Wedbush analysts say the recent selloff of International Business Machines Corp (NYSE:IBM) shares is overdone and presents a buying opportunity, arguing that fears around AI-driven disruption to the company’s legacy business are being misinterpreted.
IBM stock fell 13% on Monday, its largest drop in more than 25 years, compared with about a 1% decline for the S&P 500. The move followed heightened competition concerns after Anthropic said its Claude Code tool could help modernize legacy systems that run on COBOL, a language central to IBM’s mainframe ecosystem.
Shares recovered partially on Tuesday, trading at about $230.
Wedbush maintained an ‘Outperform’ rating and a $340 12-month price target on the stock, saying the market reaction reflects a broader “sell first, ask questions later” mentality surrounding AI disruption.
The analysts acknowledged the perceived risk, noting that IBM’s mainframe systems are widely used for large-scale transaction processing where COBOL remains prevalent, as about 95% of US ATM transactions rely on the language. However, they argue the threat is being overstated.
“While this is a risk to its older business unit, IBM’s core mainframe computer business offers a platform that provides the same quality of performance and security for various programming languages, not just COBOL, making it an adaptable platform that can capitalize on AI’s value across various businesses,” Wedbush wrote.
The firm emphasized IBM’s entrenched position in mission-critical environments. Even if AI tools speed up code modernization, enterprises will still require migration planning, compliance validation, and systems integration, areas where IBM already has long-standing customer relationships.
AI is “more likely to drive a modernization cycle than bypass the installed base as AI lowers the friction of legacy transformation projects,” Wedbush wrote.
The analysts also pointed to IBM’s existing partnership with Anthropic, under which Claude is being integrated into IBM’s software portfolio. Wedbush believes this collaboration reduces the likelihood of meaningful displacement risk, adding that “any replacement risk within IBM will not be a reality.”
IBM executives have also pushed back on the narrative. In a recent blog post, Senior Vice President of Software and Chief Commercial Officer Rob Thomas argued that the modernization debate is overly focused on programming language rather than platform capabilities.
According to Wedbush, IBM’s mainframe architecture is designed for scale and reliability, enabling 25 billion encrypted transactions per day on a single system, 450 billion AI inferences daily at 1-millisecond response times, and up to eight nines of availability with quantum-safe encryption and full utilization without breaching service-level agreements.
“Decades of integration work around hardware and software cannot be replaced by moving code as it will decouple performance across these components,” the analysts wrote.
Wedbush further noted that roughly 40% of COBOL already runs on Windows, Linux, and other distributed platforms, reinforcing the view that the platform, not the programming language, is where most enterprise value resides.