The Innovation Tax: Why Banks Aren't Actually Transforming
Every bank says it is innovating. Boards approve nine-figure tech budgets. Here is the number nobody puts on the slide: only 8-12% of that budget ever reaches something genuinely new.
"You're not paying for innovation. You're paying ransom to the 1970s."
Large-scale studies of bank IT spending estimate that 70-80% of technology budgets are consumed by maintaining legacy infrastructure and mandatory change-the-bank work, leaving 20-30% for actual innovation. Global banks now spend 6-12% of revenue on IT. JPMorgan, the archetypal tech-forward bank, will spend about $19.8bn on technology in 2026.
When a bank says "we're investing in AI and digital transformation," the honest translation is: we're spending one dollar in ten trying to build the future, while the other nine keep the past from collapsing. The innovation tax isn't a budget line. It's the structural price of inherited infrastructure and regulation.
The dinosaur problem
Around 43% of global banking and financial systems still run on COBOL, which processes about 95% of ATM transactions. Legacy core maintenance spend hit $36.7bn in 2022 and is projected to reach $57bn by 2028 if banks keep patching rather than rearchitecting.
The compliance tax
- →DORA: €5-15m per institution for implementation; per-employee costs approaching $10,000 per year for larger firms.
- →MiCA: around a third of European crypto companies now report annual compliance costs above €500,000, crowding out R&D for smaller players.
The customer is changing faster than the stack
81% of UK Gen Z already use their bank's mobile app; only 21% still use branches. Roughly 45% of millennials and Gen Z globally say they only bank digitally. In the US alone, women are expected to control much of $30 trillion in baby-boomer financial assets by 2030. The innovation tax is being charged at the precise moment your customer set fragments by generation, gender and life stage.
The exceptions tell you what actually works
UBS used the Credit Suisse absorption as a demolition permit: 700 applications decommissioned, ~56% of the acquired legacy stack retired, $9.1bn of gross cost synergies already realised on the way to $13bn by 2026.
DBS chose the opposite path: a long, deliberate strangler-fig rebuild around a microservices and API-first architecture. CEO Piyush Gupta describes DBS as "more like a technology company offering financial services than a bank."
Pictet shows the private-bank flavour: tradition plus technology, a dedicated Pictet Tech division, and AI embedded in investment and client-advisory processes. Julius Baer has invested over $1bn since 2014 and is standardising on Temenos for its Swiss core.
The actual playbook
- →Stop trying to replace the core. Wrap it.
- →Treat cores, payments and basic servicing as utilities. Reserve scarce internal engineering for the 20% clients actually experience as differentiated.
- →Measure transformation in decommissioned legacy - not new features shipped.
How we work with banks on this
The innovation tax is structural, so the response has to be structural too. Our work with banks and financial institutions re-architects the 10%: every change-the-bank euro tied to legacy decommissioning, regulatory automation or new revenue lines - not just "more features." We convene the shareholders, boards and operators who can actually approve those architectures at our London (Oct 2026) and Geneva (Apr 2027) editions.
The full long read - with all figures, sources and case studies - is published on LinkedIn.
The Innovation Tax: Why Banks Aren't Actually Transforming - on LinkedIn
Full data, sources and case studies from UBS, DBS, Pictet and Julius Baer. Subscribe to The Game Changer Intelligence to get every long read the moment it is published.