Escalating AI Costs Shouldn’t Hinder Finance Sector’s Adoption

Don’t Expect the AI Cost of Entry to Drop Anytime Soon

As we begin 2025 in earnest, the biggest stumbling blocks for AI adoption seem to be fear of risk, lack of a safe environment for testing AI services, and waiting for policies and guidelines to provide support for moving forward. So, while financial institutions continue to congregate at the AI starting line, other barriers lie just ahead; chief among them will be cost.

AI costs can be assessed per person (“per seat”), by application or as an add-on, as Microsoft charges for Copilot. AI can be expensive. At $20 per seat per month, if everyone is given a license, a bank with 2,000 employees can expect to pay $480,000 per year. At $30 per user per month, that jumps to $720,000 annually.

To complicate matters, no single AI vendor does it all. Many employees will need subscriptions to multiple AI services, similar to what consumers face when choosing entertainment streaming services.

Of course, one can argue that not every employee needs AI access. Different staff roles may need different licenses. But even with AI provisioning, costs will be going up, not down — much like Uber, which eventually stopped offering low entry pricing to gain attention and market share.

DISCOVER: Microsoft Copilot helps workers automate mundane tasks.

Profits Aren’t a Priority for AI Companies, Yet

Companies that exclusively provide AI services have enormous overhead and operating costs. These include software development, AI system learning, knowledge and information compilation and digestion, and computer hardware use, which entails unprecedented energy consumption. The entry-level teaser rates offered today are simply not sustainable.

This chart shows just some of the leading AI companies in terms of money raised through the end of 2024.

Notably absent from this list are heavyweights Google, Amazon and IBM. However, the funds raised by the six companies listed above amount to about $25 billion in venture capital.

Eventually, these companies will be expected to make a profit and satisfy their investors (it took Uber 14 years to become profitable). What might this mean for financial institutions as they price AI products and services? How will these AI companies manage their debt, remain in business and stay competitive?

Click the banner below to keep reading stories from our new publication BizTech: Financial Services.

Strategies for Blunting the Costs of AI

Banks are still playing catch-up in a post-pandemic environment where legacy systems were on full display, reaffirming the need to modernize — and this was before AI entered the picture. Technology managers are beginning to express worry about AI affordability.

Here are some possible remedies for dealing with the inevitable rising cost of AI.

1. Industry Partnerships

• Collaborating with companies to share costs and risks can make AI more accessible. For instance, companies may provide AI tools at a lower cost in exchange for data access or pilot testing in real-world finance settings.

2. Regional Cooperation

• Financial institutions can form consortia to pool resources and share AI infrastructure such as cloud computing and data sets for training, reducing individual costs.
• Shared AI procurement agreements could help firms negotiate better pricing through economies of scale.

3. Open-Source and Low-Cost Solutions

• Banks can turn to open-source AI tools and platforms, which are often less expensive and customizable.
• Nonprofit organizations and academic institutions are expected to develop open AI models and tools that can be adapted for financial use at a fraction of the cost of proprietary systems.

4. Phased Implementation

• Introducing AI incrementally allows banks to spread costs over time and focus on high-impact use cases first, reducing initial financial burdens.

UP NEXT: What is voice-activated banking.

5. Cost-Benefit Analysis and ROI Focus

• Conducting rigorous cost-benefit analyses can prioritize AI investments with the highest ROI (for example, reducing fraud or optimizing traffic).

6. Budget Reallocation and Efficiency Gains

• Redirecting funds from outdated or inefficient processes and technologies to AI initiatives can free up resources.
• Banks can also use AI to identify and eliminate inefficiencies, creating savings to reinvest.

7. Provisioned Workstations

• Create dedicated AI spaces to serve staff on a provisional basis. Employees can schedule time at workstations loaded with different AI services.

By combining these strategies, financial institutions can better position themselves to adopt and sustain AI technologies despite rising costs. It’s clear that the cost of AI will increase, presenting complex challenges. It’s not too early to lay the foundation for sensible use of AI technologies so that managers can plan for future requirements and budget accordingly.