By Andrew Miller
AI is quickly becoming the most expensive employee on many payrolls.
Every Monday in The Signet Letter, we flag something that could affect your business this week and what you can do about it. Now, back to the story…
The very same technology that was supposed to cut costs has now increased them in many cases. Many businesses are facing overblown AI budgets, including some Fortune 500 companies. For example, Uber already burned through its yearly AI coding budget in just four months.
The problem is skyrocketing token costs to use AI models.
Every new updated model from leading AI developers costs roughly twice as much as the previous token. It’s too unsustainable even for titans like Microsoft.
Microsoft is one of many employers reassessing their AI needs. They’re getting rid of tokens for employees in non-technical roles who don’t need access to advanced models. Still, power users like the company’s top engineers are racking up token costs. Rather than cut back on their AI use, Microsoft is switching engineers to cheaper, in-house models instead.
But Microsoft isn’t alone. Nearly 95% of enterprise AI usage still runs on the most expensive frontier models for simple tasks that could be done with much less power.
If your business is already paying for AI tools, it’s worth asking two questions: Who actually has access, and what are they using it for? If the answer to the second question is “basic tasks,” you’re probably paying top dollar when a cheaper tool would get the job done.
Businesses taking full advantage of AI aren’t spending less on it, they’re spending it in the right places.
P.S. The same logic applies to business financing. The most expensive option isn’t always the best fit, and the best fit is rarely the first one you find. If you’re looking for funding that matches your actual needs, let’s talk.