Canada's tech-forward companies are running into the same wall as businesses around the world: artificial intelligence is expensive, and the bill is coming due.
For the past couple of years, the corporate playbook was simple — throw AI at everything, see what sticks, and worry about the cost later. But a growing number of firms that bet big on using AI intensively have been stung by sky-high costs, and they're starting to ask a harder question: is all this actually worth it?
The 'tokenmaxxing' hangover
The culprit, in industry slang, is "tokenmaxxing" — squeezing as much work as possible out of AI models, especially for so-called agentic, chain-of-thought tasks where the software reasons through problems step by step. Every one of those steps burns through "tokens," the units AI providers charge for. The more an AI thinks out loud, the more it costs.
That approach can produce impressive results. It can also rack up eye-watering invoices when deployed at scale across an entire organization. Companies that leaned hardest into intensive AI use are now the ones feeling the most financial pain.
From experimentation to 'tokenomics'
The shift now underway is from open-ended experimentation to cold, hard accounting. Businesses are digging into what some are calling "tokenomics" — breaking down the real return on investment for big AI spending rather than treating it as a cost of staying current.
That means scrutinizing which AI tasks genuinely move the needle and which are just expensive novelties. It means asking whether a cheaper, smaller model could do the same job, or whether a process needs AI at all. After a period of near-unlimited budgets driven by fear of falling behind, the pendulum is swinging back toward discipline.
Why it matters here
For Canadian firms, the recalibration is especially relevant. Ottawa's Kanata North tech hub, Toronto's growing AI sector and a wave of homegrown startups have all been folding AI into their products and operations. With venture funding tighter than it was during the boom years, Canadian companies have less room to absorb runaway cloud and AI bills, making ROI questions urgent rather than academic.
The federal government has also poured money into AI research and infrastructure, positioning Canada as a leader in the field. But the lesson emerging from the corporate world is that the technology's promise has to be weighed against its very real operating costs — a calculation that applies to public-sector AI projects just as much as private ones.
The bottom line
None of this signals that companies are abandoning AI. Instead, the era of spending without measuring appears to be ending. The businesses that thrive in the next phase will likely be the ones that figure out where AI delivers genuine value — and where the tokens simply aren't worth it.
Source: CBC News (Top Stories), cbc.ca.
