Amazon Didn't Lose China to Alibaba. It Lost China to Seattle.

In August 2004, Amazon paid about $75 million to buy Joyo.com, then one of the largest online sellers of books, music and video in China, and Jeff Bezos announced he was 'very pleased to be entering the Chinese market.'¹² Fifteen years later, Amazon told its Chinese sellers it would 'no longer operate a marketplace on Amazon.cn,' effective July 18, 2019.³ The same company that conquered American retail, built the cloud, and turned a bookstore into a planet-spanning machine could not hold a single percentage point of the world's biggest e-commerce market. The temptation is to call it a fair fight that Amazon lost. It wasn't a fight Amazon lost. It was a fight Amazon never fully showed up for.

The official story is that Alibaba and JD.com simply out-competed a foreign giant on its home turf — better local knowledge, deeper pockets, a friendlier regulator. There is truth in that. But the story leaves out the more uncomfortable cause, the one that lives inside Amazon rather than across the table from it: Seattle never let China be Chinese.

The collapse was real even when the peak is fuzzy

Start with what we can and can't say. Amazon's high-water mark in China is genuinely disputed: one analyst cited by CNBC puts the peak near 15% in 2011–2012, while Euromonitor's head of retailing pegs it lower, at roughly 8% in 2009.⁴⁵ Pick whichever you like — the direction is the same and the destination is brutal. By around 2018–2019, Amazon held about 1% of the Chinese market.⁴⁵ A company that briefly looked like a contender ended up rounding to nothing. The argument here doesn't hinge on the peak year; it hinges on the slope. Something turned a credible entrant into a footnote, and it wasn't a single lost quarter.

Note the table, because it kills the lazy headline. Amazon did not 'leave China' in 2019. It surrendered the one part of the business where it had to localize to win — the domestic storefront, the seller services, the consumer relationship — while quietly keeping the parts where China was useful to its global machine.³ You don't keep the cloud and the export pipeline if the country defeated you. You keep them if you concluded that one specific game was unwinnable the way you were playing it.

Run the same playbook everywhere and it stops being a playbook

Here is the mechanism, worked down. Amazon's superpower is a global system — the same fulfillment logic, the same Prime promise, the same interface, replicated market to market. That sameness is an asset when you're the most sophisticated operator in the room. It becomes a liability the moment a local rival is better at the local game and you can't redesign fast enough to answer them. In China, the decisions that mattered — how aggressively to discount, how to plug into the payment habits people actually used, how to build a logistics network for Chinese cities — were the decisions that needed to be made in China, by people who lived the problem. Instead they were made, or vetoed, in Seattle, by an organization that trusted its template more than its local team. Proximity is judgment. When the judgment lives 11 time zones away, you are always answering last year's question.

That is the difference between losing on scale and losing on governance. A scale problem says: the rival was simply bigger and richer, and there was nothing to be done. A governance problem says: the resources were there, the talent was there, but the authority to use them locally was withheld. The tell is in what Amazon retained. A company beaten on raw scale doesn't keep operating AWS and the export business in the same country.³ A company that mis-governed one division does exactly that — it amputates the part it ran badly and keeps the parts it ran from headquarters just fine.

India is the same company that learned the lesson

Now look at India, where Amazon arrived later and behaved differently. It built physical reality on the ground — more than 100 fulfillment centers across 15 states — and grew into roughly 30–35% of the e-commerce market, second only to Flipkart's ~48%, with the two together holding about 79% of the field.⁸ Amazon India booked Rs 25,406 crore in revenue in FY24, up 14.5% year over year.⁸ That is not the trajectory of a company being driven out. It is the trajectory of a company that decided to actually compete on the ground rather than from a deck.

And the conviction is now backed with the one signal that doesn't lie at Amazon: capital. After roughly $40 billion already invested, the company pledged $35 billion more for India in December 2025, then raised that to a $48 billion commitment for 2026–2030 announced by CEO Andy Jassy to Prime Minister Modi in June 2026 — taking cumulative 2010–2030 investment past $88 billion.⁶⁷ In December 2025, Amazon described itself as India's largest foreign investor.⁷ Set the two countries side by side and the contrast is the whole argument: in one giant emerging market, Amazon wrote down the business; in the other, it keeps writing bigger checks.

But wasn't China just rigged against foreigners?

The honest objection is strong, and it deserves a straight answer: China was harder than India for reasons that had nothing to do with Amazon's org chart. The regulatory environment favored domestic champions, the competitors were exceptional and well-funded, and a foreign retailer faced structural headwinds no amount of local autonomy fully erases. All true. But two things keep the governance read intact. First, those headwinds explain a harder fight, not a forfeited one — and a company that retained its cloud and export operations in the same country clearly judged the consumer game lost for reasons other than being barred from the field.³ Second, India is not an easy market either; it has its own fierce local incumbent in Flipkart, which still leads, and its own thicket of foreign-investment rules. Amazon didn't win India by finding an easier country. It got further by playing the local game more seriously — and the proof is that it's pouring tens of billions into doing more of it.⁶⁸

Amazon's China story is usually filed under 'the foreign giant that couldn't crack the locals.' It belongs in a different folder: the company that brought everything to China except the one thing the market demanded, which was the trust to let China be run from China. India is the same company holding the same template — but this time loosening its grip enough to fit a country it didn't design for. The genius of Amazon's system was always its sameness. The cost of that genius is that sameness travels badly, and the first market that punishes it isn't the one you lost. It's the one you keep running from headquarters, wondering why the numbers won't move.