Global banking has just posted the most profitable year of any industry – US$1.3 trillion in net income – and still trades at the lowest price-to-book of any industry. McKinsey’s preview of its Global Banking Annual Review 2026 reads that gap plainly: investors are delighted with the results and unconvinced the advantage will last.

The report shows why. Banking’s real competitive moat was never property. It was inertia – trust built over decades, relationships too tangled to unwind, customers who never got around to chasing a better rate. Four forces reached a tipping point in 2025, and every one attacks that inertia directly. Mature fintechs have taken 17 per cent of the revenue they share with the top thousand banks. Neobanks like Nubank (131 million customers) and Revolut (69 million) now beat incumbents on growth, on returns and, increasingly, on trust. Stablecoins loosen the deposit base. And agentic AI dissolves stickiness itself: software agents that monitor balances in real time, sweep idle cash into higher-yield accounts, shift card balances and compare products while the customer sleeps. Banks survived the internet and the smartphone by moving at the pace of their older, high-value customers. That grace period is gone – generative AI reached roughly half of US working-age adults within three years, across every age group at once. When technology can dissolve your switching costs without infringing a single right you hold, what remains is what you actually own.

When inertia dissolves, ownership decides: the intellectual property strategy lesson

The durable lesson is not about banking. Any business whose advantage rests on habit, incumbency or switching costs is holding a moat that technology can drain without ever infringing it, because inertia cannot be registered, licensed or enforced.

Now look at where McKinsey says the surviving value sits – hyperpersonalised engagement built on proprietary customer data, platforms that reach beyond the core product, agentic operations, a three-speed innovation portfolio – and you are reading a list of intangible assets. Each one is protectable: the data and the models trained on it, the operational know-how as trade secrets, the platform technology as patents, and the brand that customers and their AI agents must still choose to trust as registered trade marks. It is the same shift I examined in When AI Does the Shopping, What Does Your Brand Actually Own? – when an AI intermediary sits between you and the customer, owned rights are what travel – and in When Everyone Can Run the Model: What Open-Weight AI Means for Your IP Strategy – when a capability commoditises, value moves to cost position, proprietary data and whoever owns the customer relationship. Banks already spend more on technology than the next four sectors combined; the strategic question is how much of that spend ends the year as owned, enforceable assets rather than vendor licences and undocumented practice – the discipline set out in When AI is embedded in your Workforce, Trade Secrets Become the Strategy. Three questions for your next strategy review, whatever your industry:

– Which parts of your customer relationship would survive the arrival of an AI agent acting for that customer, and which are merely habit?

– What did last year’s technology spend leave behind that you own – filings, controlled trade secrets, data rights – rather than rent?

– If a competitor can buy the same AI you can, what in your stack could they still not copy?

Investors are already pricing the answers. The time to convert record profits into owned advantage is while the profits are still records.

Read the article: Global Banking Annual Review 2026: Precision with speed, McKinsey & Company, 21 May 2026.