To go the AI route or not

A chatbot demo is easy to show off. Running AI across an entire company is something else entirely. That is where the real cost begins. Gartner expects global AI spending to hit $2.52 trillion in 2026, a 44% jump in just one year, with roughly $1.37 trillion going into infrastructure alone. In Europe, companies are also working under tighter rules shaped by the EU AI Act.

At the same time, another pressure keeps building in the background: the rising prices of compute, data centers, integration work and energy. And that changes the entire management question.

Many organizations are still asking how fast they can scale AI, but the more useful question is less exciting: Does this problem really need AI at all? Some issues do. Many do not. In quite a few cases, the better move is to simplify a workflow, automate a repetitive step, improve search, or give teams a dashboard they can actually use.

The cost profile also changes faster than people expect.

A pilot can look manageable. Production is another story. The bill quickly stops being just about model access and starts including integration, monitoring, cybersecurity, evaluation, human oversight, vendor dependency, and compute capacity. That is often where an initiative that looked compelling in a slide deck begins to feel slower, heavier, and more fragile in real life.

The wider market data tells the same story. Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027. Reuters has also reported that major technology companies are preparing to spend more than $600 billion on AI in 2026. Amazon, meanwhile, warned that in Europe a power-grid connection can take up to seven years, even though a data center itself may take about two years to build. At that point AI is no longer only a software question. It is also an infrastructure question, and very clearly a capital allocation question. This is why responsible AI starts before compliance. Compliance matters, yes, but it is not the first decision.

The companies that will benefit most from the next phase will probably not be the ones that add AI everywhere. They will be the ones that that understand that the real value lies not in AI itself, nor in efficiency for its own sake, but in the business transformation it enables. In this new paradigm, AI is not the end goal, but the catalyst for rethinking how the business creates, delivers, and captures value through better decisions, stronger customer outcomes, new sources of value, and a greater capacity to adapt and compete. And they will know where AI creates that value, where it mostly adds cost, and where a simpler solution still works better. Ultimately, what matters is not whether AI is present, but whether it generates measurable business impact.

>> Undecided about whether AI is best suited for your business? Connect with me to discuss your goals and the best way forward.