What the Rising Costs of AI Means for Government

AI Costs Can Quickly Add Up for Government Agencies

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 that provide comfort in charging ahead. So, while state and local government continues to congregate at the AI starting line, other barriers lie just ahead, and chief among them will be cost.  

AI costs can be viewed as per person (seat), by application or via add-on, such as Microsoft’s Copilot. AI can be expensive, even at $20 per seat per month, especially if everyone is afforded a license. A government with 2,000 employees can expect to pay $480,000 per year. If that per-person license were $30 per user per month, that same government entity with 2,000 employees would have to pay $720,000 annually. 

To complicate matters, no one AI vendor does it all. Many employees will need additional subscriptions to numerous 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 classifications of employees may need different types of licenses. But even with AI provisioning, AI costs will not be going down but up — like Uber, which eventually no longer needed to offer low entry pricing to gain market attention and share. 

READ MORE: 7 tips for a successful AIOps implementation.

AI Companies Will Seek to Become Profitable

Companies that provide exclusive AI services have enormous overhead and operating costs. These costs include software development, AI system learning, knowledge and information compilation and digestion, and computer hardware usage, including 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.

Of course, this list doesn’t include Google’s, Amazon’s or IBM’s AI offerings. However, the billions of dollars raised from the six companies listed above amount to about $25 billion in venture capital. 

These companies will eventually be expected to realize a profit and satisfy investor expectations. After all, it took Uber 14 years to become profitable. And what might this mean for state and local governments when pricing AI products and services? How will these AI companies manage their debt, remain in business and stay competitive? 

How to Address the Rising Costs of AI Solutions

State and local governments are still playing catch-up in a post-pandemic environment where legacy systems were on full display, reaffirming the need to modernize further. And this was before AI ever 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. Leveraging Federal Funding and Grants

• Infrastructure investment: While not a long-term strategy, governments can tap into federal programs that support technology and infrastructure, such as grants from the U.S. Department of Transportation, the U.S. Department of Commerce or the National Science Foundation for AI-related initiatives.
• Public safety and justice programs: AI tools for policing, firefighting and emergency response may qualify for federal funding under Homeland Security or community policing grants.

2. Public-Private Partnerships 

• Collaborating with private 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 public data access or pilot testing in real-world government settings.

3. Regional Cooperation

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

RELATED: NASCIO releases ‘AI blueprint’ for state agencies.

4. Open-Source and Low-Cost Solutions

• Governments 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 AI models and tools that can be adapted for public sector use at a fraction of the cost of proprietary systems.

5. Phased Implementation

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

6. Cost-Benefit Analysis and ROI Focus

• Conducting rigorous cost-benefit analyses can prioritize AI investments with the highest ROI (e.g., reducing fraud, optimizing traffic, or improving public health outcomes).
• Demonstrating ROI can help secure additional funding from taxpayers or state legislatures.

LEARN MORE: AI can transform the citizen experience.

7. Budget Reallocation and Efficiency Gains

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

8. Provisioned Workstations

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

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

UP NEXT: AI isn’t new to cybersecurity, but some of its use cases are.