Airwallex × Outsome × Bloom: Going Global from Korea

We sat down with Yuna Kwon of Airwallex and Peter Shin of Outsome to talk about how Korean startups and SMEs expand overseas.

What stuck with me most was Yuna's story. When she worked at Shopee, overseas sales by Korean sellers grew 40x in just five years. The reason: an understanding of the logistics and payment setup needed to sell abroad spread across the market, and a lot of sellers started going into foreign markets directly.

In Japan, companies usually go abroad indirectly, through big conglomerates, trading houses or overseas distributors. In Korea, most go direct. Global financial infrastructure like Airwallex is aimed mainly at SMEs, and in Japan only about 5% of SMEs sell abroad directly. In Korea it's 20%. So compared with other countries, far more Korean companies want to go out on their own. To grow revenue in several countries you have to open accounts and set up entities country by country, and I know how complicated and expensive that is. It made me appreciate again how much time and money this kind of global payments and finance infrastructure saves.

Peter is close to being a mentor to just about everyone in the Korean startup scene. What he posts isn't theory copied from somewhere. It comes from things he ran into himself, so it's practical and detailed, with a lot you can use right away. His main message was to hack the system. Two things stood out. Even moves that feel close to the line between legal and illegal in Korea count as traction in the YC way of thinking. And YC itself is an outlier in the US investment world, so it keeps looking for founders in Korea who bring a vision that won't break.

After the sessions, the roundtables ran about 40 more minutes with a lively discussion. At the end, each table's leader came up and shared what their group had worked out, and every table came at it from a completely different angle. It was fun to listen to the whole time. Thank you to Gain Choi of Olive Young (Korea's biggest health and beauty retail chain) for the great space, and to everyone in the Bloom community who stayed until the end.

181 people signed up, and 51% of them were CEOs, C-level executives or partners. 36% said they were already making revenue overseas, and when asked to pick a target market, 56% put the US and North America first. When we asked them to choose the single biggest blocker in overseas payments and transfers, regulation, paperwork and tax handling came first with 30 votes, followed by opening overseas accounts and entities at 26 and fees and FX losses at 23.

If you made it to Series A in Korea, you're already at Series D in the US

Peter Shin co-founded QueryPie and went through YC in the W20 batch. He then ran the Korea program at 500 Global for four years before starting his own accelerator, Outsome, last July. 39 companies have gone through the program so far, at every stage, from college students with just an idea to a CEO doing 20 billion won a year in revenue who is starting over quietly in the US.

Earlier this year I had a video coffee chat with him and got chewed out pretty good. His point was that I had ambitions to go overseas but wasn't specific about what I was selling. The US is huge, he said, so you have to pick West Coast or East Coast, and even the city. I brought back one line from that call and asked him what he meant by "sell the purpose, not the means."

Say you make pajamas. If there's an IoT device in those pajamas that collects biometric data, it becomes a completely different product. When we buy an iPhone Pro or Max, we're not buying a lump of CPU in a titanium chassis.

The first wall Korean founders have to get over, he said, is self-censorship, not technology. If you got to Series A in Korea, you've already cleared a difficulty level close to Series D or an IPO by US standards. Even so, founders stay stuck on the idea that the product has to be perfect, and they don't go out. He compared it to Korean archery, where the national championships are harder than the Olympics. Taking that level of difficulty straight to Silicon Valley is the arbitrage we get to enjoy.

The problem is when founders cross over without letting go of that obsession. They burn government grants plus their own money to get there, and the night before meeting local VCs and companies they're doing code review and QA. So he gives teams heading to Silicon Valley just one metric.

Go to Stanford or somewhere like that and make one friend. Dye your hair pink or yellow, whatever, so they remember at least one thing about you.

He didn't see government grants as bad in themselves. His own team went to TechCrunch on government money, and that's where they met YC. But he added that you may not need that money to bring out your instincts as a businessperson. In the US, even people coming out of FAANG need a paycheck right away, so founders start with a different intensity. Living off unemployment benefits and a pre-startup grant (a Korean government program for early founders) for six months isn't even an option.

Where outliers write the textbook

The name Outsome comes from "Outliers build something." Peter said the people he looks for, and the people who come looking for him, are all outliers in the end. His evidence was his own experience. Before getting into YC, he went to every investor in Korea and got turned down by all of them. Only after YC backed him did sales open up with KakaoBank and KakaoPay, and big-name VCs lined up. He named his own background as what made him an outlier: he grew up in Singapore and didn't come from the Ivy League or Samsung.

He went on to say YC itself was an outlier. When YC set up its Continuity Fund to make follow-on investments too, Sequoia and other Silicon Valley VCs openly opposed it, because if one firm took both the early and the follow-on rounds there'd be no room left for them. On top of that, the person who built YC came from the ranks of exited founders, not from the traditional LP world of university endowments, and the funding came from outside Silicon Valley.

Silicon Valley is a place built by outliers, where outliers write the textbook. So if you're an outcast in Korea, Silicon Valley is your real home.

I asked how you spot an outlier, and he gave two things. One is being able to handle isolation. To last three or four months in Silicon Valley without even a bowl of doenjang jjigae (Korean soybean paste stew), you have to be seriously tough. The other is a habit of executing first. Silicon Valley is a society that collects frameworks, and the good outliers are the ones who put a framework into action the moment they get it.

I answered with the bottle cap story. Korean companies have plenty of great ideas, but if you keep hearing "does that even make sense?" you end up jumping with the lid on, and even after the lid is gone you only jump that high. I took him to mean that an outlier is the person who keeps going up once the lid comes off.

The best survey is a payment

He said several times that instead of sending out surveys and watching reactions, you should watch whether money comes in. I asked what a company without a product yet should do. His answer: in this year's YC batch, one in six companies is a services company. In Korean terms that's consulting or SI (systems integration). Now that coding has gotten easy, someone with industry insight interviewing customers and signing contracts is itself proof of product strength, he explained.

This is a zone that goes back and forth across the line between legal and illegal. By Korea's self-censorship standards, selling a product that doesn't exist yet could be called fraud.

When I asked how to sell once you've built it, the first answer was that you need a thick skin. Before the channel, cold email or LinkedIn, what matters is whether you've built the muscle to get your insight across in front of other people. Nobody is born with that muscle. You build it by writing on LinkedIn, giving talks, and hosting events yourself starting from nothing.

I also asked how far to narrow the target. He said don't stop at the company level; go all the way down to the job title. Your customer isn't US healthcare. It's the hospital selling a new diabetes drug, or the insurer setting the reimbursement rate for that drug, and then you go one step further and write down which title at that company is your customer.

He was firm that in B2B you have to show up in person. Skip the famous events like CES and go to places like an architecture conference in New York or a conference just for dentists, where you can see that America is a nation of nerds. AI has lowered the barrier to entry, so while others make calls with agents and meet twice on Zoom, the side that shows up in person one more time has the edge. His evidence was Minsoo Seo, CEO of Taewoong Medical, which sold to Olympus for about 500 billion won, and whom Peter considers a mentor. Their highest-revenue hospital was the one closest to their sales rep's house.

You don't have to lead with "I'm from Korea"

I asked whether being a foreigner would be a trust problem, and he flipped the premise. Korea is homogeneous, so someone from Vietnam stands out right away. In the US, nobody knows whether a person speaking English with a Vietnamese accent is a citizen or not.

If nobody asks, don't bring it up. Don't say you're from Korea and you're only here from November 2 to 7.

The moment you tell them when you're leaving, they have no reason to open up about their problems. So don't carry business cards; use a website. And at the bottom, don't put your Bongeunsa-ro address (a street in Seoul's Gangnam district) in Korean. Change it to English or take it off. He also advised against bringing up BTS and kimchi, because locals see people like that every day and won't take you seriously.

On when to set up a US entity, his rule was clear. Do it when a contract is signed, you've actually received $3,000 or $5,000 into your Korean account, and it looks like other customers will buy the same way. The moment you form a Delaware entity, you have to file three or four things, including the Delaware franchise tax and state taxes, and a local tax accountant alone runs 2 to 3 million won a year. There's no reason to burn that money with no revenue.

Never set up a US entity when you don't have revenue.

Two audience questions fit right into this. A CEO who also runs a US accounting firm that issues SOC2 reports asked whether big tech prime contractors really require the certification. The answer: for a software company going after deals of $100,000 or more, SOC2 Type 2 is basically a given. Another attendee had quickly added agentic commerce and gotten traction, but the proxy-purchase setup ran into the payment provider's terms of service. They asked whether it's OK to take actions that cross legal lines.

One of the frameworks Silicon Valley VCs really love is "hacking the system."

He didn't stop there, though. You can't keep going like that forever, he added, so the far more important work is interviewing the people who pushed through the barrier and bought again, and finding out why. I walked out of this session with two questions for myself. Can we write down what we sell down to the job title we're selling to, not just the company name? And are we putting off the first payment because we're busy building a perfect product?

A company that started in a cafe and is now worth 17 trillion won

The main session was with Yuna Kwon of Airwallex. She spent four years and ten months at BCG, going back and forth between Seoul and Melbourne, then seven years and nine months at Shopee running cross-border business for Korea and Japan. She joined Airwallex this May and leads its Korea business and partnerships with financial institutions.

What stood out was how consistent her career criteria have been. Around college graduation she didn't have a clear idea of what she wanted to do, so she chose consulting as the option that would help her next career most. After her MBA, once she knew what she wanted, she changed the criteria.

I wanted to build something from nothing, like running my own business, but I didn't want my finances or my survival on the line. So I decided to look for a business unit that was just getting started at a company with a lot of money.

Shopee fit that exactly, and Airwallex is the same setup. The company has money, and the Korea business hadn't started yet, so she's starting from nothing. Airwallex is valued at 17 trillion won and processes 440 trillion won a year, with APAC as its biggest revenue region. This year it secured payment gateway, prepaid electronic payment and foreign exchange licenses in Korea, and it's preparing to set up a Korean entity by the end of the year. This event was the Korea team's first time meeting the community.

The company started in 2015 in a cafe in Melbourne. The founders were buying coffee beans from overseas, and bank SWIFT transfers couldn't keep up with how fast they ran out of stock. Timing orders against transfers was so painful that they started building the infrastructure themselves. When I asked her to sum up the company in one line, the answer was short.

We handle everything in finance except investing and investment brokerage.

In Japan, not messing up matters more than doing well

Her stories from covering both Korea and Japan at Shopee were interesting. What sellers in the two countries have in common: a strong work ethic and a relatively conservative streak, which gives them high trust ratings on the platform. The differences came after that.

Japan seems like a market where not doing badly matters much more than doing well.

Japanese sellers focus on not making mistakes and not causing returns, she said, but they're conservative about investing to grow sales. They're good at manufacturing but weaker than Korea at marketing. And because the domestic market is big, few small companies export directly; they go out through trading houses or local distributors. On the other hand, they strongly trust third-party specialists, so logistics and marketing service providers are far more varied and specialized than in Korea. Her summary: the ecosystem is solid, but individual sellers are relatively less capable.

The numbers show the difference too. In Japan, SMEs account for about 5% of cross-border trade. In Korea it's 20%. That's why Airwallex had wanted Korea for more than five years. But bringing in a third party would hurt its cost structure and competitiveness, and because it insisted on entering directly, it took more than a year to close the acquisition of a Korean company early this year.

I asked what was behind Korean sellers' 40x growth in five years. There was demand for Korean products and good supply, she said, but the bridge in the middle, logistics, was scary, and that blocked things. Before, sellers were split into two extremes: send each order by Korea Post EMS for 20,000 to 30,000 won, or ship 500 million won of inventory by boat to an Amazon warehouse. Once a middle option appeared, where a small company could ship for 1,000 to 2,000 won and pre-stock just 50 million won of inventory, the numbers jumped all at once.

You should only exchange currency when you want to

I asked a simple question: is opening an account really that hard? Her answer was clear. The moment you become a foreign business in the eyes of the financial system, you pay a penalty, in currency exchange costs and in the time transfers take. With one 100,000 won transfer a month you won't feel it. But as frequency goes up, suppliers have to work around your transfer cycle, and that becomes the bottleneck.

When I asked her to pick just one of the two, she chose fees and exchange.

You should only exchange currency when you want to, but right now most people end up doing it because they have no choice.

Her example was concrete. When a Vietnamese buyer purchases a Korean product on Shopee, they pay in Vietnamese dong and Shopee converts it to USD. If the seller only has a won account in Korea, it gets forcibly converted one more time. The US tariff refunds this spring were an even clearer case. To get a refund you needed a local account; they wouldn't send it to an overseas account at all. When the payer sets the conditions, the receiver can't choose the currency.

On when you actually need this, she did the math. Opening an account and getting it running takes two days on average; put the value of those two days at a few hundred thousand won. Avoiding forced conversions saves about 2% each way, so 4% to 5% round trip. Working backward, the break-even is 5 to 6 million won a month in transaction volume. But if you set it up only after you hit 5 million, you pay switching costs, so it's better to do it before then. I was running the numbers in my head as she talked. In the end the test is whether more than 5 million won a month is moving in and out of the country, and whether you're already losing 4% to 5% round trip to conversions you didn't want.

Someone asked whether you need an overseas entity to use the service. Right now, she said, you can do more with a Singapore, Hong Kong or US entity, and a Korean entity alone has limits. This year's acquisition was meant to close that gap, but it isn't live yet. Starting next year, they're preparing to let companies with only a Korean entity take payments directly from overseas. Her conclusion matched what Peter Shin said earlier.

If you don't have to set up an overseas entity, it's better not to.

One question from the floor was how Airwallex differs from Stripe. Stripe is a gateway for receiving overseas payments, while Airwallex adds global accounts as its main product, so the scope is different. Then came a specific case. With Klarna, which has a large share of buy now, pay later in the US and Europe, if a French customer pays in euros and somewhere in the middle it gets converted into the currency of the merchant's home country, Klarna cancels the payment. The deciding factor is whether the currency the customer sees, the settlement currency and the currency the Korean business receives all stay the same.

When I asked how things looked from the inside compared with the outside, her answer came down to one thing. In Korea, meeting in person matters unusually much, and Koreans are most comfortable doing that with other Koreans. It isn't just language, and foreign teams find the Korea business especially hard. My last question was what those of us going abroad should do. Her answer:

I hope you don't think of yourselves too much as foreign businesses. Just act like a local business, and don't go out of your way to look like a foreign one.

Real bottlenecks from tables of six

Roundtables are the best part of a Bloom event. Six people form a group and share their name, company and why they came. Whoever most recently traveled abroad for work or fun becomes the table leader, and at the end they come up front and present what the group came up with.

The first group agreed that the bottleneck is finding customers. Someone at that table also wondered if so many people are starting companies now that startups themselves are becoming a red ocean. That group included two high school students who said they want to start a company in the US and came to soak up the founders' energy.

The stories from day-to-day payments work were specific. An attendee who does marketing at a retail company shared how, because overseas payment fraud was so common, they deliberately added steps like passport checks instead of removing friction. A sole proprietor who can't use Stripe without a US account said they chain together services that act as the merchant of record on their behalf.

Several tables shared the same worry: deals don't close. One group concluded that Korean politeness actually works against you in business. Because the culture avoids telling someone to their face that it won't work, people say everything sounds great and then the contract never happens. In the US, someone pointed out, people just tell you no. A more senior founder at that table told of a stalled deal that closed right away after he invited the client to dinner with wine at a hotel. He summed it up by saying business sometimes needs you to drop the habit of thinking in numbers and set your ego aside.

The clearest takeaway was about how trust gets bought. The leader of one table builds a service that automates entry paperwork for international students. The day before, they had come back from Uzbekistan after signing contracts with seven universities. That happened, they said, because of an introduction from a local vice governor they had met in Korea first, not because the universities trusted them. When they sold to Korean universities, contracts came only after they brought a university president onto the team. In deals worth tens of millions of won, what the buyers actually bought was the person they trusted, not the product. The same group shared a B2C example: when expanding into Spain, they built a long, detailed landing page Korean style, and the conversion rate went down. Videos also didn't load properly because of internet speeds.

One group collected differences by country. China has no concept of pre-registration at all, so people ask why they should sign up for something that hasn't launched. Marketing there is basically all content, and only video-based work gets traffic. In Japan, if you aren't home to receive a delivery at the scheduled time, they take it back, so the condition of the item comes before the consumer's convenience. That's different from Korea. Their conclusion: if you don't know each country's cultural codes, commerce goes wrong.

The group that took on the language barrier had a memorable takeaway. Koreans speak better English than they think, so if you speak slowly or say up front that your English isn't great, the lowered expectations mean people end up saying your English is good. One attendee concluded that without a local network, the only option is to build one, and decided to go to Singapore in person.

The table that debated when to set up an entity confirmed what both sessions said. Before revenue, it's not time. After revenue starts, there will be a step in collecting payment that requires an entity, and that's exactly when to set one up. The same table also said that a founder's vision and values aren't there from the start; they take shape through experience and trial and error.

Another group heard about companies that pay and get paid in stablecoins instead of fiat because of taxes and regulation, and another said that verifying whether security certifications are valid can help you tell which vendors are frauds. One table leader closed their presentation by saying they'd learned that where there's a bottleneck, there's also a way to solve it.

That's close to what Bloom is trying to do: keep making room for people to talk about what's still stuck, more than what went well. Going global isn't a topic you finish in one event, so we plan to dig into it again from a different angle.

Originally written in Korean on Brunch, September 23, 2026.