The COBOL Guy | Making COBOL Cool Again 🚀 | WE help corporations maintain & secure their mission-critical COBOL systems with AI Domain specific tools - because stability beats risky rewriting.
COBOL isn't dead. In fact, it's running your money right now. The reality nobody talks about: → 95% of ATM transactions run on COBOL → 43% of banking systems depend on it → $3 trillion in daily transactions processed → 70% of business systems still use it But here's the real problem... The average COBOL developer is 55+ years old. They're retiring fast. New grads aren't learning it. Most companies are panicking, rushing to "modernize." But the smart ones know better... Because rewriting these systems costs millions more than maintaining them. And most modernization projects? They fail. While others scramble tomorrow, the prepared companies are quietly investing in COBOL expertise today. That's how the biggest organizations stay ahead - they see opportunity where others see obsolescence.
Ok bro.
Well said Juan Lucas
To my opinion traditional COBOL banking applications typically run with excellent performance, achieving well under one CPU second for individual transactions. The bookings are then processed in an end-of-day run. However, modern applications increasingly target a single LUW transaction with intraday bookings. In my experience, such an LUW transaction takes 15–25 seconds, even with optimal programming. I believe the desired intraday booking is often fictitious, so any errors that occur often have to be corrected in end-of-day processing. This is often so slow that in some cases, more than 24 hours of processing time are required. 😉
I'm sure AI will be able to maintain all those systems, not.
Don't forget about how much more it'll cost to troubleshoot, bug-fix, and maintain a java replacement...
It appears this process is happening where I work. Can’t wait to see how it turns out lol
The irony is that COBOL’s biggest weakness today isn’t technical—it’s generational. The code runs, the systems are rock-solid, but the talent pipeline is vanishing. That’s a people problem, not a tech problem.
I programmed in COBOL with JCL/ECL from 1969 thru 1994. Tandem COBOL and SCOBOl from 1986 to 1989. I programmed in Assembler from 1965 thru 1974. .Next the following languages: Pascal, C Language, Tandem's Transaction Application Language (TAL), C Language, VBScript/ASP, PERL, Unix Shell Languages, some VB Dot Net and Java. From 1997 to present I have been a Full Stack Web Developer. I also developed using Oracle Forms and Reports (3 years) and Oracle Application Express (APEX) from 2004 to present. I have 20 + years developing Oracle SQL and PL/SQL. I have 5 years developing with T-SQL I cannot get hired because of age discrimination now. I would very much like to work again. I am confident that with a little refreshing, I could be very useful as a COBOL developer. I live in Pensacola Florida and can be reached at: dennis.williamson@gmail.com
IT Programming and Solutions Consultant
2moWhile the 'upgrade costs more than our entire annual revenue' is a bit of an exaggeration, the truth is with every single Cobol system that I have witnessed being replaced with a more 'modern' system, the new system costs at least 4 times more just in software and infrastructure costs, and far more in training costs and loss of on-site expertise. And half of these new systems had reduced functionality, which required additional software to address the shortcomings. Cobol systems are not for every enterprise, but a well-designed Cobol system should not be replaced without looking at all factors, including costs.