models & Commoditization 2 ep Hybrid Cloud 2 ep AI Compute Costs 1 ep Mainframe Business 1 ep Artificial General Intelligence (AGI) 1 ep National Security 1 ep Job Displacement 1 ep

Notable quotes

“I also believe on a second one that many of the largest models are going to be a commodity... If there is low switching cost that means you can have a margin but it not going to be a margin with a massive moat around it.”

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

“Arvind Krishna predicts that the cost of AI compute will become 1,000 times cheaper over the next five years.”

Arvind Krishna · IBM CEO Arvind Krishna says there is no AI bubble after all ｜ Decoder

“If you say that that's got a five to seven-year payback, you are going to need an extra $1 to $2 trillion a year of revenue... So that much incremental revenue, I don't believe, is there.”

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

“We know that quantum computers can do what is called Shor's algorithm. What does it do? It helps you decrypt most of today's encryption.”

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

Report last updated: May 10, 2026

Get started free

Back to Entities Intelligence Report