Enterprise

IBM CEO: Only 2% of Software Portfolio Faces AI Disruption Risk

Arvind Krishna pushes back on investor concerns after disappointing quarterly results and a 30% share price decline in 2026.

Omega Editorial· July 23, 2026· 3 min read

IBM CEO Arvind Krishna is working to calm investor fears about artificial intelligence disrupting his company's software business, claiming that only 2% of IBM's software portfolio is vulnerable to replacement by AI-built applications.

The reassurance comes at a critical moment for IBM. The company's shares have fallen roughly 30% in 2026, and its second-quarter results disappointed Wall Street as customers shifted spending priorities away from mainframe computers toward servers and storage infrastructure.

Software revenue guidance cut sharply

The financial impact has been significant. IBM's mainframe infrastructure revenue dropped 42% in the second quarter, while transaction processing software declined 9%. This marked a sharp reversal from the first quarter, when mainframe revenue had grown 48%.

CFO Jim Kavanaugh revised IBM's software revenue growth forecast for 2026 down to 6-8%, abandoning the double-digit growth projection he had expressed confidence in back in January. Software accounts for 45% of IBM's total revenue and delivers the company's strongest profit margins.

For every dollar IBM generates from mainframe infrastructure, it collects three dollars in related software revenue — making the hardware slowdown particularly consequential.

Why it matters

IBM's struggle illustrates a broader challenge facing enterprise software companies as AI capabilities advance. Wall Street has grown increasingly skeptical of traditional software business models, with the iShares Expanded Tech-Software Sector ETF down 17% in 2026. Krishna's 2% claim attempts to draw a clear line between application software (vulnerable) and infrastructure software (protected), but investors remain cautious. The company maintained its $1 billion free cash flow guidance despite the revenue headwinds, suggesting management believes current challenges are temporary rather than structural.

Mainframe customers delay software spending

Krishna pointed to specific examples of the software at risk. Starbucks, which spends approximately $2 million annually on IBM software, is replacing IBM's Tririga lease management software — a product IBM acquired in 2011 and plans to discontinue support for in 2027. Krishna characterized this as representative of the vulnerable 2% segment.

The CEO argued that the remaining 98% of IBM's software serves a different purpose: helping enterprises prepare for AI by managing data in real time, reducing infrastructure complexity, and operating across hybrid cloud environments. He described this infrastructure software as positioned to benefit from AI adoption rather than face disruption.

Krishna told CNBC that mainframe hardware capacity continues to grow, but software revenue typically lags behind hardware deployment. He predicted the software business would recover within a year as customers scale up their usage.

About 75% of deals that slipped out of the second quarter should close before year-end, according to Krishna. Analysts at Jefferies, who maintain a buy rating on IBM, wrote that they would withhold full confidence in the company's maintained guidance until more of that delayed business appears in reported results.

These details were first reported by CNBC.

#ibm#enterprise software#mainframe computing#ai disruption#arvind krishna#software revenue

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Enterprise

Enterprise· 4 min read

AI Now Shapes 60% of Search Journeys, Reshaping Brand Visibility

New research shows most consumers encounter AI-generated answers during product research, forcing marketers to rethink organic search strategy.

Via AI Watch · Jul 23, 2026
Enterprise· 3 min read

Why B2B Buyers Stall Deals in 2026: Fear, Not Disinterest

A sales veteran who interviewed 300+ founders reveals how AI changed buying behavior and what actually unsticks stalled enterprise deals.

Via AI Watch · Jul 23, 2026
Enterprise· 3 min read

Brands Generate AI Content Faster Than They Can Govern It

Three-quarters of enterprise content now involves AI, but visibility and control systems haven't kept pace with production speed.

Via AI Watch · Jul 23, 2026