Put the run rate over the revenue and the shape of the commitment becomes visible: technology is running at roughly 11 percent of M&T's 2025 top line, and close to half a percent of its assets [1][7]. Held at that level across the eight years the effort has run, the cumulative figure is on the order of $8bn, and the early years were almost certainly smaller than the current pace [2]. That is a standing claim on the expense base for the better part of a decade, which is a different kind of approval from a program budget.
Wisler calls the early years "the hard stuff that nobody cares about" [7], and the throughput number is worth reading mechanically to see why. Quadrupling releases at constant headcount works out to a factor of about 4.3, or something near 260 releases on an average working day [3][4]. Both of the things his team attacked released capacity rather than adding it: an outage baseline that a 90 percent reduction puts at roughly ten a year [5], and the 30 percent of assets the bank had classified as non-destination, now in the low single digits [12]. Effort that had gone into incident response and into systems the bank intended to retire became effort that shipped.
The claim on the table is Wisler's own, that technology has "moved from some version of a liability to a real asset for the bank" [6], and a skeptic has two fair objections ready. These figures come from one executive describing his own record in a Forbes interview, with no audited disclosure behind them [19], and release count is among the easiest metrics to move by cutting batch sizes rather than delivering more. Both points stand. What survives them is the sequence, which is the transferable part, and the fact that the baseline Wisler volunteers, well over 100 outages a year on arrival, is unflattering to the institution whose technology he now runs [8].
The staffing choice is the clearest tradeoff in the account. Roughly one engineer in six is now external [6], where the bank had previously leaned heavily on contractors and allocated capacity through annual competitions for a few large projects [9]. Contract labor is a cost you can stop inside a quarter; payroll engineers are a cost you carry through the cycle, and more than 350 computer science graduates hired through the Technology Development Program are a commitment to keep training people who are not yet productive [15]. M&T bought retention of capability and gave up the ability to flex spending down quickly, which a board should choose deliberately rather than discover in a bad quarter.
The order of operations is where this record argues with most transformation plans. Wisler says he resisted the build-versus-buy, cloud-timing and agile questions until leadership had settled what it believed about how banks now compete [14], and by his account the next phase only became obvious once debt was down and resiliency, speed and quality were up [17]. An agenda laid on top of a hundred outages a year inherits the hundred outages, so the reliability bill comes due before anything sitting above it is worth funding [8]. This quarter's decision to defer that spending is next quarter's decision to explain why the new capability does not work when a customer touches it.
