JPMorgan has tightened its artificial intelligence spending by implementing stricter internal cost controls and shifting from unlimited AI experimentation to disciplined budgeting. [1, 2, 3, 4, 5] The bank is actively hiring engineers and data architects in Singapore to develop its proprietary AI systems. [2, 3, 4, 5]
AI computing costs, measured in so-called “tokens,” have become a significant and unpredictable factor in corporate technology budgets, even as the price per token declines. [1, 2, 3, 4, 5] To manage this, JPMorgan is aligning token usage with the financial returns of AI projects, reserving expensive tokens for complex, revenue-linked applications. [1, 2, 3, 4, 5] As co-head of JPMorgan Global Payments Max Neukirchen explained in a June interview in Singapore, "The days where everybody could just do anything are probably over. We are managing the AI prompt generation with a lot more discipline and thoughtfulness." [1]
Routine, lower-value tasks like summarizing news utilize cheaper AI models, while high-value functions such as real-time corporate treasury cash flow forecasting get allocated higher-cost tokens. [1, 2, 3, 4, 5] JPMorgan is also boosting resources to combat AI-driven fraud and partners with Anthropic’s Project Glasswing to access cutting-edge cybersecurity AI models. [1, 2, 3, 4, 5]
JPMorgan CEO Jamie Dimon stressed on CNBC in late July that companies must treat AI token spending as a business resource. "They all see the costs going up rapidly. So, of course, we're all going to be rational about it like any other resource we use," he said. [6] Dimon also said the bank is "very protective of our data and our IP," emphasizing its commitment to safeguarding customer information while negotiating AI vendor contracts closely. [6]
The financial sector overall is expected to introduce tighter AI cost controls as AI expenses rise. [1, 3, 4, 5] JPMorgan’s efforts reflect this trend, combining spending discipline with strategic hiring in growth hubs like Singapore.
In June, Neukirchen detailed the shift during his Singapore interview, and Dimon’s CNBC remarks in late July reinforced the bank’s measured approach. Hiring of Singapore-based AI specialists continues as JPMorgan builds internal capacity for proprietary AI development. [1, 2, 3, 4, 5, 6]
