According to TheRegister.com, the UK’s National Savings & Investments (NS&I) bank has a transformation program that is now a staggering £1.3 billion over budget and four years late. Chief executive Dax Harkins told the Public Accounts Committee the project is forecast to blow past both its March 2028 deadline and its agreed budget from last year’s Spending Review. The program, which includes a new core banking engine, was originally called Project Rainbow and began in 2020. In a new letter to MPs, Harkins said options to fix the mess include reducing connections to older legacy systems and reprioritizing work. The total cost is expected to hit £3 billion by 2030-31, with the main contractor being Atos, which holds a £474.4 million contract, while systems integrator Capgemini is involved in the rescue planning.
The Legacy Anchor
Here’s the core of the problem, and it’s a classic one in big tech migrations: legacy systems. The plan to intricately connect brand new services with NS&I’s old 20-year-old banking platforms is sinking the whole project. It’s so complex and risky that the National Audit Office slammed it last year, warning it carries an “extremely high level of risk.” So now, the “option” on the table is basically to cut the cord. Less integration. Fewer links. Maybe even leaving some old tech in a corner to gather dust. It’s a desperate move, but when you’re this deep in a hole, you stop digging.
The Incumbent Trap
And then there’s the supplier drama. Look at Atos. They were the incumbent, running the old outsourced system. Their contract kept getting extended without competition—first to 2024, then to 2028. Then, when NS&I went to market for a new core banking system, who won the nearly half-billion-pound contract? Atos. Why? Because, as NS&I said, they were the only ones who knew the “interdependent leveraged technology” of the old platforms. It’s the ultimate vendor lock-in story. You hire the company that built the maze to lead you out of it, and then seem surprised when they can’t. Now Capgemini is being brought in to help find a way out. It’s a mess.
What Happens Next?
So what’s the plan? More analysis, of course. Harkins says they’re working with Capgemini to finish reviewing options by March 2026, hoping for Treasury approval by June 2026. That timeline itself is telling—we’re talking about another year and a half just to figure out *what* the new plan even is. The final decision goes to ministers, which adds another layer of political risk. Basically, they’ve burned through a mountain of cash and half a decade to get to a point where they might start simplifying the original, failed plan. It’s a brutal lesson in how not to run a digital transformation, especially for a critical piece of public financial infrastructure. You have to wonder, at what point does someone just pull the plug and start over?
