A single blog post from Anthropic wiped more than $30 billion off IBM’s market value in a single session. Why, you may ask? It’s all linked to the COBOL programming language, which has been central to IBM’s operations for more than 60 years. So what exactly is COBOL and what did Anthropic say to cause investors to panic? Let’s find out below.
What is COBOL?
COBOL, which stands for Common Business-Oriented Language, is a programming language created in 1959 that is still widely used for processing business and financial data. The language was created by a consortium called the Conference on Data Systems Languages (CODASYL), with the first version released in 1960.
Read also | What is Claude Code Security? The AI tool that wiped out billions of cyber stocks
The language was intended to serve as a temporary solution, but the US Department of Defense quickly realized its usefulness and mandated computer manufacturers to offer it.
Despite more than 65 years of existence, COBOL has remained popular, and according to an IBM blog post, it supports 80% of personal credit card transactions and 95% of all ATM transactions, generating over $3 billion in business each day.
According to the company, an estimated 250 billion lines of COBOL code are used in production across sectors such as financial services, government, logistics, manufacturing and retail.
Read also | Who is Dario Amodei? Did you know that the CEO of Anthropic was a former employee of OpenAI?
However, IBM admits in the same blog post that the number of COBOL professionals is dwindling and many COBOL specialists are retiring, while current software developers have been trained in modern languages.
In addition, legacy COBOL systems may not always meet modern security and regulatory requirements. There are also issues with integrating legacy COBOL applications with modern cloud platforms and software packages.
What did Anthropic say?
In a blog post on Monday, Anthropic talked about COBOL’s continued relevance across critical systems in finance, aerospace and government. The company also discussed the dwindling number of people who understand COBOL and the language’s tiny replacement ratios.
“Modernizing a COBOL system once required armies of consultants to spend years mapping out workflows. This led to extensive timelines and high costs that few were willing to take on,” Anthropic wrote. blog post.
However, the company says that modern AI tools like Claude Code can “automate the exploration and analysis phases” that consume most of COBOL’s modernization time.
According to Anthropic, AI systems can map dependencies across large codebases, document workflows, identify risks and offer opportunities for refactoring, helping teams better understand legacy systems before attempting to migrate.
“Thanks to artificial intelligence, teams can modernize their COBOL codebase in quarters instead of years,” Anthropic said.
“AI excels at streamlining tasks that made COBOL modernization cost-prohibitive. It allows your team to focus on strategy, risk assessment and business logic while AI automates code analysis and implementation,” the company added.
Read also | The New Delhi Declaration on AI Impact reached 91 signatories as three more nations joined
Shortly after the blog post, IBM shares had their worst market day in more than 25 years, with the company down 13.2%. IBM shares fell 27% in February and are on track for their biggest monthly decline since at least 1968, according to Bloomberg data.
The write-off was prompted by investor concerns that if artificial intelligence were able to automate work that previously required years of consulting, it could adversely affect IBM’s revenue.
IBM, however, its position in blog post where the company’s Senior Vice President and Chief Commercial Officer, Rob Thomas, wrote: “The value IBM mainframes bring has nothing to do with COBOL. It has to do with what the platform is: a purpose-built architecture from the silicon through the operating system for unmatched transactional resiliency, security, performance and efficiency at a scale that no other distributed environment has been able to deliver.”
