IBM stock price plunges; AI impact on the software industry continues to send shockwaves

On February 23rd, local time, IBM’s stock price plummeted, becoming the latest visible victim of the rapid development of artificial intelligence technology. Previously, Anthropic announced that its Claude Code tool could be used to modernize legacy systems running COBOL. For a long time, IBM has been selling large mainframe systems optimized for high-volume transaction processing, which commonly use COBOL.

Notably, since February, a market panic triggered by advances in AI technology has been spreading globally, with the software sector bearing the brunt. Industry insiders reveal that the impact of AI on the business models of software companies is intensifying. Coupled with rising financing costs and stricter lender scrutiny, many software firms are delaying debt issuance plans. However, some analysts suggest this could be a “misfire,” and existing software giants might actually benefit and evolve from this situation.

Stock Prices Decline Across the Board

On February 23rd, the three major U.S. stock indices all closed down more than 1%, with most large tech stocks falling. IBM’s stock closed at $223.35, down 13.15%, marking its largest decline since 2000. With this drop, IBM’s stock has fallen a total of 27% in February alone, possibly the largest single-month decline since at least 1968.

IBM’s sharp stock decline has made it the latest clear victim of the rapid growth of AI technology. Previously, AI company Anthropic announced that its Claude Code product introduced new programming features that automate much of the research and analysis work for the COBOL programming language, raising concerns about the future of IBM’s mainframe business. COBOL-related services are a core part of IBM’s business. For years, IBM has sold large mainframe systems optimized for high-volume transaction processing, which generally use COBOL.

In a blog post, Anthropic explained that about 95% of ATM transactions in the U.S. use COBOL, making it a key area for AI-driven disruption at lower costs.

“Billions of lines of COBOL code run in production environments daily, supporting critical systems in finance, aviation, and government sectors. Yet, the number of people who truly understand this language is decreasing year by year. AI is very good at simplifying those tasks that once made COBOL modernization costs prohibitively high,” Anthropic wrote.

Anthropic stated that in the past, modernizing COBOL systems required extensive consulting work over several years to map out workflows. Tools like Claude Code can automate the most labor-intensive phases of exploration and analysis during COBOL modernization.

At the end of last month, IBM released its Q4 2025 financial report, showing revenue of $19.69 billion for the quarter, up 12% year-over-year, with software revenue at $9.03 billion, up 14%. It’s clear that the core driver of IBM’s quarterly performance was its software division, which accounted for 45.8% of total revenue, with a high gross margin (85%), significantly boosting overall profitability.

However, IBM is just the latest stock to fall recently due to AI-related concerns. As AI panic rekindles, Asian software stocks also declined on Tuesday, following the overnight drop in the U.S. markets.

Previous fears about AI have shaken the market and continued to disturb investors. In early February, anxiety over large-scale AI capital expenditures prompted declines in tech and software stocks, further fueling a turbulent “sell first, ask questions later” trading environment.

Industry Facing Challenges

Industry insiders reveal that the impact of AI on the business models of software companies is intensifying. Coupled with rising financing costs and stricter lender scrutiny, many software firms are postponing debt issuance plans.

Influenced by expectations that AI could disrupt the industry, software companies in the U.S. and globally have paused or delayed financing activities. This concern is especially evident in the loan market, where high-risk corporate bonds are beginning to factor in more default expectations. The market panic caused by AI has also affected private equity firm Blue Owl, which recently announced the sale of $1.4 billion in assets to return funds to investors, leading to a drop in its stock price.

Matthew Mish, head of credit strategy at UBS, said, “We expect that from 2026 to early 2027, the disruptive risks brought by AI will become more apparent in the market, especially for lower-rated credit sectors with high refinancing needs — and the impact on the U.S. market will be greater than in Europe.”

“This disruption will gradually become evident over the next two years,” Mish added. “We ultimately believe the market will digest most (but not all) of the predicted default risks.”

A banker noted that even companies with higher credit ratings, considered less affected by AI, are waiting for market conditions to improve before re-entering the financing scene.

According to informed sources, the market is closely watching investor feedback on the financing plans of well-known software maker Qualtrics — which will raise $5.3 billion in acquisition financing next month to buy its competitor Press Ganey Forsta.

Currently, there are no leveraged loan transactions in progress for software companies. Since late January, when AI-related fears intensified, the trading prices of existing debt in the software industry have continued to decline, with companies and banks waiting for market conditions to recover.

A Moody’s January report indicated that for upcoming maturing lower-rated companies, “refinancing pressures and default risks could increase significantly in 2026.”

Jeremy Burton, portfolio manager at PineBridge Investments’ leveraged finance team, said, “In the next year, I don’t think software and business services will become hot sectors for financing. The pace of technological change is so rapid that investors need strong confidence to step in.”

Overreaction to Panic?

Wall Street’s fears that “AI will disrupt the software industry” have led to heavy selling in the sector. However, some analysts argue that this widespread panic is an overreaction. AI may not replace all existing software companies; instead, it could strengthen many established firms by enhancing their services and better serving customers.

Keith Weiss of Morgan Stanley believes that companies are unlikely to allocate their IT resources to develop and maintain custom open-source AI models unless it provides a long-term competitive advantage. “The key point is that when organizations consider whether to automate business processes with their own custom code or collaborate with third-party providers, initial development costs are only part of the equation,” Weiss wrote in a research report.

“He pointed out that open-source software has no licensing costs (essentially free to develop), and has been used by customers to build their own applications for over 20 years. Despite this, the third-party software market has continued to thrive during that time,” Weiss added.

Regarding whether model builders will steal market share from existing software companies, Weiss stated that large language models (LLMs) are unlikely to “become efficient data warehouses or message servers like solutions built specifically for such functions.” Instead, he explained, various AI functionalities are likely to be integrated into existing software, expanding their capabilities and customer value.

Additionally, AI companies face another challenge: data governance. Some firms tend to be protective of their proprietary data and may be less willing to share their data repositories with AI models or AI companies compared to their long-term partners.

However, this does not mean all software companies will escape unscathed. Some may struggle to keep pace with the evolving AI landscape and inevitably fall behind. Those that can adapt to AI advancements are likely to continue moving forward.

Senior industry analyst and Ding Technology founder Ding Shaojiang believes that AI will not simply replace traditional software but will reshape the industry, phasing out inefficient, single-function products and forcing giants to upgrade their technology. Software companies need to embrace AI to reconstruct their products and services, shift toward AI-driven solutions, focus on differentiated technologies and practical applications, and use technological evolution to strengthen their competitive edge.