Thirteen systems have to go [16], and the numbers doing the rounds in the trade press, which the bank itself has not put its name to, put the consolidation spend at roughly EUR600m [8]. That is about EUR46m for each system switched off [17], a reasonable price for an estate assembled over decades, and it is being spent under an agreement that runs ten years [6]. On the calendar the source gives, the deal begins in 2026 [19], so the partnership outlives the 2028 deadline by eight years; if the promised run-rate saving held across those eight years it would come to about EUR2.4bn against the EUR600m outlay [20].
What decides whether the EUR300m arrives is the decommissioning calendar rather than the platform. Legacy systems and Vault Core are to run side by side through the transition to protect service continuity [5], products begin moving across in phases only in 2027 [4], and a cost line falls when an old system is turned off rather than when a new one goes live, which compresses thirteen decommissionings into a window of roughly two years [21]. Or rather, the more interesting version of the question is what counts as one of the fifteen, because an estate the bank defined is an estate the bank can recount, and the first products in the queue are everyday accounts such as call-money deposits [9], the least demanding entries in the book.
Yiping Li, the division's chief operating officer, called the selection the moment strategy turns into delivery [13]. Delivery here is GFT Technologies, appointed transformation and systems-integration partner on the strength of more than 25 years of work with Deutsche Bank, running a build-and-migration team across onshore, nearshore and offshore locations [7]. This is probably wrong, but I read the still-unnamed second platform [10] as holding the harder half of the estate rather than the residue, because the products that resist a clean split between product configuration and core ledger [12] are the ones you name last. An alternative reading holds that the second slot stays open on purpose, so the first vendor's build sets the integration standard and the second negotiation comes in cheaper. Either way the test is visible from outside the bank. Call-money deposits live on Vault Core during 2027, plus a named second platform, would make the 2028 figure real; another year of testing with no name attached pushes the EUR300m figure into 2029.
