Every conversation about Chinese AI companies going global is a conversation about the United States. Which is exactly why almost nobody noticed what’s happening in Australia.
Chris Liu arrived there in 2024, after a career that reads like a tour of Chinese internet history and its casualties. Product manager at Weibo and Alipay. Then three startups: a restaurant SaaS company, a garbage disposal venture, a livestream commerce supply chain business. His own bio says serial entrepreneur, never successful.
He’s unusually clear-eyed about all three.
“The first one I did looking at opportunity — where’s the demand, where’s the vacuum, where’s the money. The second I did on passion.” That was the garbage disposal company, an environmental play nobody around him understood, including his investors. “I bought into Musk’s line — he says even today he doesn’t think EVs were a good opportunity, but he wasn’t willing to sit around waiting for someone else to do it. Smart people wouldn’t touch it, so I did. Didn’t work out either.”
The third he did looking at what accumulates. “Once you have a family you think more about what this thing builds. That’s my criterion now.”
His mother’s assessment: You didn’t make much money, but you spent a lot — so you contributed to the country.
Undercover in a Furniture Store
Chris spent his first six months in Australia mostly not working. Then he started looking for something to do.
His first instinct was furniture. He’d noticed that assembling a few beds in Australia costs around 2,000 AUD — close to 10,000 RMB. Having spent years in supply chain and e-commerce in China, his reaction was: this is easy money.
An older friend heard about the plan and sent word through a mutual contact: Tell Chris — for six months, don’t do anything. Especially anything that involves spending money.
He didn’t spend the money. Instead he went to work in a furniture store for two weeks to see the business from the inside, and found it more competitive, more logistics-heavy, and slower-moving than it looked from outside.
Then a friend brought him along to paint walls.
The economics were startling. A laborer makes 300 AUD a day — already above most white-collar salaries. A skilled tradesman, 700 to 800. Installing a set of window screens ran him 4,000-plus dollars for three hours of work; net of materials, that’s 1,000 an hour.
“I painted for a few days and it became clear. I’m not here to do something where I can see the whole road from where I’m standing.”
He put down the roller and picked up the laptop.
Closing Deals Is Easier Here
Through a chain of coincidences Chris ended up working with UMART — a 26-year-old Australian consumer electronics e-commerce platform, number one in its category.
What surprised him was the speed.
“In China, if you want to charge a company even a modest retainer for operations work, it’s a whole project. You send a proposal, you meet several times, you negotiate budget. In Australia it’s basically one or two conversations to know whether this is happening.”
The tradeoff runs the other way on budget. Australian companies are cautious with initial spend — nothing like the millions Chinese brands throw at campaigns. But once trust is established and there’s a defensible measurement framework, the money becomes hard to spend fast enough.
“You do a halfway decent job and you’re already in the top few in the industry.”
The Numbers Nobody Quotes
Australia’s GDP per capita is roughly 65,000 USD — second only to the United States in the entire Western world. Ahead of the UK, France, Japan, and South Korea.
Twenty-seven million people, with Sydney, Melbourne and Brisbane holding about half of them. Cover three cities and you’ve effectively covered one unified market — a very different proposition from the fragmentation of Southeast Asia or Europe.
It also has the world’s highest minimum wage, currently 26.4 AUD an hour and rising, and among the most expensive logistics and warehousing anywhere. Four banks. Two supermarket chains with over a thousand stores each. Heavy concentration at the top, and 97 percent of businesses with fewer than 20 employees.
Australia’s 2025 national AI strategy also declined to legislate — no restrictive framework, no exclusion of foreign systems. At a moment when regulatory posture is becoming a primary variable in market entry, that’s a real advantage.
Already on the Shelves
This isn’t a forecast. Chinese AI hardware is in Australian retail right now — not on Amazon, but in Harvey Norman and JB HiFi, the mainstream chains. One AI fitness product is on track to do roughly four million USD in Australian sales this year.
Which brings Chris to the reason he thinks Australia matters strategically, beyond its own size.
“Australian consumer habits, market rules, and culture are very close to the US. But the competition is a fraction of it, and the cost of getting things wrong is far lower. If you can clear compliance here, you’ve properly entered a Western market. And if you can’t make it work in Australia, you definitely can’t make it work in the US.”
The American market is competitive enough that everyone in it dismisses Australia as too small to bother with — which is precisely the opening.
Why You Can’t Sell AI Tools There
Here’s the part that inverts the standard playbook.
When 97 percent of your addressable market has under 20 employees, selling an efficiency tool doesn’t work.
“You can’t reflexively sell AI as a tool. A tool still needs a person driving it — all you’ve done is improve efficiency. But the margin in this market is what it is. Improve efficiency and you’ve just created surplus efficiency.”
There’s no headcount to redeploy on a five-person team. The gain doesn’t convert to money.
What works is replacing the function outright.
“If you’re doing AI accounting, go compete with accounting firms as an accounting firm — lower price, better service, more accurate reports, real-time if you want. Don’t explain how AI helps them do things. Tell them what you’ll do for them.”
The Nine-Month Lesson
Chris’s clearest argument for Australia comes from a number he found while running his restaurant SaaS company in China: the average lifespan of a Chinese restaurant business is under nine months.
“You haven’t even earned back the value-added services before they’re gone. Your sales cost is still sitting there, the amortization isn’t finished, the upsell never happened.”
Australia is the inverse. Less brutal competition, continuous population inflow, far longer business lifespans. Customer lifetime value is long enough that a high upfront acquisition cost amortizes comfortably.
And there’s a second structural difference he considers underrated. Chinese companies, once they reach a certain size, build internal IT teams — outside vendors start looking like a bad deal. Australian companies, in his words, stay put. If an external provider can do it, they don’t hire.
“For a B2B AI company that matters enormously. Your customer doesn’t grow up and replace you.”
Where the Two Kinds of People Meet
Chris’s closing observation is one I’ve heard versions of from nearly every guest, but he puts it most concretely.
There are two kinds of people in the AI era. One comes out of internet companies — instinct for the tools, understanding of what a workflow is underneath. The other has spent years inside a specific industry and knows exactly which steps hurt and where the change needs to happen.
“These two kinds of people need to meet at the summit.”
He practiced this literally at his restaurant SaaS company: product managers sent to wait tables, designers put on the cash register, engineers watching from the side.
“Clicking through the system during a rush is not the same as clicking through it on a quiet afternoon. Do one shift on the register at peak and you come back knowing exactly which interactions are wrong.”
Which is also, in retrospect, what the furniture store and the paint roller were.
“Doing the Australian market — or any small market — comes down to whether you find the right partner, plus strategic patience, plus respecting how long things take.”
This article is adapted from 离线时间 EP17.