今天早上我拿出古坑的水洗豆，研磨，注水，等待。這動作我每天重複，但2026年的台灣咖啡市場，已經不讓人有心情只想著水溫和萃取時間。
信義A11排了一個月的隊。印尼獨角獸凱娜克（Kenangan Coffee）四月插旗新光三越A11館，試營運期間週末單日就吸引350人次。這家市值約12億美元的東南亞連鎖，靠著APP精準推播、自動化設備、預訂自取等待時間壓在兩分鐘以內，85%門市走「抓了就走」小店型——它的對手不是台灣的精品咖啡師，是台灣人每天帶著手機決定喝什麼的那個瞬間。
四方同時壓境，但各自在打不同的仗
凱娜克帶來的是資本戰邏輯：背後有紅杉資本、維港投資，數位會員CRM是護城河，目標2030年在台展店300間。蜜雪冰城代表的低價連鎖路線，在越南市場退燒後仍持續南下試探台灣。這兩股力量本質相同——以規模與補貼換市佔，以數據取代職人判斷。
台灣本土的回應很直接。手搖精品品牌BLIKE把單杯定價拉到388元，不解釋，不打折。這是一個宣示：工藝有它自己的座標系，不在低價戰場裡計算。另一個方向更具體——台灣咖啡師林紹興（Bala）在2026年WCC世界咖啡大賽世界拉花項目中擊敗33國選手奪冠。他服務的品牌COFFEE LOVERs PLANET，已宣佈即將在台北忠孝東路505號B1開設個人品牌據點。世界冠軍頭銜，第一次有了實體出口。
43,000噸進口，800噸本地，這個比例說明了什麼
台灣每年進口咖啡豆43,000噸，本土生產僅約800噸。常見的批評從這個數字出發：台灣的咖啡文化建立在進口原料上，技術再好也只是代工，談不上文化輸出。這個說法看似有力，卻搞錯了問題的框架。
法國沒有可可豆種植園，但沒有人質疑法國的巧克力工藝是否算文化主權。日本沒有波爾多葡萄園，但清酒與葡萄酒的侍酒師文化在東京自成體系。台灣的問題從來不是原料產地，而是詮釋能力——一個產區、一種處理法、一個溫度曲線，能不能在杯子裡說出那片土地的語言。
2016年吳則霖在愛爾蘭都柏林拿下世界咖啡師大賽冠軍，成為亞洲第一位。2024年台灣首次主辦卓越杯（Cup of Excellence），成為全球第17個主辦國，10國33位國際評審盲測，4支台灣本土豆得分逾90分。阿里山海拔1000至1600公尺，日夜溫差20度的山地，讓鄒宿山農園的豆子在SCAA評鑑中拿到90分。這些成績不依賴進口豆，依賴的是對台灣土地特性的精準理解與表達。
機器可以把沖煮流程標準化，但無法把日夜溫差20度的山坡記憶裝進APP推播裡。這就是凱娜克的自動化設備和林紹興的拉花技術永遠不在同一個戰場的原因。
冠軍之後的問題才是真問題
台灣人均咖啡消費量從2000年的28杯增至2020年的200杯，二十年成長超過七倍。路易莎咖啡524家店超越星巴克500家成為台灣最大咖啡連鎖，7-Eleven City Cafe年銷2.8億杯。這是一個真實的消費大國。但路易莎董事長公開說台灣是「全球最難市場」，市場難，難在哪裡？難在本土品牌的數位化能力仍停留在「寄杯」的半數位模式，而凱娜克已經在用CRM系統預判你下午三點想喝什麼。
林紹興從2018年開始參賽，六年內三度拿下台灣選拔賽冠軍，2026年才終於站上世界最高位。這條路沒有捷徑，也沒有APP可以加速。他宣佈開設個人品牌據點，是台灣咖啡師第一次試圖把競賽榮譽轉化為可複製的品牌邏輯——技術主權能不能變成商業護城河，這才是接下來真正值得看的考題。
古坑的農民還在砂質火山土上等著今年的收成。他們不知道信義A11排了多少人，也不需要知道。但如果台灣的精品咖啡圈沒有辦法把農民的土地和冠軍的技術連成一條說得清楚的故事線，388元的溢價就只是一個數字，不是護城河。
— 陳映彤
延伸閱讀
Taiwan’s Coffee War Cannot Be Won by Price Alone
Every morning I grind beans from Gukeng and wait for the water to bloom through the grounds. The ritual takes four minutes. In 2026, four minutes is roughly how long it takes Kenangan Coffee’s app to push a discount notification to a thousand users in Xinyi District.
The Indonesian coffee unicorn — valued at roughly USD 1.2 billion — opened its Taiwan flagship at Xinyi A11 in April. Trial-run weekends drew 350 visitors per day. Its system is precise: automated equipment, digital CRM, mobile pre-order with under two minutes’ wait time, and 85% of its outlets designed as grab-and-go formats. Backed by Sequoia Capital and GIC, its stated target is 300 Taiwan locations by 2030. This is not a restaurant chain. It is a data infrastructure company that happens to sell coffee.
Four Forces, Four Logics
The competitive map in Taiwan’s coffee market now has four distinct players operating under different rules. Kenangan Coffee runs on capital-scale logic. Chinese-origin budget chains such as Mixue, whose low-price model cooled in Vietnam, are testing the Taiwan market from the south. Domestic premium hand-shake brand BLIKE priced a single cup at NT$388 with no apology. And barista champion Lin Shao-Hsing (Bala) beat competitors from 33 countries at the 2026 WCC World Coffee Championship latte art event to claim the world title — then announced plans to open a COFFEE LOVERs PLANET personal flagship at B1, 505 Zhongxiao East Road, Taipei.
Each of these four forces is fighting a different war. Only one of them involves soil.
The 43,000-Ton Question
Taiwan imports roughly 43,000 metric tons of coffee beans annually. Domestic production is approximately 800 tons. Critics use this ratio to argue that Taiwan’s coffee culture is built on foreign raw materials, making technical excellence more akin to skilled processing than genuine cultural authorship.
The argument sounds rigorous and misses the point entirely. France has no cacao plantations; no serious observer doubts French chocolatiers exercise cultural authority over their craft. Japan grows no wine grapes, yet Tokyo’s sommelier culture is globally referenced. The question was never about origin. It was always about interpretation — whether a roaster or barista can translate the memory of a specific altitude, a specific soil chemistry, into something legible in the cup.
Taiwan’s answer, documented: Wu Ze-Lin won the World Barista Championship in Dublin in 2016, the first Asian champion in the competition’s history. In 2024, Taiwan hosted the Cup of Excellence for the first time, becoming the 17th country to do so; 33 international judges from 10 countries evaluated blind submissions, and four Taiwanese-grown beans scored above 90 points. Alishan-grown beans, cultivated at elevations between 1,000 and 1,600 meters with a 20°C diurnal temperature range, produced a SCAA-evaluated score of 90 from Zou Su Mountain farm. These results do not require imported beans. They require knowing what the mountain does to the cherry.
Automated equipment can standardize extraction parameters. It cannot standardize what a 20-degree night-temperature drop does to a Typica at 1,200 meters. That is the gap that separates Lin Shao-Hsing’s latte art from Kenangan’s CRM dashboard — they are not competing; they are operating in different ontologies.
After the Championship, the Hard Part
Taiwan’s per-capita coffee consumption grew from 28 cups per year in 2000 to 200 cups in 2020. Louisa Coffee’s 524 locations now outnumber Starbucks Taiwan’s 500. 7-Eleven’s City Cafe moves 280 million cups annually. The volume is there. The structural gap is digital: domestic brands remain largely on semi-digital models while international entrants deploy real-time behavioral data.
Lin Shao-Hsing competed for six years before winning the world title. His decision to open a personal brand space signals something specific: the championship podium is being treated as a brand launchpad, not a finish line. Whether that conversion — from technical supremacy to scalable brand — succeeds will matter more for Taiwan’s coffee future than any foreign chain’s expansion target.
The Gukeng farmers are tending their volcanic-soil plots at 600 to 1,200 meters, indifferent to the Xinyi District queue. The question is whether the people between the mountain and the cup can build a coherent enough story that NT$388 means something other than a price point.
— 陳映彤
Related Posts
https://justfly.idv.tw/s/Au898Dn