AI Strategy
-16 April 2026
-4 min read
By Patrick Rechsteiner, Founder rechsteiner.io
The most common AI cost objection in business is also the most revealing one. The problem isn't the $30. It's the framing.
Many business leaders evaluating AI tools can be forgiven for asking some of the wrong questions. Everything is new, it's moving fast, and the old ways of thinking are as legacy as mainframe code.
When it comes to cost and ROI assessment, the most common objection I hear is that the "extra $30/month" is too expensive and holding back progress. This is most certainly the wrong question - or at least the absolute wrong framing of it.
Jensen Huang of Nvidia made this explicit recently in interviews. He said he would be deeply alarmed if a $500k/year Nvidia engineer spent less than $250k/year on AI tokens. He went on to compare an engineer who avoids AI tools to a chip designer insisting on paper and pencil. The numbers are extreme by design, but the principle is directly applicable to every organisation sitting on the fence about a $30 seat licence.
The math is simple. An employee earning $120k/year equates to $10k/month. You have already decided the role is worth at least that amount. A $30 monthly AI tool is 0.3% of that role's monthly cost - and the opportunity exists, at the click of a few buttons and some investment in training, to increase the output of that role dramatically.
Conservative estimates put productivity gains from consistent AI tool use at 20–30% for non-technical roles. And these are very conservative estimates in my view. Even if you only create a 10% productivity gain, the return is 33x the investment. At 20% it is 66x. At 30% it is 99x. The debate isn't even close.
The reason leaders hesitate is not financial. It is a framing problem. They are evaluating a $30 AI subscription against other $30 software licences, applying cost heuristics built for decisions that have no bearing on this one. The correct comparison is against the cost of the role itself - which is the same frame Huang applies to $500k engineering roles.
The question is not "can we afford this tool?"
The correct question is: "What is it costing us not to use it?"
I'll save you the time. It's a lot.
Frequently Asked Questions
Keep exploring
Work With Patrick
Patrick works directly with leadership teams on AI strategy, digital retail, and commercial growth. If something here resonates, start a conversation.
Get in Touch