Is it worth learning COBOL in 2026?

Yes, with conditions. There is genuine demand driven by retirements, and there is far less competition than in mainstream languages. But COBOL alone is a narrow bet.

What makes it a good bet

• The code is not going away quickly in banking, insurance and government.
• The people who understand specific applications are retiring faster than replacements arrive.
• Employers in this space frequently train from zero, which is unusual.

What makes it a poor bet on its own

• Roles are geographically concentrated and industry-concentrated.
• Pure maintenance work can be repetitive and limit your growth.
• Some organisations are genuinely migrating away, on long timescales.