Everyone Is Paying for the AI Boom. Few Are Collecting.

AI & TechPoint of View
Written By Adam Taufiq

The Federal Reserve's latest meeting notes, released Oct. 7, did something new. They named the AI buildout, alongside oil and tariffs, as a reason inflation is still above target. Several officials said the effect of tariffs on prices is fading while the effect of AI spending is growing, and most expect another rate increase this year. In the same week, Samsung estimated a quarterly operating profit of about $80 billion, which would be the largest any technology company has reported, and the 10-year Treasury yield touched its highest level since 2002.

A boom growing this fast squeezes everyone else. Chips, servers and data centres take enormous amounts of capital. JPMorgan's CEO expects the largest cloud companies to spend about $700 billion this year and close to $1 trillion next year, and much of that is borrowed. It comes out of the same pool of capital every other borrower draws on, so the price of borrowing rises for all of them. The same thing is happening with physical goods. Chipmakers are shifting production toward AI, memory chip prices rose close to 20% last quarter, and PC makers have confirmed laptop price increases of 15-20%.

So a mid-market firm or a public agency is already paying for AI through its financing costs and its hardware budget, whether or not AI has changed anything about how it works. For most, it hasn't. The Bank of Canada found that only 8% of Canadian businesses use AI significantly in their core operations. Statistics Canada puts the share using it at all to produce goods or deliver services at about one in five.

What holds the rest back is some of everything: the technology, the strategy, the ability to change how work gets done, and the capital to fund it. In one survey of 300 U.S. and Canadian executives, 78% said they already know what their organization needs to do to scale AI. Only 19% had a company-wide plan to do it. The barriers they named most often were skilled talent and budget. Access to the tools didn't make the list.

The ground also keeps moving. OpenAI rolled GPT-6 out across ChatGPT this week, a month after the first version reached paying customers. New models and new ways of working arrive faster than most organizations can evaluate them, and staying current is real work that someone has to own.

For a COO setting next year's budget, four moves:

  1. Budget for AI in every part of the organization, under one plan. AI isn't an IT line item. Each function needs funding for it, tied to a single plan so that what works in one area gets scaled to the others.

  2. Hold a reserve. Tools and prices are changing too quickly to commit every dollar in January. Keep an unallocated amount so you can pivot when a better option appears or a cost jumps.

  3. Check that your technology can carry it. Before buying more, evaluate whether your current systems, data and hardware can support what you're planning. In a survey of 1,330 board members and executives, integration with existing systems ranked among the top priorities for getting value from AI. With hardware prices rising, decide on purpose what gets replaced and what gets stretched.

  4. Fund training as seriously as the tools. In Workday's latest workforce research, 79% of employees said they know which skills they need, and only 66% said their employer helps them build them. The efficiency you're budgeting for shows up only when people know how to use what you bought.

The squeeze may ease. Fed staff expect AI's effect on prices to fade over the next two years, and if the productivity gains arrive, the investment pays for itself. Next year's budget can't wait to find out.

The costs of the AI boom arrive on their own. The returns have to be built.

Working out where AI belongs in next year's budget? We're glad to talk it through: eightygroupllc.com/contact

Sources

Adam Taufiq
Next

The Job Market Froze. Your People Noticed.