Finance Bets Big on AI, But Bills Keep Rising

Artificial intelligence adoption is surging across Europe's finance industry, with banks, fund managers, and traders experimenting with AI for tasks ranging from analyst recommendation compilation to code generation for quant traders. However, a significant hurdle has emerged: the escalating cost of AI, primarily driven by supply constraints in computing power. Users of popular AI assistants like Anthropic's Claude are experiencing substantial price increases, with annual bills for a single firm potentially reaching millions of dollars.

This rapid adoption and the associated high expenses are creating new challenges for financial institutions. While AI agents offer advantages, their cost adds pressure on profitability, necessitating cost-cutting elsewhere. There are also growing concerns about becoming dependent on a few dominant AI companies, particularly American ones, as core IT skills shift to external suppliers. This situation mirrors initial industry reactions to cloud computing, where some in-house development eventually occurred without crippling provider profits.

In response to these mounting costs, finance firms are beginning to re-evaluate their AI strategies. Some are shifting towards developing in-house models for tasks that do not require cutting-edge, large language models. This approach could foster a more mature phase of AI integration, moving away from a 'token-maxxing' mindset of unchecked spending. There is even speculation that firms might collaborate to share AI costs and expertise, a departure from the industry's traditional rivalry, potentially leading to consolidation, especially in Europe's fragmented banking sector.