Starbucks opened in Ho Chi Minh City in January 2013. Thirteen years later it has about 150 stores in a country of 100 million people. Highlands Coffee, a local chain, has more than 1,060. And Café Amazon, the Thai giant with 4,400+ stores across Southeast Asia, closed its last Vietnamese shop in November 2025 after five years and 12 stores.
This is not a story about Vietnamese people disliking foreign brands. It is a story about cost, habit and speed. Here is how it works, and what a founder can take from it.
The market: a slower pie, a faster slice
The whole market is slowing
Vietnam's food and beverage market was worth about US$28 billion in 2025, according to iPOS and Nestlé Professional. It grew 16.6% in 2024, then only about 5.5% in 2025, and the forecast for 2026 is 4.6%. Nearly 60% of F&B businesses reported lower revenue in 2024.
Chains are the fast-growing slice
Chains are the exception. Momentum Works puts Vietnam's coffee-chain market at about US$725 million in 2025, up 27% and the third largest in Southeast Asia. That is roughly five times the growth of F&B overall (my arithmetic). Independent cafés close; chains with a repeatable format take their customers.
Highlands: from a bag of coffee to 1,000+ stores
The timeline
Growth by the numbers
Stores grew about 19 times from 56 in 2012 to 1,062 in mid-2026 (my arithmetic). In 2024 an average store sold VND15–16 million a day, about US$630. Growth is speeding up again: in Q2 2026 Jollibee reported Highlands' system-wide sales up 46.7% and same-store sales in Vietnam up 11.5%.
The founder who said no to Starbucks
The founder story matters here. Thai says Starbucks once offered to buy Highlands. After about six months of talks he learned Starbucks would run both brands, with the best locations going to Starbucks. He declined and chose to compete instead. Fourteen years after selling a minority stake to Jollibee, he is back in control.
Two local strategies global brands cannot copy
The two playbooks side by side
Strategy 1: coverage (Highlands)
Highlands wins on coverage. It put small stores wherever people already walk, kept the menu short and local, and used kiosks to test sites cheaply. A 2023 press estimate puts rent at roughly VND9–10,000 per cup sold at Highlands against about VND14,000 at Starbucks.
Strategy 2: experience (Katinat)
Katinat wins on experience. It started in 2016 and grew from 10 stores at the end of 2021 to about 115 by December 2025, with around 120 planned for 2026, by taking prime corners, giving each store its own design, and releasing seasonal drinks that young customers post about. It charges more than the mass chains, and customers accept it because the space is part of the product.
A third model: Milano
A third model, Milano, shows the same logic from the other end: about 2,500 franchised outlets in small 30–70 sqm stores. Three local models, three ways to cost less to serve than a global brand.
Why Starbucks and Amazon could not follow
Starbucks: growing, but slowly
Starbucks has been steady but slow. It is still growing, but Highlands has about seven times as many stores. Its 2020 revenue of just over US$30 million was roughly a third of Highlands' (my arithmetic). In 2018, the latest comparison available, a 200 sqm Starbucks in Saigon cost about US$215,000 to open, against US$86,000 for The Coffee House. When Starbucks arrived, a drip coffee at a corner café cost VND10–15,000, under US$1.
Vietnam is the one large SEA market where Starbucks never broke out, even while it was opening hundreds of stores next door.
Café Amazon: a clean exit
Café Amazon is the cleaner failure. In Thailand it wins through petrol-station sites and sheer scale. Neither edge travelled. It entered Vietnam in November 2020 through a joint venture, opened about 12 stores, and Thai partner Centel dissolved the venture in 2025 citing sustained competitive pressures. Highlands had 928 stores when Amazon left, about 77 times as many.
Same owner, different results
The sharpest test is inside one company. Jollibee owns Coffee Bean & Tea Leaf, a global brand that has had only about 15 Vietnamese stores since 2006. Jollibee also holds a large stake in Highlands (60% until September 2026, now 49%). Same parent, same capital, very different results.
Five barriers
Five barriers explain the pattern:
Cost per cup. Prime-site rent and imported beans raise the price of every drink.
Price anchor. Corner cafés already sell coffee for a fraction of a global-brand price.
Taste. Vietnam grows robusta and drinks it strong, with condensed milk. Highlands sells that; global menus sell something else.
Habit. Neighbourhood cafés with Wi-Fi are already the third place.
Capital discipline. A licence-based operator must test each prime site carefully, so it expands slowly.
My take
Out-scaled, not out-branded
Global brands do not lose in Vietnam because of taste or nationalism. They lose because a local chain can serve a cup of coffee for much less money, in more places, in a format people already know. Highlands did not out-brand Starbucks; it out-scaled it.
Five takeaways for founders
Five things a founder can use:
Scale a local habit; do not import a global identity. Highlands turned an existing routine into a chain. Starbucks tried to sell an aspiration.
Win on cost to serve before you win on brand. Work out your cost per unit after rent, duties and imports. If a local rival serves the same customer for half, brand will not close the gap.
Treat format as a moat. Small stores, kiosks and franchise let a company grow in cities where land is expensive.
Pick a lane. Mid-price scale (Highlands), premium experience (Katinat) or low-cost franchise (Milano) all work. The lane global brands often choose, premium with prime rent, is the most expensive one.
If you sell, keep control or a path back. Thai sold part of Highlands in 2012, kept running it, and bought control back in 2026.
A quick test before you enter a new SEA market
What does the nearest local substitute cost today?
What is my cost per unit after rent, duties and imports?
Can I open at one-third of my home-market cost?
Who already owns this habit, and how many places do they cover?
Would I rather have 5 great stores or 500 good enough ones?
What's next
This Wednesday: What 250 Cafe Owner Told Reddit, I analyze over 250+ comments from Cafe Owner on Reddit and come up with pain points, solution and opportunities for you.
Frontier is built on beehiiv. If you've been meaning to start a newsletter, this is where I'd begin.
— Jetni
