Summary

Airwallex’s founder-market fit was forged through financial insecurity rather than fintech pedigree. After losing family support at 16 in Australia, Jack Zhang funded $24,000-a-year tuition by carrying lemon boxes in 40-degree heat, washing dishes, bartending, and working overnight petrol-station shifts—sometimes 16 hours straight. His operating baseline became simple: “You don’t have a choice,” and he says early hardship made him resilient.
The company emerged only after Zhang proved that financial success without technological scale did not satisfy him. Alongside a roughly $200,000 job as a developer and algorithmic trader, he built more than 10 businesses, generated $2 million-$5 million annually from side ventures, and had made over $10 million by 28 or 29. He kept coding despite earning perhaps 10 times more elsewhere because “I need to write code every day”—the side businesses supplied security, but Airwallex supplied the mission.
A coffee-shop payment trapped for two months exposed the infrastructure problem Airwallex would attack. A transfer under co-founder Max Li’s name hit an OFAC false positive, while SWIFT’s intermediary chain, fees, and 140-character messages made small-country payments slower and more expensive. Zhang’s first answer—netting opposing flows point to point—failed because it required billions in balanced volume, forcing Airwallex to build licenses, clearing connections, and institutional FX access the hard way.
The financing history shows how violently startup outcomes can diverge from conventional underwriting. Lucy, a first-time investor Zhang met over dinner, wired $1 million into his personal account before incorporation for 20%; on the episode’s diluted $9 billion valuation math, that became roughly $1 billion. Matrix Partners signed and then withdrew a $2 million term sheet at a $10 million post-money valuation, while Australian funds that initially refused meetings later invested at $6.2 billion.
Airwallex survived repeated product failure before two customer groups unlocked explosive volume. Peer-to-peer matching failed, Australian SMB invoicing never overcame acquisition costs, and Tencent and Mastercard delivered far less business than expected after anchoring Series A. Square Peg’s $6 million extension bought time; tuition-payment companies and SHEIN then carried Airwallex from zero to $1 billion in transaction volume in nine months, including a 40-hour weekend integration when a partner threatened to cut off SHEIN, which was kind of 90% of the business.
Rejecting Stripe’s roughly $1.2 billion offer transformed Airwallex from a payments rail into a lifetime company. The proposed package included about $800 million for the cap table, $350 million for Zhang and his co-founders, and roughly $25 million-$50 million for core employees; he was “70%” inclined to accept but feared finishing a five-year lockup near 40 and starting over. After roughly 90% of senior leaders voted to continue, Patrick Collison’s pledge to build Stripe for “the next 20, 30, 40 years” became Zhang’s model for committing his own life to Airwallex.
Post-offer capital abundance created Airwallex’s largest strategic bet and its clearest operating mistake. Zhang raised another $100 million, expanded headcount from roughly 100 to 600-700 without a budget, and put most of the capital into cards, acquiring, and banking products that would generate no revenue for three to four years. The platform bet eventually produced revenue; opening international offices before local product-market fit did not, with the UK taking three years and exposing weak commercial hiring.
The current numbers make durability—not raw growth—the central investor debate. Zhang says Airwallex grew at least 100% annually from 2015 through 2023, remained near 90% year-on-year, and moved from roughly $500 million ARR in August last year to $600 million in November and $700 million in January or February. Harry projected more than $1 billion by year-end, but Zhang did not affirm that forecast. In the latest $300 million round at a $6.2 billion valuation, Zhang said investors emphasized roughly $450 million of gross profit this year and, after companies decelerated sharply following 2021, “give you less credit” for sustaining hypergrowth. He and co-founders are pursuing $70 million of debt to buy secondary shares.

Deep dive

1. Financial insecurity established Zhang’s operating baseline

Zhang arrived in Australia around 15 and lost family financial support at 16. Returning to China’s university system no longer seemed viable, while international tuition cost roughly $24,000 annually, so survival and education depended on “working as many jobs as possible” in a country where he had no safety net.
One summer meant several hours of trains and buses to a mountain lemon factory, carrying thousands of boxes for 12 hours in 40-degree heat without lunch. The $14-an-hour Australian wage was respectable; the enduring thought was, “One day, I don’t want to do this anymore. I want to write code.”
At university he washed dishes, bartended from 4:00 to 11:30 p.m., then worked a petrol station from midnight to 8:00 a.m., four or five days weekly. Harry’s distinction mattered: both men worked extreme hours, but Zhang’s early work was physically punishing and unrelated to anything he enjoyed.
Zhang traces resilience to both hardship and an early taste of success. At 13 or 14, he helped turn unwanted merchant flyers into a student magazine carrying love stories and Counter-Strike advice; the limited print run made it popular, and he said roughly 8,000 merchants advertised over 18 months before proceeds were donated to the school.

2. Ten profitable businesses proved money was not the mission

After graduating, Zhang worked as a developer and algorithmic trader at Aviva and investment banks, earning around $200,000 annually. Nights and weekends went into exporting Australian wine and olive oil, importing Chinese textiles, reselling phone cases, and building architecture, project-management, property-development, and hospitality ventures.
He characterized the trading model as simpler than outsiders imagine: find a supplier and buyer, avoid manufacturing and consumer marketing, and broker B2B demand. The portfolio eventually produced roughly $2 million-$5 million a year, while his real-estate projects reached 40-50 apartments in an A$40 million-A$50 million project.
By 28 or 29, Zhang estimated he had made more than $10 million and achieved financial security. Harry challenged the lack of focus—why not make property 10 times larger? Zhang’s answer was that he had started more than 10 businesses and “concluded I didn’t like any of them”; profitable was not synonymous with meaningful.
The tell was that he never stopped his lower-paying engineering work: “I need to write code every day.” Watching Facebook, Google, Tencent, WeChat, Alipay, and Alibaba emerge convinced him technology could amplify impact at a scale unavailable to his cash-generative side ventures.

3. A blocked coffee payment exposed SWIFT’s structural tax

Zhang and Max initially intended their coffee business to become a chain, inspired by the prevalence of retail businesses in Australia’s fast-growth rankings. Finding no Square-like POS stack in 2013 produced one possible startup, but CTO Jacob favored QR while Zhang favored NFC; unable to agree, they abandoned it.
The decisive problem surfaced while importing beans and packaging. Because Max Li shared a name with someone on the OFAC blacklist, a personal transfer was blocked within the correspondent-banking network and returned only after two months—prompting Zhang to investigate why an ordinary supplier payment could disappear.
His reconstruction of SWIFT ran from a local bank through institutions such as Barclays, Citi, and Itaú before reaching a smaller Brazilian bank. More intermediaries meant more fees, delay, and compliance checks, while a 140-character message left too little context to eliminate false positives: “If data on the internet is real time, why isn’t money?”
Airwallex first proposed matching opposite Australia-Brazil flows and netting them locally. Building the algorithm revealed the fatal dependency: it needed “billions and billions” in balanced volume when Zhang did not know how to acquire even $100 million, so the founding concept failed before launch.

4. A first-time angel funded a company that did not yet exist

On a Friday at the coffee shop, Zhang met Lucy, a friend of Max who had left investment banking and initially wanted to discuss investing in coffee. Over dinner, he redirected the conversation toward rebuilding cross-border payments; within an hour she asked, “What if I give you two million?”
Lucy proposed $2 million for 40%. During a three-hour negotiation at the University of Melbourne law school the next morning, they settled on $1 million at a $5 million post-money valuation, or 20%, despite her husband arguing that the money was intended for a home and family.
By Monday, before Zhang had resigned, incorporated a company, or signed documents, CBA notified him that US$1 million had arrived in his personal account. He resigned the same day. He estimated that, after reasonable dilution and at a $9 billion valuation, Lucy’s first-ever investment was worth around $1 billion.
Zhang then pressured an exceptional former engineering colleague with three children to leave his bank job: “We’re not going to leave until you resign.” The founding team worked roughly 20-hour days and slept in sleeping bags inside a 10-square-meter office—an immediate conversion of financing into speed rather than comfort.

5. Rejected founders became expensive lessons for their investors

Matrix Partners signed a $2 million term sheet at a $10 million post-money valuation, then withdrew it after concluding the matching algorithm would not work and was not defensible. Zhang had already rejected alternatives; Gobi and other smaller investors eventually assembled the additional $2 million needed to proceed.
Another early investor waited roughly five months after signing before wiring, repeatedly demanding a demo—what Zhang calls Airwallex’s first near-death. Australia had only a few meaningful venture firms, and two would not initially meet him; Blackbird and AirTree later joined the cap table at a $6.2 billion valuation. Zhang said that investment was the largest check in Airwallex’s history.
Harry’s pushback was that investors asked the wrong questions: Zhang’s survival story said more about a founder’s ability to build for decades than quarterly pricing or product. DST’s Yuri Milner later focused almost entirely on Zhang’s upbringing, then agreed within roughly two hours to invest $100 million at a $1.1 billion valuation after the Stripe rejection.
Zhang’s fundraising rule is not to maximize price, but leverage across brand, recruiting, and commercial access. He believes only perhaps five venture brands materially help and says Sequoia’s name made “a huge difference,” though he still wishes he had found a Michael Moritz-like champion publicly underwriting the trillion-dollar vision from day one.

6. Failed matching gave way to expensive, defensible infrastructure

With the first $3 million disappearing and peer-to-peer matching unworkable, Airwallex chose the slow route: institutional liquidity, country-by-country licenses, and direct local-clearing connections. At 8:00 one morning, Zhang cold-called Tom, a junior Macquarie employee finishing an overnight FX shift, and persuaded him to mobilize an engineering team.
Macquarie ultimately let Airwallex execute transactions as small as $20 at roughly two basis points, versus normal interbank minimums of $500,000-$1 million. For its first years, Airwallex streamed prices through its own FX engine while trading back to back with Macquarie; Tom later became Macquarie’s head of distribution.
Coverage initially came from aggregators and future competitors including Currencycloud. Airwallex built a payment-link and invoicing product for Australian SMBs, but acquisition costs were too high and the company reached only about 100 customers; Zhang knew the product lacked product-market fit even though the underlying rails worked.
His response to commoditization was depth, not denial. Airwallex pursued licenses in Australia, Hong Kong, the UK, and the US, connected fragmented infrastructure, and was, in Zhang’s telling, essentially the only provider offering this capability across Asia-Pacific—turning regulatory reach and unit economics into the prospective moat.

7. A broken demo rescued Series A, but strategic customers disappointed

Zhang sought funding at a $60 million valuation with no revenue, pitching an alternative to SWIFT rather than pretending the SMB product worked. Sequoia said it would invest if Tencent co-led; Tencent completed its investment-committee work but passed its review committee. The proposed API pivot would have used Airwallex to power WeChat Pay’s global settlement.
After nearly three months, Zhang secured a January 5, 2017 meeting with Tencent strategy chief James Mitchell, calling it “the most important meeting of my life.” Mitchell’s prior work on PayPal’s IPO meant he already understood cross-border payments’ economics, but when Zhang demonstrated the invoice product, the payment button returned a 404.
Zhang improvised that Tencent’s firewall might have blocked the link—though the underlying payment provider could also have failed. Mitchell replied that it did not matter, convinced Pony Ma to approve the investment, and unlocked a roughly $13 million Series A from Tencent, Sequoia, Mastercard, and others.
Harry challenged Sequoia’s conditional conviction; Zhang defended it as rational for a Series A with zero revenue because Tencent and Mastercard were supposed to become anchor customers. Neither delivered as expected: Tencent took three years, built much internally, and used Airwallex only as a liquidity vendor; Mastercard promised $1 billion of volume but supplied under $1 million of high-risk transactions Airwallex eventually offboarded.

8. Two customers converted three years of failure into product-market fit

By late 2017, Airwallex had spent another year building an enterprise API without winning customers and was again close to running out of cash. Square Peg’s founder—an Australian entrepreneurial idol for Zhang—backed the mission and led a $6 million Series A extension that created the runway for another iteration.
In January 2018, Airwallex onboarded large tuition-payment businesses and SHEIN, then rode their international growth from zero to $1 billion of transaction volume within nine months. Product-market fit arrived not through a broad SMB base, but through roughly 100 comparatively large customers whose cross-border needs were already scaling.
When SHEIN sent $20 million for Chinese suppliers, Airwallex’s local partner recognized that its own customer had moved volume to Airwallex and threatened to shut the rails off. With SHEIN representing kind of 90% of Airwallex, Zhang told a replacement partner, “Let’s figure out the legal contract later”; engineers worked 40 hours from Friday night and launched Sunday.
The inflection rapidly changed financing. Sequoia and Tencent led an $80 million round at a $400 million pre-money, $480 million post-money valuation despite minimal revenue, because Airwallex had found large customers with rapidly scaling transaction needs.

9. Stripe’s offer clarified that Airwallex was a lifetime project

Stripe approached in October 2018, when Airwallex’s real product-market fit had existed for roughly 10 months. Patrick Collison flew to Shanghai for a full day: Airwallex had payout and FX rails and wanted acquiring, while Stripe had global pay-ins but needed payouts, making partnership, competition, or acquisition the logical choices.
Their shared ambition was an “AWS of financial services.” Zhang was struck less by a collaborative Google document than Collison’s intellectual honesty and ability to go deeply into quantum physics, biology, and Chinese history: “I’m Chinese. I feel ashamed that he knows more Chinese history than me.”
The proposed deal placed roughly $800 million on the cap table, approximately $350 million with Zhang and his co-founders, and another roughly $25 million-$50 million with core employees—close to $1.2 billion altogether. Stripe sent senior staff to Melbourne for diligence, and Zhang became about 70% convinced he should accept.
Back in Melbourne, the five-year lockup became decisive: Zhang would be close to 40 before starting again, and money held little appeal in a city where he believed $200,000 could fund a good life. Roughly 90% of leadership voted to keep building, while Collison’s commitment to Stripe for “20, 30, 40 years” inspired Zhang to make the same choice. Zhang said he probably would have made at least $3 billion if Stripe reached a $100 billion valuation.

10. Capital abundance funded the platform—and exposed operating weakness

DST and other investors subsequently supplied another $100 million, removing constraints Zhang did not yet know how to manage. “I don’t know what financial discipline is,” he admitted: without a budget, Airwallex went from roughly 100 employees to 600-700 in a year, while volume grew faster than revenue and cash again ran low.
Zhang put most of the $100 million into corporate-card issuance, merchant acquiring, and banking infrastructure that would produce no revenue for three or four years. He viewed that as existential after rejecting Stripe: Airwallex had to evolve from money movement into an end-to-end global banking platform before it lost its role in the market.
International expansion was mistimed. Airwallex opened UK, US, and other offices before possessing local product-market fit or a repeatable commercial organization; Zhang lacked his Asia-Pacific venture network and did not know how to sell enterprises locally. He kept trying rather than cutting immediately, and the UK took three years to work.
Culture compounded the delay. Experienced bank and SWIFT hires repeatedly told the founders their ideas could not work, yet Zhang says none of those hires worked out; Airwallex eventually fired them, endured negative press, and shifted toward curiosity, determination, resilience, belief in the vision, and passion. He would still hire the first 100 carefully—but use an early recruiter instead of sending the messages himself.

11. A pandemic close converted another near-death into eight years of doubling

SoftBank prepared to lead a 2019 round, but the WeWork crisis halted investing and the responsible partner was fired. DST and Tencent supplied a roughly $70 million-$100 million convertible lifeline; Hedosophia then led a roughly $150 million round at a $1.7 billion valuation, with Ian Osborne contributing around $75 million-$80 million.
Closing coincided with COVID-19 and a roughly 30% market decline. Airwallex knew the danger from employees in locked-down China, and tuition and travel represented about half its revenue. Neil Shen told Zhang nobody could know whether the round would close. Osborne funded it and subsequently doubled down across five or six rounds.
Zhang gave the revenue chronology as roughly $2 million in 2018, $10 million in 2019, and $20 million in 2020, followed by approximately 2.5 times growth in 2021. The cards, acquiring, and banking investments only began contributing revenue in 2022.
Capital flooded in during 2021: Greenoaks invested at $2.6 billion, Lone Pine at $4 billion, valuation reached roughly $5.5 billion, and Airwallex raised about $400 million. That cushion absorbed close to $200 million of annual burn and enabled a flat 2022 round. Airwallex continued growing without layoffs, but Zhang said it was not profitable in 2023.

12. At $700 million ARR, durability matters more than possibility

Zhang says Airwallex never grew below 100% annually from 2015 through 2023 and still expanded about 90% year-on-year. ARR advanced from roughly $500 million in August last year to $600 million in November and $700 million in January or February.
Harry projected more than $1 billion by year-end. Zhang did not affirm that forecast; he answered that Airwallex had raised $300 million at a $6.2 billion valuation. He said investors now emphasize gross profit—about $450 million this year, implying roughly 13 times—and use public-market comparables. After companies that grew 100% in 2021 later fell to 20%, “people give you less credit” for sustaining hypergrowth.
His own signal is unusually direct but not yet completed: Zhang and co-founders are pursuing approximately $70 million of debt to buy Airwallex secondary shares. He advises late-stage founders above $1 billion to take enough liquidity for housing and family—perhaps $20 million-$30 million in London—while Harry’s caveat was equally firm: timing matters, and early founders should not extract that sum.
Zhang dislikes the everyday CEO work of personnel issues, policies, and procedures, but wants control over the company’s destiny and accepts decisions with “more than 50% probability” of being wrong. By 2035, he wants Airwallex larger than Citi or HSBC, serving millions of businesses. He said going public is something the company will inevitably have to consider, but for now it is secondary to building the infrastructure.

7 More Healthy Years: What We Can Learn from Super Agers Graham Weaver - Building Alpine - [Invest Like the Best, EP.425]