AI Disruption Fear Slams IBM: Claude Code Challenges Profit Model

On February 23, IBM’s stock fell over 13% in a single session, marking its largest daily decline in nearly 25 years. The immediate trigger was Anthropic’s launch of Claude Code, a tool that claims to automate the most complex and time-consuming stages of COBOL system modernization—code analysis and refactoring—potentially reducing the workload of traditional consulting and system migration projects.

For IBM, legacy system services have long provided stable cash flow. If demand for COBOL engineers declines and “manual labor” is replaced by AI efficiency, the sustainability of IBM’s ecosystem and profit model has become a focal point for the market.

The broader impact rippled across U.S. tech sectors, with software, consulting, and cybersecurity stocks broadly weakening, and major indices closing lower. Traders are now asking whether this sell-off is simply pricing in potential risks or an overreaction driven by market sentiment. More broadly, does this signal a shift in the “buy AI” narrative?

Profit Reassessment: IBM’s Short-Selling Logic

From a fundamental perspective, IBM’s profitability is not entirely dependent on maintaining and upgrading COBOL systems, but its mainframe ecosystem and enterprise-level system services remain core profit drivers. Claude Code could compress the pricing power of these services and their labor-based revenue model, which is the main reason for the short-term market sell-off.

What’s more noteworthy is that the narrative of “AI disrupting traditional business models” is spreading. Over the past two years, the market has primarily viewed AI as a growth engine—boosting efficiency, creating demand, and increasing valuation premiums. This time, however, traders are considering another possibility: if AI improves efficiency but simultaneously erodes service pricing power, could profit margins at traditional tech and consulting firms, such as Accenture and Cognizant, face structural revaluation?

If the market gradually accepts this logic, pressure could spill over into broader tech sectors and U.S. equity indices. For now, however, it appears more like a localized risk re-pricing rather than a systemic deleveraging.

Sentiment Shock, Not a Trend Reversal

IBM saw a sharp surge in trading volume and a spike in volatility over a short period, driven by concentrated short-selling activity. This sell-off appears driven more by the event and market sentiment than a fundamental challenge to the “buy AI” narrative.

On the one hand, IBM’s business remains solid. Its Q4 2025 earnings report, released in January, showed revenue and profit exceeding expectations, with generative AI orders surpassing $12.5 billion and demand for next-generation mainframes still strong.

On the other hand, there is often a lag between technological breakthroughs like Claude Code and their commercial adoption. Core systems operated by large banks, government agencies, and insurance companies require high levels of security, stability, and compliance. Even if AI can rewrite code, enterprises may not immediately migrate entirely.

In other words, the stock’s reaction is pricing in potential adjustments to the profit model rather than an actual cash flow collapse. The market is equating “efficiency gains” with “profit compression” at a sentiment-driven level. This logic is not entirely baseless, but it overlooks important commercial realities in between.

Moreover, IBM is not a passive observer. The company has already invested in enterprise AI platforms and hybrid cloud solutions, such as Watsonx. If AI accelerates enterprise system modernization, IBM could benefit rather than be solely at risk. Short-term market sentiment has not fully priced in this bidirectional potential.

Watching AI Adoption and Earnings Signals

At its core, IBM’s sharp decline reflects a potential reshaping of future profit distribution. The market is shifting from indiscriminate AI integration to a more nuanced evaluation—distinguishing which companies will truly benefit and which may face pressure.

For traders, short-term volatility presents potential trading opportunities. If the sell-off is confirmed to be sentiment-driven and fundamentals have not materially deteriorated, rebound or swing trades may be considered.

However, positions must be carefully managed, as any downward revision of profit expectations could accelerate valuation compression. The post-market release of NVIDIA’s earnings on February 25 will also be a key variable in stabilizing sentiment.

In the medium to long term, whether IBM faces sustained short-selling pressure depends on two factors.

Externally, attention should be paid to the actual adoption rate of Claude Code among enterprise clients and whether consulting firms like Accenture choose cooperation or resistance.

Internally, it is important to monitor IBM’s upcoming earnings reports for signs of slowing order growth or profit pressure, and whether the company accelerates the deployment of its own AI coding assistant or takes targeted countermeasures.

If IBM maintains resilient performance over the next few quarters and Claude Code adoption remains limited, this sell-off may ultimately prove to be an overreaction, with stock price having room for further recovery.

Conversely, if IBM’s management acknowledges that AI tools are reshaping pricing structures, it would mark a shift from a sentiment-driven shock to a fundamental profit logic inflection point.