With global IT spending expected to reach $6.15 trillion in 2026, according to Gartner, our focus is starting to shift from “What can AI do for us?” “How much does AI cost?”
deloitte warns that AI is now the fastest growing IT spend, consuming up to half of IT budgets in some companies. When employees use five, ten, or more AI tools, often without IT support or central oversight, costs skyrocket, adding not only dollars but reputational, administrative, and hidden risks.
Untracked Spending and the Real Cost of AI
To reduce costs, teams often look to: open source modelcompanies limit who can access it. Every day we hear about employees relying on unapproved AI tools to do their jobs more efficiently, increasing risk unchecked. Anyone can instantly talk to ChatGPT via Siri on their iPhone. But if an employee does that with sensitive company data, who’s in control? Not you.
Shadow AI Risks are increasing, and poor access and data management are leading to large-scale losses. Violate. Supply Chain From malicious code injected into open source large-scale language models such as LiteLLM compromiseto viral AI tools like Moltbook. Thousands of API keys leaked To hackers. IBM reports that the global average cost of a data breach is $4.4 million. AI adoption is outpacing AI governance.
McKinsey has found that the use of AI is on the rise. 88% of companies Now we use technology. If you’re not already using AI, you should heed philosopher Meister Eckhart’s warning that “the cost of doing nothing is far greater than the cost of making a mistake.”
Notable costs
Per-user pricing is the status quo that major companies want to maintain. Their apps, interfaces, and your data. “We spend about $3,500 a month,” said one airline IT executive. [about] With 60 employees, we are always at our limit and have to wait. ” Reaching your spending limit is like eating a hot dog and having to stop halfway, only to receive a notification that you have to pay extra.
In my experience, it reduces productivity. When you finally get things going, you get the dreaded message: “You’ve reached your plan limit. Your plan will reset in 3 hours.” Power users like me will hit the limit, light users will barely hit the limit, but the cost is the same. We are losing productivity and cash.
I spoke to a leading global provider of IT services and digital solutions with over 3,500 employees in 36 countries. We discussed the increasing costs of AI and what customers are seeing and experiencing. They found that most tasks performed by AI can be performed by lower-cost models, but accessing those models in addition to premium models creates management overload.
Direct implementation of premium models is best suited for manufacturing and logistics applications where machines make decisions and manipulate data, but they also see a spike in costs. I spoke with a popular yogurt company that partners with one of the major providers. GenAI Automation in manufacturing and logistics processes. He said his company had to spend millions of dollars just to set up the data correctly for model ingestion.
What if a competing provider has a better-performing or lower-cost solution? You’ve exhausted your data management budget and are locked in. I hope there’s some reuse out there.
The problem is not the AI itself.
Big vendors want users to be locked into their models, pricing, and data silos. Everyone is touting their AI as the best at this or that. deep seek has been proven You don’t need billions to compete. There are some basic steps you can take today to avoid the worst problems and give your company the best that these tools have to offer.
Step 1. Tackle Shadow AI head-on.
This is a cost you are already paying in risk and inefficiency. Talk to your employees, perform a complete audit, and gather feedback on the tools they’re using both on and off the books. The goal here is not to eliminate or limit what your employees are using, but to identify and provide sanctioned options that eliminate the need to look for external tools.
Step 2. Build for flexibility, not rigidity
Can I change the model? Are you always in control of your data? Can you integrate multiple providers and tools? If the answer is no, move on. Many third-party providers are emerging that integrate multiple options into a single tool or interface.
Step 3. Prioritize governance over gatekeeping
The technology industry moves fast. Your goal is to remain flexible and secure, and move away from a “how do we deploy AI?” mindset. “How do you manage AI?” If you understand how your team uses AI, AI integration can potentially reduce your SaaS bill by 20% to 40%. The companies reaping the most benefits from AI are not the ones spending the most. They are people who spend their money wisely. Ex ante governance has long-term benefits.
AI hangovers are real, but preventable.
Spending on AI is hurting companies across all sectors. With the rise of AI-supported data centers, electricity and transportation costs will hit everyone. 7% increase Until December. By leveraging AI wisely, we can all ease these burdens. It starts with honest conversations about what we’re paying for, its benefits, how and why shadow AI thrives, and what pricing models reward rather than penalize productivity.
