The entire tech world seems to have gone collectively mad for Artificial Intelligence, throwing money at anything that smells like a neural network. The narrative is dominated by the hyperscalers—the giants building the colossal digital brains. But what if we’re all looking in the wrong direction? What if the real winners of this AI gold rush aren’t the ones selling the picks and shovels, but the ones who quietly use them to find gold?
This isn’t just a hypothetical; it’s a compelling argument for a significant AI banking advantage. As John Bailer, manager of the £1.2bn BNY Mellon US Equity Income fund, pointed out in a recent Trustnet piece, the current frenzy has an unnerving echo of the dot-com bubble. We’re seeing a land grab in banking infrastructure for AI, but the long-term value might lie elsewhere entirely.
The Dot-Com Déjà Vu: A Lesson in Infrastructure
Remember the late 1990s? Companies spent billions laying fibre optic cables, convinced they were building the very motorways of the new internet economy. For a while, their stock prices went through the roof. Then, reality hit. As Bailer notes, “You just invest too much, and then prices collapse because there’s too much competition”. The motorway builders went bust, but the companies that used those motorways to build incredible businesses—think Amazon and Google—went on to define the next two decades.
This is the core of Bailer’s argument, and frankly, it’s a compelling one. He believes the same pattern is unfolding now with AI. The hyperscalers are in an arms race to build the biggest, most powerful AI models, spending eye-watering sums on chips and data centres. They are today’s fibre optic cable companies.
The real, sustainable value, however, often accrues to the users of the technology, not the builders. It’s the organisations that can integrate AI into their existing workflows to become more efficient, create better products, and serve customers more effectively. And who is perfectly positioned to do just that? The financial sector.
AI is Not Disruption; It’s an Accelerator
For banks, AI isn’t a threat of disruption; it’s a tool for acceleration. This is the crucial point of competitive differentiation. A bank doesn’t need to build its own foundational model from scratch, any more than Amazon needed to manufacture its own fibre optic cables. Instead, it can license the technology and apply it where it has a unique advantage: its vast troves of proprietary data, its established customer relationships, and its deep understanding of risk.
This is where we see potent income generation models emerging within financial sector AI. Consider these applications:
– Hyper-Personalised Services: Using AI to analyse a customer’s spending habits to offer tailored financial advice, loan products, or investment opportunities. This drives engagement and loyalty.
– Enhanced Underwriting: AI models can process thousands more data points than a human underwriter, leading to more accurate risk assessments for loans and insurance. This directly improves the bottom line by reducing defaults.
– Fraud Detection: AI algorithms can spot anomalous transactions in real-time, saving banks and their customers billions annually.
– Operational Efficiency: Automating back-office tasks, compliance checks, and customer service inquiries frees up human capital to focus on higher-value work.
None of this requires a bank to compete with Google on AI research. It simply requires them to be smart integrators of technology to fortify their existing business. This is the essence of the AI banking advantage: leveraging new tools to do the old job better, faster, and more profitably.
Why Some ‘Tech’ Bets Are Smarter Than Others
This perspective also reshapes how we view tech investments. Bailer’s fund, for instance, has largely sidestepped the ‘Magnificent Seven’ this year. While those stocks are down about 4%, his top seven holdings are up 8%, according to his interview with Trustnet.
He highlights Cisco as a prime example of a smart ‘picks and shovels’ play. Why? Because Cisco’s success isn’t just tied to the hyperscaler arms race. It sells essential networking gear to a broad base of enterprise customers—including banks—that are upgrading their own internal banking infrastructure to handle new AI workloads. It’s a more diversified and, arguably, more durable business model than simply betting on who wins the AI model war.
This approach—backing the “enablers” and the “users” over the pure infrastructure builders—seems to be paying off. It’s a strategy that prizes tangible productivity gains over speculative hype.
A Regulatory Tailwind on the Horizon?
There’s another factor that could supercharge the AI banking advantage: a changing regulatory environment. In the United States, there’s a growing expectation of regulatory easing, particularly for regional banks. This could unlock a wave of mergers and acquisitions (M&A).
What does that have to do with AI? Everything. AI tools are perfectly suited to streamline the integration process of two merging institutions. From consolidating customer data to harmonising IT systems and identifying operational redundancies, AI can make M&A cheaper, faster, and more effective.
If this wave of consolidation happens, the banks that have already mastered AI integration will be in pole position. They’ll be the acquirers, able to absorb competitors efficiently and scale their superior operating models. This creates a powerful flywheel effect, further cementing their market leadership.
The Real AI Revolution Will Be…Profitable
So, while the headlines are filled with the drama of warring AI labs and trillion-dollar valuations for chipmakers, the real revolution might be happening more quietly inside the balance sheets of the world’s financial institutions. The true measure of AI’s impact won’t be the size of the model, but the tangible return on investment it generates.
The story of the next decade might not be about which hyperscaler wins. It might be about how centuries-old banks used this new technology to achieve a new level of profitability and competitive differentiation. They have the data, the customers, and the business case. All they need to do is execute.
The question for investors and industry watchers isn’t just “Who is building the best AI?”. It should also be, “Who is using AI the best?”. Based on the historical precedent and the current landscape, the financial sector has a surprisingly strong claim to the throne. What do you think—are banks the dark horses in this AI race?
