---
title: "Tesla, Huawei, Starbucks: Three Foreign Market Entries"
description: "Tesla in Shanghai, Huawei in Thailand and Starbucks in Australia compared by size of commitment, earlier steps and cost of leaving, with three questions to ask."
url: "https://staging.khoi.co.uk/en/blog/tesla-huawei-starbucks-ba-cach-vao"
lang: en-GB
author: "Lê Quốc Khôi"
date: 2026-11-03
updated: 2026-11-03
topics: ["Strategy & leadership"]
tags: ["Entry modes", "Market exit", "Tesla", "Huawei", "Starbucks"]
---

# Tesla, Huawei, Starbucks: three ways in, three sizes of bet

## Key takeaways

- Tesla broke ground on its Shanghai plant on 7 January 2019: investment of over CNY 50bn (USD 7bn) and capacity of 500,000 cars a year, against 245,240 cars delivered worldwide in 2018 (Xinhua, 2019).
- On 11 November 2020 Huawei Technologies Thailand announced plans to invest THB 700m (about USD 23m) in a third data centre, which Xinhua words as equal to the first two combined (Xinhua, 2020).
- In August 2008 Starbucks closed 61 company-operated shops in Australia, leaving 23; the restructuring charge was USD 16.9m, of which USD 11.6m was lease exit costs (Starbucks Corporation, 2008). BBC offers at least three explanations, none of them measured (BBC News, 2008).
- The library's entry-mode table cannot tell the three firms apart [Foreign market entry modes](https://staging.khoi.co.uk/en/framework/market-entry-modes). What separates them is the size and order of the capital steps, as I read it.
- The bubble-tea chain near the end is invented, made up by me.

In August 2008 Starbucks closed 61 of its shops in Australia and kept 23. A little over ten years later, on 7 January 2019, Tesla broke ground in Shanghai on a plant it priced at USD 7bn. In November 2020 Huawei Technologies Thailand announced plans to spend about USD 23m on another data centre. All three were investing on the spot, yet the amounts at stake in each step, and the price of leaving, were very different. This post sets them side by side. The sources are news reports, a company report and press material, so it can measure size but not results.

## Where do the three sit on the entry-mode table?

Tesla and Starbucks sit in the same column: a subsidiary the firm owns outright. Huawei is a direct investment by a local subsidiary. The table therefore cannot separate them.

Xinhua says Tesla was "the first to benefit from a new policy allowing foreign carmakers to set up wholly-owned subsidiaries in China" (Xinhua, 2019). Starbucks' fiscal 2008 report says it "continues to have wholly owned operations in Australia", with 23 company-operated stores (Starbucks Corporation, 2008). For Huawei, Xinhua gives the capital figure for Huawei Technologies Thailand but not the ownership split (Xinhua, 2020), so I place it in no column.

The library's [foreign market entry modes](https://staging.khoi.co.uk/en/framework/market-entry-modes) framework rates a wholly owned subsidiary high on control, resource commitment and capital risk, as a textbook tendency and not a measurement. The full table is in the post on [choosing an entry mode](https://staging.khoi.co.uk/en/blog/phuong-thuc-tham-nhap-thi-truong), and an acquisition that went wrong is in the [Walmart in Germany](https://staging.khoi.co.uk/en/blog/walmart-o-duc-thich-nghi-sai) post. This one takes a different line: same column, three different bills.

> **Figure:** (compare) The author's synthesis of Xinhua (2019; 2020), Huawei (2017), Starbucks (2008) and BBC (2008). Figures are as the sources give them.

```json
{
  "type": "compare",
  "columns": ["Tesla, Shanghai", "Huawei, Thailand", "Starbucks, Australia"],
  "rows": [
    { "label": "Entry, per the source", "cells": ["Wholly owned subsidiary, new plant", "Local subsidiary investing in data centres; ownership split not given", "Wholly owned operations, company-operated shops"] },
    { "label": "Size the source gives", "cells": ["Over CNY 50bn (USD 7bn), 500,000 cars a year", "THB 700m (USD 23m) for the third centre", "87 company-operated shops on 30 September 2007"] },
    { "label": "Steps before", "cells": ["Selling cars, charging network, R&D centre in Beijing (October 2018)", "Data centres in 2018 and 2019; OpenLab Bangkok existed by December 2017", "Eight years selling in Australia (BBC)"] },
    { "label": "What the source says about the outcome", "cells": ["Only the groundbreaking report", "Only the announcement of a plan", "61 shops closed, 23 left; restructuring charge USD 16.9m"] }
  ]
}
```

## Tesla bet on one plant: how much, and what came first?

More than CNY 50bn, or USD 7bn, according to Xinhua, the largest foreign-invested manufacturing project in Shanghai's history. The plant's capacity is 500,000 cars a year, while Tesla delivered 245,240 cars worldwide in all of 2018 (Xinhua, 2019). By my arithmetic the plant alone could build a little over twice the company's whole 2018 volume. Capacity is not sales.

That bet was not the first step. The Model 3 was already on sale in China at CNY 500,000 to 560,000 (about USD 72,980 to 81,738). Over six years Tesla had built more than 1,500 fast-charging stations and 1,850 destination-charging points, and in October 2018 it opened its first overseas R&D centre, in Beijing (Xinhua, 2019). So the USD 7bn step came after six years of building a charging network. The report does not say Tesla saw it as a sequence of growing commitments; that reading is mine.

The menu was also drawn by law. In the same report BMW had announced it would raise its stake in BMW Brilliance, its Shenyang joint venture, from 50 to 75 per cent and extend the partnership to 2040, and the approval rules for car joint ventures were due to switch to simple filing from 10 January (Xinhua, 2019). At the moment Tesla could enter with 100 per cent of the equity, the joint-venture route was still open and getting easier.

This size of commitment suits you if you already have customers there and can afford to commit once. It does not suit you if you do not yet know what customers will pay, because capacity built ahead of sales is fixed cost. That is my inference. The source stops at groundbreaking day; see also the [Tesla in Shanghai case](https://staging.khoi.co.uk/en/case/tesla-shanghai-gigafactory).

## Huawei in Thailand: committing one data centre at a time

The latest step is THB 700m, about USD 23m, for a third data centre in 2021. Xinhua says this equals the total investment in the first (Eastern Economic Corridor, 2018) and the second (Bangkok, 2019) (Xinhua, 2020). Against Tesla's USD 7bn it is roughly one three-hundredth.

Huawei ties this step to the host country's agenda. Its Thailand chief executive, Abel Deng, cited Thailand 4.0, the 5G licence auctions in February and an expectation that the digital economy will contribute 30 per cent of GDP by 2030 (Xinhua, 2020). That is a company executive speaking at the company's own press conference, not an independent assessment.

The same stepwise pattern appears in Egypt. According to Huawei's December 2017 release, the Cairo OpenLab had run as an R&D unit for about a year before opening to customers, at roughly 400 square metres. The release lists Bangkok among seven OpenLabs that predate Cairo (Huawei, 2017), which puts it before the first Thai data centre. The [Uppsala model](https://staging.khoi.co.uk/en/framework/uppsala-model) centres on gradually gained knowledge of foreign markets and increasing commitment (Johanson and Vahlne, 1977). A lab followed by data centres one by one fits that description, though no source says Huawei meant to follow the model.

The gain is that each step is small enough to stop. The price is speed, and each step still carries its own fixed costs. See also the cases on the [Thai data centre](https://staging.khoi.co.uk/en/case/huawei-thailand-data-centre) and the [Cairo OpenLab](https://staging.khoi.co.uk/en/case/huawei-egypt-openlab).

## Starbucks in Australia: 87 shops down to 23, and the cause stays open

The company closed 61 company-operated shops in the fourth quarter of fiscal 2008. The store table in its report shows 87 shops in Australia on 30 September 2007 and 23 on 28 September 2008, a net fall of 64, more than 61; I found no explanation of the gap in the report (Starbucks Corporation, 2008). BBC, on 4 August 2008, counted 61 of 85 shops closed after eight years (BBC News, 2008). The two counts fall on different dates, so I cannot reconcile 85 with 87.

What does leaving a company-operated network cost? The restructuring charge was USD 16.9m: USD 1.5m of asset impairment, USD 11.6m of lease exit costs and USD 3.8m of severance (Starbucks Corporation, 2008). Leases were about seven-tenths of it. That is an accounting charge booked at closure, not everything the company lost in Australia over eight years. The same report shows another route: 43 licensed shops in New Zealand, 814 in Japan, and Australia not named in the licensed list. Starbucks writes that licensing partners give access to good sites and market knowledge (Starbucks Corporation, 2008). No source says licensing would have saved Australia.

The causes, as the sources give them:

| Who says it | What they say | Kind of evidence |
|:--|:--|:--|
| Michael Edwardson, consumer psychologist (BBC) | The chain was "maybe too standardised" | One expert's opinion |
| John Roberts, University of New South Wales (BBC) | It wanted to be the global, local store, a positioning he calls "slightly schizophrenic" | One university-based voice |
| Unnamed "analysts" (BBC) | Expanded too quickly, too much debt | No figures in the article |
| Starbucks management (BBC) | Underperforming shops; refocus on three big cities | The company's own account |
| Fiscal 2008 report | Closed to address difficulties specific to its Australia market | No further detail |

The first two comments use the two axes of the [global integration and local responsiveness grid](https://staging.khoi.co.uk/en/framework/integration-responsiveness-grid) but do not reach the same conclusion: one says too standardised, the other says it wanted to be global and local at once. I do not pick between the explanations, because nothing in the sources measures any cause. See also the [Starbucks exit from Australia case](https://staging.khoi.co.uk/en/case/starbucks-australia).

## How big should your first step be? Three questions from three cases

Ask three questions before signing a first commitment. They are my inference from the three cases and have not been checked against data on Vietnamese firms.

| Question | Case it comes from | What to do |
|:--|:--|:--|
| Which ways in does the law give you? | Tesla and BMW in China, 2019 | Ask a local lawyer about foreign ownership limits in your sector |
| What is the smallest step you can lose without hurting home? | Huawei: a lab, then data centres one by one | Split the plan into steps and write down the conditions to stop or go on |
| What keeps running after you stop? | Starbucks: lease exit was USD 11.6m of 16.9m | Negotiate lease length and exit clauses before opening |

An invented example, made up by me: a Ho Chi Minh City chain of 20 bubble-tea shops wants to enter Sydney. Opening 10 shops at once means 10 leases, each a cost that keeps running if you stop. Opening 2 on short leases and deciding after a set number of months, fixed in advance, buys information before it buys premises. The second suits you if you do not yet know what customers there will pay; it does not suit you if the good sites are about to be taken.

## Going deeper: what did I read, and what did I not?

I read in full the Xinhua report on Tesla (via the Pudong government's English site), the Xinhua report on Huawei (via China Daily), the 2008 BBC article and Huawei's 2017 press release. For Starbucks' "fiscal 2008 annual report" in the library, I read the Form 10-K for the year ended 28 September 2008, filed with the US Securities and Exchange Commission on 24 November 2008 ([full text on EDGAR](https://www.sec.gov/Archives/edgar/data/829224/000095013408021134/v50316e10vk.htm)). I did not compare it with the shareholder report, and I give no page numbers because it is an HTML file.

For Johanson and Vahlne (1977) I read only the abstract. The two frameworks are the library's summaries; I did not read the books they cite.

Three points I am less sure of. First, Xinhua's sentence about Huawei can be read as "equal to the first two combined" (my reading) or as "equal to each". Second, "a little over twice" and "one three-hundredth" are my own arithmetic from figures in the reports. Third, Starbucks also had a plan to close about 600 US stores that year (Starbucks Corporation, 2008), so I cannot separate general effects from those specific to Australia.

## Frequently asked questions

### How does a wholly owned subsidiary differ from a joint venture, and which was Tesla Shanghai?

In this site's glossary, a joint venture is a new organisation two sides set up together to share cost, risk and expertise, while a wholly owned subsidiary is a foreign unit the parent owns outright: maximum control, but it needs a lot of capital. Tesla's Shanghai plant is the second kind. Xinhua says Tesla was the first to benefit from a new policy allowing foreign carmakers to set up wholly-owned subsidiaries in China, and the same report says BMW had announced it would raise its stake in the BMW Brilliance joint venture from 50 to 75 per cent (Xinhua, 2019).

### Why did Starbucks close 61 shops in Australia?

The sources I read do not give one cause. BBC quotes a consumer psychologist saying the chain was "maybe too standardised", a University of New South Wales voice who says it wanted to be "the global, local store", and analysts who say it expanded too quickly and took on too much debt (BBC News, 2008). Starbucks itself says it closed the shops to address "the difficulties specific to its Australia market" (Starbucks Corporation, 2008). Nobody measured any of these, so treat them as hypotheses.

### Did Starbucks leave Australia altogether?

Not in 2008. Its fiscal 2008 report says it kept 23 company-operated stores there on 28 September 2008, and BBC says it was refocusing on Sydney, Melbourne and Brisbane (Starbucks Corporation, 2008; BBC News, 2008). I have no source on what happened afterwards, so the post says nothing about today.

### Is investing in steps always safer than one big commitment?

No. The Uppsala model describes firms gaining knowledge of a foreign market gradually and raising their commitment in small increments (Johanson and Vahlne, 1977). Tesla put more than CNY 50bn into a single plant, though it had spent six years building a charging network in China (Xinhua, 2019). Steps keep the loss small but are slower; one big step is faster and costs more if it is wrong. The sources do not compare the results of the two.

## References

- BBC News (2008) *Shunned Starbucks in Aussie exit (4 August 2008)*. BBC News. Available at: http://news.bbc.co.uk/2/hi/7540480.stm.
- Huawei (2017) *Huawei announces new OpenLab in Cairo to build ICT ecosystem in Northern Africa (11 December 2017)*. Available at: https://www.huawei.com/en/news/2017/12/huawei-new-openlab-cairo.
- Johanson, J. and Vahlne, J.-E. (1977) ‘The internationalization process of the firm: a model of knowledge development and increasing foreign market commitments’, *Journal of International Business Studies*, 8(1), pp. 23–32. https://doi.org/10.1057/palgrave.jibs.8490676
- Starbucks Corporation (2008) *Fiscal 2008 annual report*. Seattle: Starbucks Corporation.
- Xinhua (2019) *Report on the groundbreaking of Tesla's Shanghai factory (7 January 2019)*. Shanghai Pudong government English site. Available at: https://english.pudong.gov.cn/2019-01/08/c_314114.htm.
- Xinhua (2020) *Report on Huawei Technologies Thailand's plan for a third data centre (12 November 2020)*. China Daily. Available at: http://global.chinadaily.com.cn/a/202011/12/WS5facc67aa31024ad0ba93ae9.html.

---

Canonical HTML version: https://staging.khoi.co.uk/en/blog/tesla-huawei-starbucks-ba-cach-vao