Anthropic’s Claude Code vs COBOL: IBM Investors React

COBOL, or Common Business-Oriented Language, is a core component of today’s finance sector. This 60-year-old code is the backbone of nearly 80% of global business transactions. Here’s how this coding slowly became the standard amongst the financial sector, why it’s been so hard to upgrade, and what AI may do to improve these factors.

What Is COBOL? Why Banks Still Depend on It

The history of Common Business-Oriented Language dates back to 1959, when the U.S. The Department of Defense and IBM (IBM +4.37%) met with 41 industry experts with the goal of creating a universal code standard.

The need for this standard had become evident, as up to that time it was customary for the growing number of computers to use machine-dependent coding. This standardized language enabled more efficient business data processing while providing an alternative to FORTRAN, developed by IBM in 1957 and regarded as the first high-level programming language.

Geared Towards Non-Technical Users

One of the core aspects of COBOL was its design around readability, making it instrumental in laying the foundation for future business programming. The use of English-based syntax was seen as a major upgrade that improved accessibility for non-technical users.

COBOL’s Early Design: Divisions, Files, and Batch Processing

The first COBOL supported hierarchical data description, control flow, sequential file operations, and program metadata. Its ecosystem was split into four sections. Each division represented a core business task.

These divisions included the Identification sector, which dealt with program metadata. The environment sector, which handled the system specifications. There was also the Data division, which handled file storage, and the executable logic feature labeled the Procedure division.​

The original program used punched-card input, as that was the technology at the time. It supported line-printer output, and data could be stored on magnetic tape. There were several minor upgrades made in COBOL-61 to COBOL-68.

These upgrades include improving the foundational syntax operations. IBM also added support for more advanced arithmetic and more detailed reporting. By COBOL-68, the language had become the first ANSI standard to support advanced table handling, library facilities, and more.

COBOL-85 to COBOL 2023: Modernization Without Replacement

After COBOL-74 and COBOL-85, later standards (notably 2002, 2014, and 2023) introduced incremental modernization—such as object-oriented features and improved data interchange support—without changing COBOL’s fundamental role in high-reliability transaction processing.

COBOL’s core design is decades old, but the language has continued to evolve through formal standards updates. The current COBOL standard is ISO/IEC 1989:2023, which replaces the 2014 edition with technical revisions intended to keep the language viable in modern environments.

How Much of Finance Still Runs on COBOL? (What We Can Prove)

To put COBOL’s importance into perspective, you need to understand that most global financial institutions run on legacy systems. Reports show that up to 95% of all ATM transactions in the US use this coding.

Kenly, every time you swipe your credit card, there’s an 80% chance the system relies on COBOL to process. In total, COBOL is responsible for 85% of global transactions and is used by 90% of the Fortune 500. Additionally, it’s leveraged in 70% of the industry’s core financial logic.

How COBOL Got Popular and Why Banks Still Use It Today

COBOL remains integral to banking due to its proven reliability in handling trillions of daily transaction volume securely. These systems are battle-tested. Plus, they can scale and offer unmatched uptime.

Also, COBOL has been optimized to operate smoothly within today’s legacy systems. This optimization includes improving syntax, verifying data, and ensuring strong security. Additionally, it was adapted to work with today’s mainframes.

Why COBOL Modernization Is Hard: Cost, Risk, and Compliance

As with any system that you have relied on for decades, upgrading is a lot more than switching out a few lines of code. It would require a complete revamp of the entire financial data systems while retaining backward compatibility.

There are also technical risks associated with using a new system for the first time. Any technical issues could result in major financial interruptions and market volatility. As such, upgrades would need to be gradual and tested.

It’s Not a Simple Port

The sheer costs of migrating the entire financial system to a new platform would be staggering. In the past, efforts to upgrade COBOL systems to more modern languages like Python or Java required the developer to write billions of lines of code and have resulted in short-term risks that have outweighed the long-term benefits.

Data Loss or Corruption

Another vital aspect of a migration is the fact that this financial data can’t be corrupted. Additionally, there are compatibility issues, like what to do if certain data won’t port into new systems. These operational hazards, coupled with a lack of COBOL developers, have left many institutions to retain their 60-year-old systems.

Shortage of Programmers

COBOL is an old programming language, and its programmers are also getting up there in age, with the average being 58 years old. This demographic is right around retirement age, meaning that the number of COBOL programmers is about to take a major nose-dive in the coming years.

Additionally, it’s become increasingly difficult to find schools teaching COBOL programming. This scenario leaves companies to fend for themselves or consider making the major upgrade in the coming years.

Have Some Banks Already Replaced It?

Not all financial institutions have taken a passive approach to the COBOL programming issues. There are some rare examples of banks and institutions shifting their operations to new protocols with varying success.

Bank of China

The Bank of China took a gradual approach to its COBOL upgrade. They began with the integration of API layers that enhanced the system’s core operations. This approach was smoother than its predecessors because it didn’t alter core operations or require massive overhauls.

Global Investment Bank Case

The Global Investment Bank Case successfully upgraded from COBOL to a modernized Linux cluster and Oracle RAC database. This maneuver saved the company +$125M yearly and improved uptime to 99%. In this instance, the company recouped its expenses in 14 months.

DBS

The DBS implementation is another example worth discussing. This bank migrated its operations to a hybrid cloud-mainframe that enabled the company to provide better global access to its operations and services. This enhancement did not completely replace COBOL but upgraded it with digital services.

Why IBM Stock Dropped: Anthropic’s COBOL Modernization Claim

It’s important to note that IBM has enjoyed decades of revenue from the use of COBOL. Perhaps that’s why IBM investors reacted so hard to news that a new COBOL AI programming system was launching. Notably,  Anthropics Cloud Code will support several key functions.

What Anthropic’s Claude Code Says It Can Do for COBOL

According to company documentation, this algorithm will support full automation, analysis, and documentation. Additionally, it will offer complete support for converting COBOL databases into modern languages like Python.

AI-powered COBOL Conversion Tool

The fact that this tool was specifically designed to recode databases into modern options is a sign that the end is approaching for COBOL. According to Anthropic, the system works by first scanning the entire COBOL code.

This step is used to map the dependencies, data flows, relationships, and execution paths of the systems in use. It also leverages sub-agents that run in parallel to enhance performance. According to Anthropic, the system is set up to convert COBOL logic to idiomatic Java with design patterns.

IBM’s Counter-Strike: watsonx Code Assistant for Z

While Anthropic’s Claude Code grabbed headlines, IBM has been deploying its own generative AI solution, Watsonx Code Assistant for Z, designed to modernize the very mainframes it built. Unlike general-purpose AI, watsonx is purpose-built for the IBM Z architecture, focusing on a “highly governed” transition.

• Syntax vs. System: IBM argues that while Claude can translate syntax, Watsonx understands the underlying z/OS ecosystem, including transaction processing monitors (CICS) and hierarchical databases (IMS) that are invisible to outside tools.
• Incremental Refactoring: The tool allows developers to selectively refactor COBOL into high-quality Java while maintaining “semantic equivalence.” This means the new code is guaranteed to produce the exact same financial results as the 40-year-old logic it replaced.
• The “Agentic” Shift: In early 2026, IBM updated the suite with agentic workflows that proactively map dependencies across millions of lines of code, reducing the “discovery phase” of modernization projects by up to 90%.

The Testing Bottleneck: Proving Equivalence in Production Systems

Perhaps the most important feature of the new ecosystem is the test environment. This option allows institutions to test their new database coding before sharing it with their clients. This comprehensive test suite enables companies to guarantee full interoperability and functionality, avoiding any potentially embarrassing launches.

Where IBM Makes Money in Legacy Systems (and Where It Doesn’t)

The introduction of COBOL helped IBM retain its dominance in the financial data sector. This programming language remains the industry standard and is one of the company’s most reliable revenue streams.

Is the End of COBOL Near?

The news of the AI system hit IBM investors like a ton of bricks. IBM shares fell about 13.2% on Feb 23, 2026, their steepest single-day percentage decline since Oct 18, 2000, after investors reacted to claims that AI tooling could accelerate COBOL modernization. The market reaction was clear – there’s strong demand for a better alternative to COBOL, and now that it’s become affordable, most companies will make the switch.

Additionally, the system can accomplish these tasks for a fraction of the cost of traditional COBOL developers, who can range from $500 a day and up. These lower costs and efficiency will erode many of the services that COBOL offers and cut into any new businesses seeking to enter the market.

This switch would leave IBM shut out of an industry that it dominated for over half a century. However, it’s vital to note that many believe this may be an overreaction, as there are still many inherent costs and risks that migration involves, regardless of improved coding ports.

Timeline Reality Check: Why Migration Could Still Take 5–10 Years

Despite the sudden market movements, most analysts agree that this investor’s knee-jerk reaction is a bit overplayed. The use of COBOL is so widespread and the systems it controls so vital to core aspects of society that any migration to a new system will undoubtedly be slowly undertaken.

The amount of testing that will be required to finalize a migration across the entire sector will also take some time. As such, even with more capable alternatives and with an AI-powered conversion tool, it could be nearly a decade before real changes are made.

Innovative Companies in the Database Sector

Market research estimates the global database market at roughly $171B in 2026, with continued growth expected over the next several years as AI workloads expand data infrastructure demands. The success of the industry can be narrowed down to a few top-performing companies and firms that continually invest in R&D, such as Oracle (ORCL +2.63%) and Microsoft SQL Server (MSFT +2.23%).

IBM

The history of IBM began over 100 years ago in 1911. That’s when the firm launched as the Computing-Tabulating-Recording Company (CTR). It was founded by a well-known investor, Charles Ranlett Flint, with the goal of pushing the emerging technology forward.

In 1924, the company rebranded to IBM (International Business Machines). This new name was part of a broader shift towards business computing under the direction of Thomas J. Watson Sr. By 1951, the company had succeeded in launching the first scientific computer.

IBM dominated the computing sector for decades. It only began to lose dominance when computers shifted from large room-sized devices towards smaller options. IBM was heavily invested and was leading the market in these larger devices, so there was little economic incentive to change its model.

Today, IBM is just a glimmer of its former glory. However, the company is still considered an innovative force in the market by investors and has a strong presence in the financial and government sector where legacy systems still reign supreme.

Those seeking a computing company with a rich history that spans over a century should do more research into IBM shares. The firm has recently expanded into AI, blockchain, and quantum computing, signaling future innovations to the market.

Latest IBM (IBM) Stock News and Developments

What is COBOL and Why are IBM Investors Panicking | Conclusion

It’s easy to see why IBM could potentially lose some funding in the future due to this AI system. For years, COBOL programming has remained limited to an exclusive and aging class of programmers who charge premium rates for their services. Now, any company will have the ability to modernize these systems affordably.

Learn about other important market news here.