Summary

Din Tai Fung is the top-grossing chain restaurant in the U.S. on a per-store basis. Average annual revenue per location is $27.4M, twice that of the No. 2 high-end steakhouse, despite having only 21 U.S. stores and reaching as many as 19 table turns in a day. Amy traces the formula to 1996, when Takashimaya helped the brand standardize and internationalize its operations—18 pleats per dumpling and open kitchens included. From 2000 to 2013, Din Tai Fung opened only 4 U.S. stores; after 2016, it abandoned Chinese-American neighborhoods entirely and went only into malls with strong foot traffic and brand pull. Its 2024 New York flagship, spanning 2,000+ sq m with 450 seats, validated the large-store format.
Gong cha has vanished from mainland China but is the No. 1 tea-drink franchise in the U.S. It has nearly 300 stores and has topped Entrepreneur’s Franchise 500 tea-drink rankings for 5 consecutive years. The story is one of successive capital handoffs: in 2012, Korean homemaker 金汝贞 and her husband brought the brand to Korea and upgraded it; 2 years later, they sold 70% of their stake to Japanese PE firm Unison Capital; in 2019, Boston private-equity firm TA Associates acquired it for $288M, moved headquarters to London and positioned it as the Starbucks of tea; and “a few days ago,” the company was reportedly seeking a whole-company sale at $2B.
The guests see U.S. foodservice as a real-estate opportunity. Landlords hold the upper hand, leases can run 10-15 years, and Starbucks’ exclusivity agreements can shut tea-drink brands out of an entire shopping center. In the Bay Area, the assignment fee for a good bubble-tea store can reach $1M. Gong cha and Kung Fu Tea’s first-mover advantage was securing prime locations early; later entrants can only wait for a downturn to free up space or take over 2nd-generation stores.
U.S. consumers’ tastes change slowly. “The supply side has not yet become broad enough to change demand-side tastes”: Quickly, which uses non-dairy creamer, can still draw lines at new Bay Area stores, while fresh milk and real fruit are not expensive in the U.S. and therefore “are not inherently something that feels especially valuable.” What Chinese food and tea brands sell in North America is definitely not flavor; it is setting, experience, culture and identity. Malatang, with its high consumer-education burden, “can only serve Chinese and Asian customers.”
Korea and Japan have exported food culture through a joint push from government and capital. In 2019, Bibigo spent nearly $2B to acquire a U.S. frozen-food giant and gain access to Walmart and other mainstream channels; data found by 泓君, though explicitly described as unverified, reportedly put Bibigo’s North American frozen-dumpling share at 42%. Thailand’s government has put real money behind “food diplomacy,” helping Thai cuisine account for 11% of U.S. Asian restaurants even though Thai Americans make up only 2% of Asian Americans. Chinese food, by contrast, has “no favorable wind at its back”: Cosmo took an order from the South Korean government to promote Korean pears but says he has “never received a single order from any Chinese government” to promote Chinese food.
Capabilities honed by China’s hypercompetitive market do not all transfer directly. Digital operations are a “dimensionality-reduction strike,” but labor productivity does not travel: “The one thing you absolutely cannot say here is, ‘If you don’t want the job, someone else will do it’—because if you leave, there really may be no one to replace you.” A PR promise to open dozens of stores before “202X” was dismissed as unrealistic: approvals take 6-12 months, and Panda has opened only 2,000+ stores globally in 50 years.
The dark horse may be Molly Tea, a domestic 2nd-tier brand. Its U.S. store count is second only to HEYTEA, and its best store reportedly does $500K in monthly sales, with Asian and mainstream customers seemingly split about 50-50. It was brought in by local franchisees who understood American culture, and its light, minimalist Chinese aesthetic carries little cultural friction. CHAGEE’s LV-style campaign feels “really far removed” to Amy, while Wallace’s pure value-for-money pitch “cannot” win. The biggest taboo in overseas expansion is one line: avoid self-hype—respect consumers and make time your friend.

Deep dive

1. Din Tai Fung: a 50-year North American per-store sales run, with the key leap in Japan

The opening data point is counterintuitive: the Chinese restaurant most loved by mainstream Americans is Din Tai Fung, which once withdrew from the North China market. Average annual revenue per store is $27.4M, the highest among U.S. chain restaurants and twice that of the No. 2 high-end steakhouse.
Amy’s history lesson: in 1958, 杨秉彝, a Shanxi native, started by selling cooking oil in Taiwan. When he shifted to xiaolongbao in the 1970s, “no one paid attention”; it took roughly 10 years and a wave of newspaper coverage before customers began lining up. Being named one of the New York Times’ “10 Best Restaurants in the World” in 1993 was the turning point.
Entering Japan in 1996 was the decisive internationalization move. Japan was in an economic downturn, while department stores had just started using restaurants to drive traffic, and Takashimaya saw the potential in Din Tai Fung. The standards established in Japan—18 pleats on every xiaolongbao, open kitchens that let customers watch the wrapping process, and consistent output—directly supported the high table turns of its later U.S. stores.

2. U.S. expansion: 4 stores in 13 years, then a complete rewrite of the site-selection logic

The first Din Tai Fung opened in Arcadia, Los Angeles, in 2000, targeting a Chinese-American neighborhood. By 2013, the chain had only 4 U.S. stores; early expansion was “extremely, extremely slow.” The inflection came in 2016, when it moved into Arcadia’s best mall. The site-selection logic changed completely: no more Chinese-American neighborhoods, only malls with heavy traffic and meaningful brand pull.
The 2024 flagship in central Manhattan covers roughly 2,000 sq m and seats 450, delivering “a highly successful validation of the large-store model.” The newly opened Arizona store should have its reservations fully booked for 2 weeks. The U.S. network still has only 21 stores, most of them on the West Coast.
Table turns are the engine behind the sales record: the chain can reach 19 turns in a single day. 泓君 noted that a turn rate of 10 is already extraordinary; 19 is “an astronomical number.” Amy’s explanation is straightforward: full-service steakhouses turn tables extremely slowly, while Din Tai Fung’s limited store count creates an overflow effect. At Times Square, visitors and foot traffic from every demographic flow into the restaurants.
泓君 added a governance detail from the family’s 3rd generation: “Luckily, we never went public.” Otherwise, the company would have been pressured to open dozens of stores a year and would have lost the sense that Din Tai Fung is a place people aspire to visit.

3. Category is destiny: xiaolongbao as “a dumpling variant + Lava Cake”

Amy’s judgment is blunt: Din Tai Fung chose the right category. “If it had made squirrel-shaped mandarin fish, I don’t think it would have become this popular, because Americans don’t eat fish bones.” Xiaolongbao is a variation on the dumpling, a format Americans already understand, while the burst of soup resembles the molten center of the familiar Lava Cake—simultaneously foreign and familiar.
The local adaptations are real. Chocolate-filled xiaolongbao dipped in salted-cream cheese went viral on TikTok, while ube-purple-sweet-potato xiaolongbao also performed well. The menu is “extremely, extremely smart”: small portions of cucumber salad and red-oil wontons let Americans order a broad spread, eat quickly and drop in even at 3 p.m., shedding the formal-dining label and keeping traffic flowing.
泓君’s summary comes down to 4 factors: the category fits mainstream eating habits; the brand standardized early through its move into Japan; the early years focused on product quality and per-store economics; and the menu underwent extensive local experimentation.

4. Gong cha: a brand that vanished from the mainland, lifted to the top of U.S. tea franchising by 3 rounds of capital

泓君’s finding: the largest bubble-tea chains in North America are not HEYTEA or Mixue but Kung Fu Tea and Gong cha. Gong cha now has 300+ stores, has expanded deep into areas without concentrated Chinese populations, and has ranked No. 1 among tea brands in Entrepreneur’s Franchise 500 for 5 consecutive years.
Henry offered the episode’s fullest capital timeline: Gong cha was founded in Taiwan in 2006, then brought to Korea in 2012 by Korean homemaker 金汝贞 after she tried it in Singapore and saw that it was “sweet, addictive, easy to standardize and highly profitable.” Her husband, Martin Berry, was a financial operator. Together they established UCK Partners, executed the first stage of internationalization, upgraded the Korean aesthetic and added celebrity marketing. 2 years later, they sold 70% of UCK to Japanese PE firm Unison Capital, whose Korean arm was a formidable consumer operator. In 2014, Gong cha entered the U.S. through the country’s established franchise system and scaled rapidly.
In 2019, Boston private-equity veteran TA Associates acquired Gong cha for $288M, or about ₩350B, moved headquarters to London and began operating it toward a global “Starbucks of tea” ambition. “A few days ago,” the company was reportedly seeking buyers for the entire business at $2B. Amy’s point about the model: it looks like a high-investment, high-profile rollout, but in practice it avoids heavy company-owned-store spending and relies on strong local franchisees.

5. Location moat: U.S. foodservice “is a real-estate opportunity”

Cosmo and Amy explained why landlords hold so much power. Good locations are far scarcer in the U.S. than in China, and a restaurant built from scratch must clear a complicated government-approval process. That makes an existing restaurant property the preferred site; operators may even pay heavily to buy the business outright. “I’m buying time.” Leases can run 10-15 years—“once you open, you can operate for 30 years without moving”—and the assignment fee for a good Bay Area bubble-tea store “can reach $1M.”
Non-compete protections are an invisible barrier. Starbucks’ exclusivity agreements can keep other tea-drink brands out of an entire shopping center. When Gong cha entered with capital in 2014, it converted the first wave of failed coffee and juice stores into bubble-tea shops. “Once you secure the location, you can do business there for a long time.”
The route for later entrants is the 2nd-generation store: a location that was already a bubble-tea shop and is now being transferred. Taking over a store that the previous operator could not make work lets a new brand inherit the existing customer base while adding its own. But the diagnosis matters: was the problem the location, or an aging operator and concept? Many such stores belong to an older generation that no longer wants to work, while the 2nd generation may be a lawyer with no interest in taking over.

6. Don’t think about U.S. business with a Chinese palate: slow taste cycles are structural

泓君 asked the pointed question: Gong cha still uses non-dairy creamer and large pearls, a legacy-generation profile. Can mainstream consumers really tell the difference from modern tea brands? Amy’s answer: a meaningful minority can, “but most American consumers actually like, and are even accustomed to, the taste of the previous generation of tea drinks. The supply side simply has not become broad enough to change demand-side taste.”
The best evidence is Quickly, an even older brand that recently opened a Bay Area store and still drew lines with its traditional non-dairy-creamer formula. Cosmo compared it with Chinese consumers’ first exposure to steak: a griddle steak at Hao Ke Lai, a concept that does not exist in the U.S. “We’re still using our own frame of reference to think about this.” Fresh milk and real fruit are already inexpensive in the U.S.; they are “not inherently something that makes you feel you are getting a lot of value.”

7. They are selling the occasion—not flavor—and a quantifiable definition of going mainstream

Henry’s subjective threshold is clear. In a Chinese-heavy neighborhood, having 40-50% of customers speak English is already “a remarkable result.” But in a predominantly white neighborhood, if 40% of customers are Chinese, “that means the brand has not actually entered the local market”; it is still relying on hometown flavors and destination visits. An overseas brand has to decide whose business it wants before translating that decision into site selection.
The answer to what these brands are selling is blunt: “Definitely not flavor.” Gen Z discovers modern tea drinks on TikTok, and buying a matcha or even HEYTEA is an expression of identity. Cosmo’s mirror image is Western fast food in China: McDonald’s is an everyday meal in the U.S., but in China it became a place to go for a birthday or a premium occasion. “What is an everyday experience for Chinese people cannot immediately become an everyday experience for Americans. At first, it is simply an experience of novelty.”
Malatang is the concrete counterexample. Cosmo’s verdict is that it “can only serve Chinese and Asian customers”: Americans have no concept of blanching food at the table, broccoli is eaten raw—“I went to my son’s classmate’s party, and we ate the broccoli raw”—and customers do not know how to order. People compare it with Chipotle, but at Chipotle you can simply take the food and eat it. Japanese ramen offers an unexpected bridge: Asian, Black and white customers can treat malatang like ramen because they recognize noodles, meat and corn, while tripe and braised pig’s feet are “things they would not even touch.”

8. Fei Da Chu changes its signature dish: replace A with B, or run A and B together?

Fei Da Chu replaced its signature stir-fried pork with chili peppers with beef and broccoli for the U.S. market. The guests drew a distinction: broccoli and beef is “definitely better than stir-fried pork with chili peppers” because Americans are more comfortable with it and many cannot handle the heat. But the U.S. still has not seen a stir-fried dish that makes people travel thousands of miles to eat it; the commercial answer can only be tested after the stores open.
Henry’s framework is the core point: “Replacing” and “adding” are 2 completely different decisions. If a brand simply swaps out the original, it loses its memory hook. “If Din Tai Fung changed all of its pork fillings into fillings Americans love, it would no longer be Din Tai Fung. Jollibee would only become the Filipino version of KFC.” The better approach is to start with the core customer base and iterate through localized SKUs, continually widening the consumer’s bandwidth.
Henry, as a Hunan native, offered a personal example: if he saw no stir-fried pork with chili peppers on the menu, “obviously I would not go.” That layer of customers is the most loyal to the brand; discarding the hometown memory likely means giving up the base. UCSD alone has roughly 3,800-5,200 Chinese students, so the initial customer base is already large enough.

9. Korea and Japan provide the cultural scaffolding; Bibigo provides the channel lesson

Henry cited 2 Korean examples. CJ Group’s Bibigo regularly sponsors the KCON K-pop concert series, converting fans of Korean culture into fans of the brand. Shin Ramyun immediately caught the momentum behind the Netflix animation K-POP DEMON (“K-POP Demon Hunters”) in North America and launched limited-edition packaging. “When your daughter points at the K-POP Demon Hunters package and says, ‘Mom, I want that ramen,’ it is hard to say no.” Japan started earlier: if an anime is set in Japan, “there is always ramen.” Brands have deliberately or instinctively aligned their restaurant settings with the worlds depicted in Midnight Diner and Naruto: “I take a bite of ramen, and I become a ninja in Naruto.”
Bibigo’s 42% share of the North American frozen-dumpling market came from business execution, not culture. In 2019, it spent nearly $2B to acquire a U.S. frozen-food giant and immediately gained access to Walmart and other mainstream grocers. “Distribution is almost the lifeline.” Americans did not suddenly develop a taste for Korean dumplings; when they wanted dumplings, Bibigo became the package they could actually pick up.
There is a hidden product-design detail: the same dumpling can be pan-fried, boiled or baked because “Americans really do not use boiling water for dumplings very much.” They generally use a microwave or oven. Many Chinese manufacturers entering the North American frozen-food market have not even figured that out.

10. The Thailand lesson: government-backed “food diplomacy”

Thai restaurants account for 11% of all Asian restaurants in the U.S., making Thai food the No. 3 most common Asian cuisine nationwide, even though Thai Americans represent only 2% of Asian Americans. Henry attributes the gap to a Thai government program launched in the early 21st century: the state funded chef training, English education and embassy-backed restaurant events; new restaurants could even apply for government support, while official relationships helped with practical matters such as site selection. “Government power was pushing from behind.”
Other factors helped. Southeast Asia is an affordable travel destination for Americans, who stay an average of 12-15 days and are already familiar with the flavor profile. North American Thai restaurants have also broadly standardized a milder flavor profile and more localized dish names. Restaurants that adhere completely to the original taste do not represent the mainstream.

11. Why Chinese food cannot connect culturally: no “favorable wind,” no coordinated push

Henry identified 3 areas where China falls short. First, “I cannot find the favorable wind.” China’s strongest current cultural products are either costume romances or short dramas about domineering CEOs. The food shown in costume dramas cannot be replicated in modern life, while the CEOs “all eat red wine and steak.” “So I have no favorable wind at my back; sometimes it is hard to rise with the current.” Second, Sichuan, Cantonese and Hunan cuisines each have their own agenda and are difficult to consolidate into the shared culinary memory that Korea and Japan have built. Third, the industry swings between 2 extremes: excessive fidelity to regional identity or total abandonment of brand memory in pursuit of an imagined mainstream palate.
On the claim that Chinese cuisine is too fragmented, Henry pushed back: Thailand and Italy are also full of regional cuisines, with southerners looking down on what northerners eat. “Fragmentation has no particularly direct relationship with whether a cuisine can be commercialized overseas.” The way forward is to set aside excessive pride, follow the market and deliberately scale dishes with the broadest potential appeal. “Whether the dish originated in northern Shanxi or southern Sichuan is not that important to local consumers. Succeed commercially first; the work of tracing its cultural roots can come later.”
Cosmo contrasted China with the NRA restaurant show. Japan builds a small pavilion; Italy hands out a cheese-and-wine scorecard at the entrance. Behind them are capital, government, civil organizations and companies acting together. Chinese exhibitors are “scattered, with a random roll-up banner printed somewhere,” without a unified presence. The sharpest line: “Years ago, we took an order from the South Korean government to explain how delicious Korean pears were. I have never received a single order from any Chinese government to promote local Chinese food. Not once.”
There are 2 positive examples. Haidilao has built around the scarcity of experience and destination consumption in North America: face-changing performances, individual hotpots and staff handing customers hair ties turn hotpot into a social, cultural experience. In HEYTEA’s lab stores, “people of different skin colors are all relaxed inside, meeting friends and chatting,” creating a cross-ethnic, lightly social tea-drink setting—a second space much like Starbucks.

12. Which China-honed capabilities transfer to the U.S.: digital operations do, labor productivity and “chasing novelty” do not

The transferable advantage is digital: apps, mini-programs and private-domain customer operations are “far ahead of many U.S. local brands,” amounting to a decisive advantage. Labor productivity does not transfer. In the U.S., “the employer may be Party B and the employee Party A,” and “one careless move can get us sued.” Cosmo says he spends most of his time running a restaurant on emotional counseling. “The one thing you absolutely cannot say here is, ‘If you don’t want the job, someone else will do it’—because if you leave, there really may be no one to replace you.” Even influencers will turn down a brand: “It is not about the money; they simply think your brand is no good.”
A PR claim to open several dozen stores before “202X” was dismissed on the spot as unrealistic. Fire and health approvals take 6-12 months, while landlords want to know whether the operator has a local team and prior operating experience. Panda spent 10 years refining its model after launching in 1973 and still has only 2,000+ stores globally. “You have already missed the earliest batch of locations”; the next opportunity is to wait for established U.S. restaurant brands to weaken and exit.
Competing on innovation “may not work very well.” Cosmo recalled coming to the U.S. more than 10 years ago from Renren—“back then, we were just copying Facebook”—after China had already cycled through SNS, O2O and P2P, while U.S. social media remained Facebook, Instagram and YouTube, with only TikTok added. “Americans are not that obsessed with the new.” Henry added that the core customer base is much older than the domestic new-tea-drink audience; innovating for its own sake wastes huge amounts of labor. Better to refine the 4 best-selling core SKUs. The winners use “micro-innovation” around mature products: chocolate xiaolongbao, mochi doughnuts and waffles with fried chicken—familiar foods recombined with a few Asian elements.

13. Molly Tea’s dark-horse logic vs. CHAGEE’s fashion ambition

泓君 specifically asked about Molly Tea rather than HEYTEA. Based on data she found, the brand may belong to China’s 2nd tier, yet its U.S. store count is second only to HEYTEA. She had heard that its best store generates $500K in monthly sales, with Asian and mainstream customers seemingly split about 50-50. Henry’s explanation: the brand’s product and aesthetic carry “no particularly heavy cultural barrier”—clean, modern and easy to communicate. Its marketing goes “2 steps outward, but not many more; if you go too far, you pull the whole thing apart.” It precisely targets pan-Asian customers, ABCs and people familiar with 1.0/2.0 bubble tea; whether it has truly reached white and Black consumers “remains a question mark.”
Amy added the key mechanism: Molly Tea was originally brought to the U.S. by local franchisees who understood American culture, and “it grew big almost by accident.” Marketing and influencer selection were handled locally. That answers 泓君’s central question: smaller brands that do not depend on layers of approval from a Chinese headquarters may actually have better odds in North America. A telling detail is the brand’s minimalist Eastern negative space, which stands apart from Chinatown’s red-and-gold palette. Its fragrance merchandise hits the standard markers of a middle-class lifestyle, and some white college students are curious because “Molly” is a familiar drug-slang term.
CHAGEE’s assessment is mixed. Henry praised its decision to simplify after entering the U.S.: dropping Peking opera references and complex patterns in pursuit of something not only modern but fashionable, with the ambition that “the cup in the hand of a sparkling Los Angeles girl is not Starbucks but tea.” It is “a highly challenging image.” Amy’s direct assessment was harsher: the early LV-style TV campaign had “no appetite appeal for me at all” and felt “really far removed.” “Milk tea only needs to be affordable luxury; if you take it too high, people stop going.” She does, however, see the brand changing.

14. Wallace’s value-for-money illusion, and the final advice: avoid self-hype

Can Wallace win on value in the “home turf of fried chicken”? The answer was unequivocal: “No.” Cosmo used Jollibee as the comparison. When it entered the U.S. 20 years ago, it had a base of roughly 4.2M Filipino immigrants: “Wherever there are Filipinos, there is Jollibee.” Sweet spaghetti and Southeast Asian-style fried chicken supplied differentiation and a powerful emotional link to home, so “even without advertising, the base was stable.” Only later did Jollibee add Angus burgers and chicken tenders that better matched American habits, taking share from McDonald’s and Chick-fil-A at prices 10-15% below comparable McDonald’s meals. Wallace, by contrast, sells a domestic “substitute” product, goes straight after the mainstream and has no cultural connection: Chinese customers may not go, while Americans may not be won over. Henry noted that Mixue may actually occupy the “homegrown hero” position, with many customers identifying with its value proposition.
The closing advice was consistent. Cosmo: “Avoid self-hype.” Henry: “Respect consumers. North American consumers are ordinary people too. If you lower your head, understand clearly who they are and how they are willing to consume your food, commercial success will come sooner or later.” Cosmo’s synthesis: the U.S. is a market with longer time cycles, so “make time your friend and get your hands dirty”—without a foundation, a tree has no roots. His final jab: “Send a professional manager over to study the market for 2 months, then go back and open the store—I simply cannot understand it.”

(Preview) Mythos and Project Glasswing, The Year of Anthropic Continues Apace, Q&A on the NYT, Altman, De-globalization Vikings, Ragnar, Berserkers, Valhalla & the Warriors of the Viking Age | Lex Fridman Podcast #495