Silicon Valley’s AI bill is landing in Australian small business inboxes. Here’s how to fight back.

The numbers in this morning’s SMH story are worth sitting with for a moment. Sydney software company Dovetail is spending $43,000 a month on AI coding tools, six times their January bill, with no extra headcount to show for it. Tommy Gregory, a 20-year-old running a marketing agency in Sydney, pays $550 a month across his AI stack and is blunt about it: “I’m pretty much hostage to them.”

That’s the big-business version of the problem. But the more common version looks different.

For most Australian small business owners, the AI cost creep isn’t a single $43K line item. It’s six subscriptions at $40–80 each, two of which you barely opened last month, one that auto-renewed after a trial you forgot about, and another that got quietly upgraded to a “Pro” tier your contractor signed up for. The Weel Australian AI Spending Index found 30.8% of Australian SMBs are now using AI (up from 22.1% in January), with the average Anthropic customer spending $1,082 a month. That’s 3.6x what the average OpenAI customer spends.

The costs are real. The question is whether they’re working.

Three things to do this week

First: audit your stack. List every AI-related subscription your business is paying for. Not just the obvious ones (ChatGPT Plus, Canva Pro, Xero AI) but the ones buried in expense reports, the one your social media person signed up for in March, the Grammarly renewal nobody questioned. Most business owners I’ve spoken to find they’re paying for two or three tools doing the same thing.

Second: set a monthly cap per tool before you need one. The me&u CEO put it well: “What we’re seeing right now is that AI cost is moving week to week, not year to year.” API costs especially can compound fast. If you or your team are building on top of AI APIs (even small integrations), you want a hard ceiling on monthly spend, not a line item that surprises you mid-quarter.

Third: apply the token test before you add anything new. Build Club founder Annie Liao said this week that the companies that win will be the ones that know which tokens are worth burning, not the ones burning the most. Before any new AI subscription, ask: what specific output do I need, how will I measure it, and what happens if I cancel?

The dependency question

The Fertility2Family story from the same SMH piece is the one worth sitting with. The Hobart company is spending $100 a month on OpenAI for work that previously cost $30,000 through an agency. Enormous savings. But their founder’s observation cuts through the feel-good math: “AI did not remove our dependence. It moved it from agencies we hired project by project to an overseas platform we rent every month.”

That’s not a reason to stop using AI. It’s a reason to know the cost clearly, have a contingency if pricing spikes, and not build your entire operation on a single platform you don’t control.

The businesses getting burned right now are the ones that treated AI spend like it was exempt from the same scrutiny they’d apply to any other operating cost. When spend goes uncapped, it compounds fast, and fast mistakes are expensive to reverse.

Your AI stack isn’t special. Audit it, cap it, make it earn its keep.