IBM Shares Plunge 13% in Worst Day Since 2000 as Anthropic’s Claude Code Targets COBOL Legacy Systems
International Business Machines Corp. (IBM) suffered its steepest single-day stock drop in over 25 years on Monday, plunging 13% to close at $223.39, after AI startup Anthropic announced a new tool designed to modernize COBOL, a legacy programming language central to IBM’s mainframe business.[1][2][3]
The dramatic sell-off erased approximately $31 billion from IBM’s market capitalization, reducing it from $240.8 billion on Friday to around $208.7 billion.[2][4] This marks IBM’s largest percentage decline since October 2000, during the dot-com bust, and positions the stock for its worst monthly performance since at least 1968, with shares down 27% in February.[2][3][4][5]
Anthropic’s Disruptive AI Update
Anthropic, the maker of the Claude AI chatbot, unveiled an update to its Claude Code tool specifically aimed at streamlining the modernization of COBOL systems. Developed in the late 1950s, COBOL (Common Business-Oriented Language) powers critical infrastructure including 95% of U.S. ATM transactions, Social Security payments, and operations for financial institutions and airlines.[2][3][5]
In a blog post, Anthropic highlighted how legacy code modernization has historically been prohibitively expensive, often costing more than rewriting systems from scratch due to the scarcity of COBOL experts—now taught by only a handful of universities.[1][4] “Legacy code modernization stalled for years because understanding legacy code cost more than rewriting it. AI flips that equation,” the company stated.[1]
The tool promises to automate exploration and analysis phases, identify risks that would take human analysts months to uncover, and reduce costs dramatically for businesses reliant on IBM mainframes, which dominate COBOL processing.[1][3][5]
Wider Software Sector Sell-Off
IBM’s tumble is part of a broader AI-driven rout in software stocks. Anthropic’s recent flurry of updates—including legal plugins earlier in February and a cybersecurity scanning tool last week—has fueled investor fears of disruption across industries.[1][2][3]
- Cybersecurity firms CrowdStrike and Zscaler dropped about 9% each on Monday following Anthropic’s vulnerability-scanning announcement.[2][3]
- Datadog fell 11.28%, and Zebra Technologies declined 9.43%.[3]
- Indian IT stocks like Infosys, HCL Tech, TCS, Coforge, and Persistent Systems slid up to 5%, with the Nifty IT index down over 3% daily and 20% monthly.[3]
A leading software-focused ETF has declined 27% year-to-date, on pace for its worst quarterly drop since the 2008 financial crisis.[5] Investors worry that AI tools enabling “vibe coding”—generating software via natural language—could erode demand for traditional vendors.[5]
Market Context and Analyst Pushback
The sell-off coincided with broader market pressures, including tariff uncertainties and online reports speculating on AI’s negative impacts.[1] Tech stocks had briefly rebounded after initial Anthropic-induced panic but resumed declining Monday.[1]
However, some analysts urge calm. LPL Financial’s Adam Turnquist attributed volatility to a “market narrative” shift rather than deteriorating fundamentals.[2] JPMorgan called fears of AI disrupting software “broken logic,” while Wedbush’s Dan Ives labeled the trade “the most disconnected… in my career,” arguing AI would ultimately boost incumbents.[2][6]
Investment firm Stocktwits maintained a ‘Buy’ rating on IBM with a $370 price target, dismissing the crash as overreaction.[6]
Implications for Legacy Tech
COBOL’s persistence underscores the challenges of legacy systems: vast, complex codebases maintained by a dwindling pool of specialists. IBM has long profited from consulting services to update these systems, but Anthropic’s tool threatens to commoditize that revenue stream.[1][4][5]
As AI advances from startups like Anthropic, OpenAI, and Alphabet accelerate, established players face mounting pressure to adapt. While short-term pain is evident, long-term outcomes remain debated amid hype and reality.[2][5]
IBM did not immediately respond to requests for comment on the Anthropic development.