It is 7:42 a.m. on a Tuesday in Houston. A retired second-grade teacher named Linda flips on a ring light, balances a stuffed panda on her shoulder, and waves at a 6-year-old in Hangzhou who is, in her timezone, halfway through dinner. The class will run twenty-five minutes. There will be a song, a flashcard, and an alphabet. Linda will be paid in U.S. dollars by a Beijing-born company that, for most of the last decade, ran the world's largest live English classroom from servers half a continent away.
The company is VIPKid. The company is also, depending on which year you ask about, a quiet revolution, a regulatory casualty, and a survivor.
01 / THE PROBLEMA planet of English learners, a thin layer of English teachers
In 2013, the math of English education in China was awful. Millions of urban parents wanted their children speaking like American second-graders by the time they hit fifth grade. The country had roughly one credentialed native English-speaking teacher for every several hundred thousand of those families. Cram schools tried; the supply was finite, the prices were rising, and the quality was uneven. Meanwhile, in Cleveland, Sacramento, and Halifax, certified teachers were grading papers at 11 p.m. and wondering how they'd cover the heating bill.
Cindy Mi - founder, CEO, and reformed teenage tutor - looked at this gap and saw not a problem but an arbitrage. The teachers existed. The students existed. What didn't exist was the wire between them. Building that wire is essentially the entire VIPKid story.
02 / THE BETA founder who started teaching at 15
Mi's biography reads like something her own marketing team would not be allowed to write. She started tutoring her classmates in English at 15. She dropped out of high school at 17 to run a small chain of brick-and-mortar English academies with her uncle. She later went back, got her MBA in Beijing, and in 2013 pitched the idea of teaching English over video to Kai-Fu Lee at Sinovation. He wrote a check more or less on the spot. The platform launched a little over a year later.
The bet was simple, and slightly absurd: that Chinese parents would pay real money to put their five-year-olds in front of a webcam with a stranger from Ohio. The bet worked because Mi got two things almost obsessively right. First, the format - twenty-five minutes, one-on-one, scripted but warm. Second, the teachers - vetted, paid well, and treated like a brand rather than a back-office cost center.
03 / THE PRODUCTTwenty-five minutes that built a marketplace
If you have never seen a VIPKid class, the choreography is hard to fake. The teacher logs in fifteen minutes early. The platform serves them a slide deck for that specific student, that specific level, that specific day. The student appears in a tiny window. There are puppets. There are rewards stars. There is a song about apples. Twenty-five minutes later, both sides log off and the platform pays the teacher, in U.S. dollars, automatically.
The product is not the lesson. The product is the wire - the matching algorithm, the curriculum stack, the payments rail, the in-class video tooling, the customer service team handling parent complaints in two timezones at once. By 2019, VIPKid was running well over 200,000 class sessions a day. Most edtech companies are still trying to run that in a week.
04 / THE PROOFMoney, partners, and a lot of star stickers
For a company that operated entirely outside the academic establishment, VIPKid collected an unreasonable number of stamps of approval. Sequoia Capital China backed it. Tencent backed it. Yunfeng Capital, Coatue, Northern Light - all of them backed it. It partnered with ETS to align its curriculum with English proficiency benchmarks, and with TESOL on teacher development. Fast Company put it on the Most Innovative list in 2018. The World Economic Forum named Mi a Young Global Leader.
The cash chart is the easiest way to see how fast belief compounded.
VIPKid Funding Rounds
A timeline, with one bad year
05 / THE PIVOTWhat happens when a country rewrites your business overnight
The honest part of any VIPKid story has to include July 2021. Beijing's "double reduction" policy, aimed at cooling an overheated tutoring economy and easing pressure on schoolchildren, did several things at once. It banned for-profit academic tutoring on weekends and holidays. It barred companies from going public on the tutoring business. And it specifically prohibited the hiring of foreign-based instructors to teach the K-9 academic curriculum - the precise wire VIPKid had spent a decade building.
By October, the core program was gone. The company moved its headquarters to Singapore in early 2022 and quietly began rebuilding around international students - kids in Southeast Asia, the Middle East, and Latin America who want to learn English from a North American teacher and don't have a regulator standing between them and a webcam. The teacher community that survived is smaller, more global, and, by most accounts, every bit as committed to a curriculum someone in Beijing wrote a decade ago.
06 / THE MISSION"Inspire and empower every child for the future"
Mission statements are usually wallpaper. VIPKid's is unusually load-bearing. The thing the company actually built - the matching, the payments, the puppet show - is in service of an old, almost unfashionable idea: that a child anywhere should be able to learn from a great teacher anywhere. The geography is incidental. The wire is the product.
You can argue about whether VIPKid was a tutoring company, a marketplace, or a media platform with very young customers. You can argue about whether the post-2021 version of the business will ever match the scale of the pre-2021 one. What you cannot argue is that, for several years in the middle of the 2010s, VIPKid taught the rest of the world what live online learning was supposed to feel like. The Zooms came later. The puppet came first.
07 / WHY IT MATTERS TOMORROWThe wire outlives the regulator
The temptation, with any company that has had a hard chapter, is to grade it on the chapter and move on. That undersells VIPKid. The infrastructure it built - the marketplace mechanics, the per-class economics, the certification rails for distributed teachers - is now the default playbook for live online tutoring everywhere. Outschool borrowed pieces. Preply borrowed pieces. Every English-from-the-Philippines startup borrowed pieces. The pandemic-era boom in live tutoring did not invent the format. It inherited it.
It is 7:42 a.m. in Houston. Linda's ring light is on. The panda is in position. The child in Hangzhou is older now - or maybe she's a child in Jakarta, or São Paulo, or Riyadh. The platform routing the call is smaller than it was, headquartered somewhere new, and run by a founder who has had to rebuild her company once already. The class is still twenty-five minutes. The first slide is still an apple. The teacher is still getting paid in U.S. dollars before the kid finishes dinner. The wire is still there.