Taiwan Has Plenty of IPOs. So Why Do So Few Startups Go Public?
台灣明明不缺 IPO，為什麼真正從新創走到上市的公司這麼少？
- 前半段為文章的英文版本 (The first half is the English version)
- 後半段為中文版本 (The second half is the Mandarin version)
====
Taiwan Has Plenty of IPOs. So Why Do So Few Startups Go Public?
Taiwan has thousands of startups, an active venture capital ecosystem, sophisticated public markets, and one of the deepest technology supply chains in the world. Yet new data reveals a surprising disconnect between the companies Taiwan creates, the companies venture capital funds, and the companies that eventually go public.
The Taiwan startup IPO pipeline presents a paradox.
Taiwan should be very good at producing startup IPOs.
It has almost every ingredient one might ask for: world-class engineers, deep pools of technical talent, sophisticated capital markets, experienced institutional investors, thousands of small and medium-sized technology companies, a growing venture capital ecosystem, and one of the most complete technology supply chains anywhere in the world.
Taiwan also has no shortage of public companies.
Between January 2016 and September 2026, 520 companies made their first public-market debut in Taiwan, according to a large-scale study by AVA Angels.
But look more closely at those 520 companies, and a very different picture emerges.
AVA compared a database of 13,227 Taiwanese startups with those 520 newly listed companies. After distinguishing independent startups from corporate restructurings, group spin-offs, and other businesses that did not genuinely originate as independent startups, the researchers identified just 44 companies that could clearly be classified as startup-origin IPOs.
That is only 8.5% of the 520 companies that went public during the period.
This creates a fascinating contradiction.
Taiwan has startups.
Taiwan has venture capital.
Taiwan has IPOs.
Yet surprisingly few companies appear to travel through the entire chain:
Startup → Venture Capital → Scale-Up → IPO
Understanding this gap is essential to understanding the Taiwan startup IPO market itself.
The answer may tell us something much deeper about how Taiwan actually builds companies—and why its traditional model of corporate growth does not always connect naturally with the venture-backed startup model.
1. The Taiwan Startup IPO Paradox: 520 IPOs, but Only 44 Startup Stories
The AVA Angels study begins with an unusually large dataset.
Its researchers assembled a universe of 13,227 Taiwanese startups from 26 categories of public sources, then compared that population with 520 companies that made their first public-market listing between January 2016 and September 2026.
The IPO universe included 210 companies listed on the Taiwan Stock Exchange’s main board, 274 on the Taipei Exchange, and 36 on the Taiwan Innovation Board.
The researchers then went further.
Using company prospectuses, they reconstructed the capital formation histories of the IPO companies, examining capital increases, financing events, founding dates, and pre-IPO financing.
That exercise produced one immediately striking number:
18.4 years.
That was the median period between company formation and public listing.
The average was even longer: 20.6 years.
Only 21.2% of companies went public within their first ten years. Another 46.0% had existed for more than twenty years before listing.
At first glance, the obvious conclusion might be that Taiwanese startups take an extraordinarily long time to IPO.
But that would be the wrong interpretation.
The 520 companies were not all startups.
Many were mature manufacturers, family businesses, established SMEs, corporate spin-offs, or businesses that had spent decades growing before eventually entering the public market.
So the more interesting conclusion is not that Taiwanese startups require 18 years to reach an IPO.
It is this:
The typical company entering Taiwan’s public markets historically was not a startup at all.
The differences between markets reinforce the point.
Companies entering the TWSE main board had a median age of 21.5 years. Those entering the Taipei Exchange had a median age of 16.8 years.
But companies entering the Taiwan Innovation Board had a median age of just 10.1 years.
The Innovation Board represented only 36 of the 520 listings in the study, so the sample remains much smaller. But the difference is large enough to raise an important question:
Is Taiwan beginning to create a different IPO pathway for younger, venture-style companies?
We will return to that later.
Source: AVA Angels analysis of 520 first-time public listings in Taiwan, January 2016–September 2026.
2. Taiwan Has Two Company-Building Machines
To understand the Taiwan startup IPO gap, it helps to stop assuming that every successful technology company follows the Silicon Valley model.
Taiwan built its industrial economy differently.
For decades, one of Taiwan’s most powerful company-building pathways has looked something like this:
Engineer / Entrepreneur
↓
Small or Medium-Sized Enterprise
↓
Customers
↓
Revenue
↓
Profits
↓
Retained Earnings + Bank Financing
↓
Scale
↓
IPO
This model produced generations of specialized manufacturers and technology suppliers.
Many started with relatively little institutional capital. They won customers, accumulated manufacturing expertise, reinvested profits, borrowed from banks, expanded production, and gradually moved deeper into global supply chains.
Some eventually became what Germany might call Mittelstand companies and what Taiwan increasingly recognizes as its own collection of hidden champions.
This system is not a weakness.
Quite the opposite.
It may be one of Taiwan’s greatest economic inventions.
It helped create the dense industrial networks behind semiconductors, electronics, machinery, components, materials, servers, networking equipment, and countless other products that now sit inside global technology supply chains.
But this is not the same machine that built Silicon Valley.
The venture-backed model looks more like this:
Founder
↓
Startup
↓
Angel / Seed Capital
↓
Series A / B / C
↓
Product-Market Fit
↓
Scale-Up
↓
Growth Capital
↓
IPO or M&A
Taiwan has this second machine too.
The question is whether it operates with the same throughput as the first.
The data suggests that it does not.
This distinction matters because comparing the Taiwan startup ecosystem directly with Silicon Valley without accounting for these different company-building mechanisms risks mistaking a structural difference for simple ecosystem weakness.
Taiwan may not have one broken company-building machine.
It may have two different machines—one highly mature, the other still developing its connections to growth capital and public markets.
3. Taiwan Starts One Kind of Company, Funds Another, and Lists Another
Perhaps the most revealing part of the AVA dataset appears when Taiwan’s startups, venture investments, and IPO companies are classified using the same industry categories.
The structures barely resemble one another.
Among Taiwan’s startup population, the largest category is consumer and lifestyle services, representing 34.8% of companies.
Among venture capital investments in the study, however, the largest category is biotechnology and healthcare, accounting for 44.2%.
And among companies that ultimately went public, the largest category is electronics, hardware, and semiconductors, representing 36.5%.
In other words:
Taiwan starts one kind of company, funds another, and lists yet another.
The statistical relationship is similarly weak.
When the industry distribution of Taiwan’s startup population is compared with that of the IPO population, AVA calculates a correlation of -0.21.
When the venture capital portfolio is compared with the overall IPO population, the correlation is just +0.06.
The three populations therefore show little industry-level alignment.
That matters because startup ecosystems are often discussed as though they were linear pipelines.
Create more startups.
Fund the strongest ones.
Help them scale.
Take the winners public.
But Taiwan’s numbers suggest something different.
Startup formation, Taiwan venture capital, and public-market listings may not simply be three consecutive stages of one system.
Historically, they have often behaved like three partially separate systems.
And that separation may help explain why the Taiwan startup IPO pipeline looks so different from the country’s broader IPO market.
Source: AVA Angels.
4. The 44-Company Paradox
Then something unexpected happens.
AVA separated the 520 IPO companies and identified 44 that could clearly be described as independent startup-origin companies rather than corporate spin-offs, restructurings, or established businesses that later went public.
When the industry composition of those 44 companies was compared with the venture capital portfolio, the correlation jumped to:
+0.88.
That is dramatically different from the +0.06 correlation between venture capital and the entire IPO population.
This may be the most important finding in the dataset.
Because it suggests that Taiwan’s venture investors are not necessarily pointing in the wrong direction.
Once traditional companies are removed from the comparison, the companies successfully completing the startup-to-IPO journey look remarkably similar, at the industry level, to the sectors receiving venture investment.
The venture pipeline, in other words, appears to make much more sense when examined on its own terms.
The problem may therefore not primarily be that Taiwan funds the wrong startups.
The problem may be that too few companies complete the journey.
There is an important methodological caution here.
Correlation across industry categories does not demonstrate that venture capital caused those 44 companies to reach the public market. Nor does it tell us which specific financing, management, or market-development factors determined success.
What it does show is something narrower but still important:
the sector composition of venture investment is highly aligned with the sector composition of startup-origin companies that eventually listed in Taiwan.
That distinction fundamentally changes the diagnosis.
If venture investors were consistently funding sectors that never appeared among successful startup listings, Taiwan might have an obvious capital-allocation problem.
But the +0.88 correlation points toward another possibility:
the direction may be broadly right, while the pathway remains narrow.
5. The Missing Middle: Does Taiwan Have a Startup Scale-Up Gap?
Where, then, does the Taiwan startup IPO pipeline narrow?
The available data cannot prove a single causal explanation.
But it raises an important hypothesis worth investigating:
Taiwan may have a scale-up gap.
Consider what the country already has.
Taiwan can create startups.
It has universities, incubators, accelerators, government programs, angel investors, corporate innovation programs, and venture capital firms.
Taiwan also has sophisticated public markets.
The TWSE and Taipei Exchange together support a large ecosystem of listed companies, institutional investors, securities firms, analysts, accountants, and investment banks.
The difficult transition may lie in the middle:
Startup
↓
Product-Market Fit
↓
Early Revenue
↓
Scale-Up
↓
Growth Capital
↓
International Expansion
↓
IPO Readiness
↓
Public Market
Moving from a promising startup to a credible public-market candidate requires much more than achieving product-market fit.
A company may need to move from NT$100 million in revenue toward NT$1 billion and eventually several billion.
It may need to build overseas sales organizations, recruit experienced executives, establish financial controls, professionalize governance, develop predictable reporting processes, survive multiple financing rounds, and convince institutional investors that its growth can continue at much greater scale.
For hardware startups, scaling may also require substantial manufacturing capital.
For software companies, the challenge may instead involve international customer acquisition, recurring-revenue scale, and access to investors willing to value growth ahead of near-term profitability.
Taiwan is exceptionally good at helping companies scale through supply chains and customer relationships.
The venture model often asks companies to scale through another mechanism as well: repeated injections of risk capital combined with rapid market expansion.
Those capabilities overlap, but they are not identical.
This is why the 44-company figure must also be treated carefully.
Dividing 44 by 13,227 produces a tiny percentage, but it would be methodologically incorrect to describe that number as Taiwan’s startup IPO success rate.
The startup universe spans different founding cohorts. Many companies are too young to have reached an exit. Others may have been acquired, remained private, closed, or pursued listings outside Taiwan.
The more defensible observation is simpler:
Only 44 of the 520 companies going public in Taiwan during the study period were clearly identifiable as independent startup-origin companies.
That tells us something about the composition of Taiwan’s IPO pipeline.
It does not by itself tell us the lifetime probability that a Taiwanese startup will eventually go public.
The next research question should therefore be what happens between startup formation and IPO readiness.
Note: 44 / 13,227 should not be interpreted as an IPO conversion rate because the two populations contain different cohorts.
6. Taiwan Startup IPOs Have a Global Geography
There is another complication.
Looking only at Taiwanese stock exchanges does not capture every successful Taiwanese startup that reaches a public market.
Some of Taiwan’s most recognizable technology startups chose different venues.
Consider four examples.
91APP, founded in 2013, became Taiwan’s first publicly listed SaaS company when its shares began trading on the Taipei Exchange on May 25, 2021.
That is an important counterexample to any claim that Taiwan’s domestic capital markets cannot accommodate software companies.
But Appier chose another route.
Appier was founded in Taipei in 2012. It expanded internationally, raised institutional venture capital—including an US$80 million Series D—and completed its listing on the Tokyo Stock Exchange’s Mothers market in March 2021. It subsequently moved to the Tokyo Stock Exchange’s Prime Market in December 2022.
Then there is Gogoro.
After developing Taiwan’s battery-swapping ecosystem and expanding internationally, Gogoro completed its business combination with Poema Global in April 2022. Its shares began trading on the Nasdaq Global Select Market under the ticker GGR on April 5, 2022.
Perfect Corp., the Taiwan-founded AI and augmented-reality beauty technology company, followed another overseas route. Following its business combination with Provident Acquisition Corp., Perfect began trading on the New York Stock Exchange on October 31, 2022 under the ticker PERF.
These cases reveal an important distinction.
There are really two questions:
How many Taiwanese startups go public in Taiwan?
And:
How many Taiwanese startups eventually become public companies somewhere in the world?
Those are not the same question.
The AVA study is particularly useful for answering the first.
A comprehensive answer to the second would require a broader dataset that tracks Taiwanese startups across overseas exchanges, acquisitions, and other exit routes.
Still, the examples above reveal why that broader question matters.
For globally oriented software, AI, platform, and consumer-technology companies, choosing an IPO venue can itself become part of corporate strategy.
Different markets may offer different investor bases, valuation frameworks, analyst ecosystems, liquidity profiles, and levels of familiarity with particular business models.
A Taiwanese semiconductor equipment company may be naturally understood by domestic investors who have spent decades analyzing semiconductor supply chains.
An AI SaaS company serving customers across Japan, Southeast Asia, Europe, and the United States may face a different set of considerations.
Four examples are not enough to establish a general rule about where Taiwanese software startups should list.
But they are enough to demonstrate that:
Going public and going public in Taiwan are not necessarily the same decision.
The geography of Taiwan startup IPOs can extend beyond Taiwan itself.
7. Can the Taiwan Innovation Board Fix the Startup IPO Gap?
Taiwan’s capital-market institutions have increasingly tried to address this mismatch.
The Taiwan Innovation Board, or TIB, was established in 2021 as an alternative public-market pathway for companies with key technologies, innovative capabilities, or innovative business models.
Its architecture differs substantially from Taiwan’s traditional listing framework.
Under the TIB’s current Criterion I, an applicant can qualify with:
- a market capitalization of at least NT$1 billion;
- aggregate operating revenue over the latest four quarters of more than NT$100 million; and
- sufficient working capital covering the twelve months following listing.
Separate criteria exist for biotechnology and medical companies and for larger innovative companies.
Most importantly, the TIB uses a market-capitalization-centered framework and does not require profitability as a condition for IPO eligibility.
That distinction matters for the Taiwan startup IPO market.
The conventional public-market model works naturally for the type of company Taiwan historically produced: mature businesses with established revenue, profits, manufacturing capacity, and long operating histories.
Venture-backed growth companies can look very different.
They may be younger.
They may reinvest aggressively.
They may prioritize market expansion over current earnings.
Their value may lie in intellectual property, software, data, platforms, or technologies whose commercial scale has not yet fully materialized.
Trying to evaluate those businesses entirely through a framework designed around mature industrial companies risks creating a structural mismatch.
Taiwan has continued to modify the TIB in response.
On January 6, 2025, restrictions limiting TIB trading to qualified investors were removed, opening the market to the broader investing public, subject to applicable risk-disclosure requirements.
By 2025, average daily TIB trading value had increased to approximately NT$277 million, compared with NT$103 million in 2024, while 15 companies applied for TIB listings during 2025—a record since the board’s establishment.
The policy ambition has also become more explicit.
In October 2025, Taiwan launched its broader Asia Innovation Capital initiative. TWSE describes the TIB as a key pillar in building a locally distinctive “Asian NASDAQ”, emphasizing its market-capitalization-based listing framework and its role in helping innovative companies access public capital earlier.
And in June 2026, TWSE again described the TIB as an important bridge connecting innovative enterprises with capital markets and as part of Taiwan’s ambition to become a global hub for technology capital.
That brings us back to one of the most interesting numbers in the AVA dataset:
TWSE Main Board: median 21.5 years to listing
Taipei Exchange: 16.8 years
Taiwan Innovation Board: 10.1 years
The TIB sample in the study remains small.
A younger median listing age does not prove that the board will become Taiwan’s answer to Nasdaq, nor does higher trading activity guarantee a self-sustaining growth-company market.
Successful public markets require more than relaxed listing standards. They need liquidity, institutional participation, analyst coverage, credible governance, high-quality issuers, and enough successful companies to create a reinforcing ecosystem.
But the early evidence does suggest something more modest—and perhaps more important:
The TIB is already attracting a meaningfully younger population of companies than Taiwan’s traditional public markets.
That makes it more than another listing board.
It makes it an experiment.
Can Taiwan create a domestic public market capable of serving companies that previously might have remained private longer—or looked toward Tokyo, New York, or other overseas markets?
We do not know yet.
But the experiment is now underway.
8. The Future of the Taiwan Startup IPO Market: Build a Better Bridge
For years, startup policy has often been measured using easily observable numbers.
How many startups were created?
How much venture capital was invested?
How many companies went public?
All are useful indicators.
But the AVA dataset raises an uncomfortable possibility:
Those numbers may be measuring three different systems.
Taiwan already creates startups.
Taiwan already has venture capital.
Taiwan already has active public markets.
What it may historically have lacked is a sufficiently wide bridge connecting them.
That changes how we should think about the Taiwan startup ecosystem.
The goal should not simply be to maximize startup formation.
Creating thousands of additional early-stage companies does little if promising companies struggle to cross the scale-up stage.
Nor should the goal simply be increasing venture capital volume.
More capital does not automatically solve problems involving international expansion, management depth, governance, customer acquisition, or public-market readiness.
And increasing the raw number of IPOs tells us little about whether the venture ecosystem itself is producing scalable companies.
Perhaps a better question is:
How many companies successfully cross the entire bridge?
Startup formation.
Venture financing.
Product-market fit.
Scale-up.
Growth capital.
International expansion.
Public-market readiness.
Exit.
Viewed this way, Taiwan’s challenge may not primarily be about creating more pieces.
It may be about strengthening the connective tissue between pieces that already exist.
And there is reason for cautious optimism.
The most encouraging number in the AVA research may not be 13,227 startups, 520 IPOs, or even the surprisingly small group of 44 startup-origin listings.
It may be +0.88.
Because among startup-origin companies that did reach Taiwan’s public market, their industry profile aligns remarkably closely with the industry profile of venture investment.
That does not prove that the entire mechanism works efficiently.
But it does suggest that the venture ecosystem and its successful public-market outcomes are not fundamentally pointing in opposite directions.
Taiwan’s traditional company-building machine took decades to develop.
It was built through industrial parks, engineering education, supplier networks, banks, export markets, multinational customers, manufacturing know-how, and generations of entrepreneurs.
The result became one of the world’s most formidable industrial ecosystems.
The venture-backed machine is younger.
Its institutions are still evolving.
Its growth-capital networks are still developing.
Its relationship with public markets is still being redesigned.
And the Taiwan Innovation Board represents one attempt to connect those worlds.
The next chapter of Taiwan’s startup story therefore may not depend on creating another accelerator, launching another startup program, or producing another headline about the number of newly founded companies.
It may depend on something less glamorous but far more consequential:
building the infrastructure that helps promising companies survive the journey from startup to scale-up.
Taiwan already knows how to build world-class companies.
The next question is whether it can build a wider bridge for a new generation of them.
If you found this analysis useful
I regularly share deeper, behind-the-scenes insights on my LinkedIn Newsletter:
Taiwan Tech Dispatch
→ Follow & subscribe on LinkedIn
Further Reading
- AVA Angel (Mandarin)
https://www.angeltoventure.com/post/startups-vc-ipo - Why Taiwan Technology Companies Build World-Class Technology — But So Few Global Tech Giants
https://whitehsu.blog/2026/09/09/taiwan-technology-companies-global-tech/ - FunNow IPO Signals a New Era for Taiwan’s Software Industry
https://whitehsu.blog/2026/08/31/funnow-ipo-taiwan-software-industry/ - Taiwan AI Startups: How AI Is Rewriting the Geography of Entrepreneurship
https://whitehsu.blog/2026/08/17/taiwan-ai-startups/
====
台灣明明不缺 IPO，為什麼真正從新創走到上市的公司這麼少？
台灣擁有數以萬計的新創企業、逐漸成熟的創投生態系、活躍的資本市場，以及全球最完整的科技供應鏈之一。然而，一份涵蓋 13,227 家新創與 520 家掛牌公司的大型研究，卻揭露了一個令人意外的現象：台灣所創造的新創、創投所投資的企業，以及最後真正進入資本市場的公司，似乎分屬三個不同的世界。
台灣的新創 IPO（Taiwan Startup IPO）存在一個很有意思的矛盾。
照理來說，台灣應該非常擅長把新創公司送進資本市場。
我們擁有世界級的工程師、大量科技人才、成熟的資本市場、專業的機構投資人、數量龐大的科技中小企業、逐漸成熟的創投生態系，以及全球最完整的科技供應鏈之一。
而且，台灣從來不缺上市櫃公司。
根據 AVA Angels 的一項大型研究，從 2016 年 1 月到 2026 年 9 月，共有 520 家公司首次進入台灣公開資本市場。
但如果仔細拆解這 520 家公司，會看到完全不同的故事。
AVA 將 13,227 家台灣新創企業與這 520 家新掛牌公司進行比對，並進一步排除企業重組、集團分拆，以及其他並非真正從獨立新創一路發展而來的公司。
最後，他們只辨識出：
44 家可以明確認定為「獨立新創出身」的 IPO 公司。
換句話說，只占這 520 家掛牌公司的 8.5%。
這就產生了一個非常有意思的矛盾。
台灣有新創。
台灣有創投。
台灣也有 IPO。
但真正完整走過下面這條路的公司，似乎並不多：
Startup → Venture Capital → Scale-Up → IPO
為什麼？
理解這個斷層，不只是理解台灣新創 IPO 市場的關鍵。
它可能還揭露了一個更深層的問題：
台灣究竟是如何「長出一家公司」的？
1. 台灣新創 IPO 的矛盾：520 家 IPO，卻只有 44 個真正的新創故事
AVA Angels 的研究建立在一個相當龐大的資料庫上。
研究團隊從 26 類公開來源整理出 13,227 家台灣新創企業，再與 2016 年 1 月至 2026 年 9 月間首次掛牌的 520 家公司進行比較。
這 520 家公司包括：
210 家台灣證券交易所一般板上市公司、
274 家上櫃公司，
以及 36 家台灣創新板公司。
研究團隊還進一步利用公開說明書，回推這些公司的資本形成歷史，包括增資、融資事件、成立時間，以及 Pre-IPO 階段的資金變化。
結果出現了一個非常驚人的數字：
18.4 年。
這是這 520 家公司從成立到進入公開市場的中位數時間。
平均值甚至更長：
20.6 年。
只有 21.2% 的公司在成立十年內掛牌。
而高達 46.0% 的公司，在成立超過二十年之後才進入公開市場。
乍看之下，我們很容易得到一個結論：
「台灣新創要花 18 年才能 IPO。」
但這其實是錯誤的解讀。
因為這 520 家公司根本不全都是新創。
其中包含大量成熟製造商、家族企業、中小企業、集團分拆公司，以及經營數十年後才決定上市櫃的企業。
所以，18.4 年真正告訴我們的，不是「台灣新創成長得太慢」。
而是：
過去進入台灣公開資本市場的典型公司，本來就不是 Startup。
不同市場之間的差距更能說明這件事情。
進入台灣證券交易所一般板的公司，成立至上市的中位數為：
21.5 年。
上櫃公司：
16.8 年。
但台灣創新板呢？
只有：
10.1 年。
當然，AVA 研究中的創新板樣本只有 36 家，遠少於其他市場，因此必須謹慎解讀。
但這個差距仍然值得注意。
它讓我們開始思考另一個問題：
台灣是不是正在建立一條讓更年輕、更加接近 Venture-style company 的企業提早進入資本市場的新道路？
後面我們會再回到這個問題。
資料來源：AVA Angels，2016 年 1 月至 2026 年 9 月 520 家首次掛牌公司的分析。
2. 台灣其實有兩套「企業製造機器」
如果想理解為什麼台灣的新創 IPO 數據如此特殊，我們必須先放下一個常見假設：
不是所有成功的科技公司，都必須按照矽谷模式成長。
台灣建立工業經濟的方式，本來就不一樣。
過去數十年，台灣最成功的一條企業成長路徑，大概長這樣：
工程師／創業者
↓
中小企業
↓
客戶
↓
營收
↓
獲利
↓
保留盈餘＋銀行融資
↓
擴大規模
↓
IPO
這套模式孕育了數以千計的專業製造商與科技供應商。
許多公司創立時並沒有大量機構資金。
它們先找到客戶。
累積製造能力。
把賺到的錢重新投入公司。
向銀行借款。
擴充產能。
然後逐漸進入全球供應鏈更核心的位置。
其中一些企業最後成為類似德國 Mittelstand 的公司，也就是我們今天愈來愈熟悉的：
台灣隱形冠軍（Hidden Champions）。
這套系統並不是台灣的弱點。
恰恰相反。
它可能是台灣過去數十年最成功的經濟制度創新之一。
半導體、電子零組件、機械、材料、伺服器、網通設備，以及無數全球科技產品背後的供應鏈，就是在這樣的企業形成模式下逐漸長出來的。
但這不是矽谷的模式。
典型的 Venture-backed model 比較像：
Founder
↓
Startup
↓
Angel / Seed
↓
Series A / B / C
↓
Product-Market Fit
↓
Scale-Up
↓
Growth Capital
↓
IPO / M&A
台灣其實也有第二套機器。
真正的問題是：
第二套機器的 throughput，是否跟第一套一樣強？
目前的數據顯示，答案很可能是否定的。
這個 distinction 非常重要。
如果我們直接拿台灣的新創生態系與矽谷比較，卻忽略兩地企業成長機制本來就不同，很容易把「結構差異」誤判成「台灣的新創生態系不夠好」。
更準確的描述可能是：
台灣不是只有一套失靈的企業成長機制，而是同時存在兩套企業製造機器——其中一套已經極度成熟，另一套仍在建立與成長資本及公開市場之間的連結。
3. 台灣創造一種公司、創投投另一種，最後上市的又是第三種
AVA 研究中最有意思的發現之一，是把：
新創企業、
創投投資，
以及 IPO 公司，
全部使用相同產業分類重新整理。
結果三者的結構幾乎不像同一個經濟體。
在台灣的新創母體中，最大類別是：
消費／生活服務：34.8%
但是在研究涵蓋的創投投資中，占比最高的卻是：
生技／醫療：44.2%
到了最後真正 IPO 的公司，最大類別又變成：
電子／硬體／半導體：36.5%
換句話說：
台灣創造一種公司，創投投資另一種公司，最後進入資本市場的又是第三種公司。
統計關係同樣非常弱。
台灣新創產業分布與 IPO 產業分布之間的相關係數：
-0.21
創投投資組合與全部 IPO 公司之間：
+0.06
幾乎看不到明顯的產業結構連動。
這非常重要。
因為我們通常會把「新創生態系」想像成一條流水線：
創造更多 Startup。
↓
挑出最好的公司投資。
↓
幫助它們成長。
↓
最後 IPO。
但台灣的數據顯示，現實可能完全不同。
Startup formation、Venture Capital 與 Public Market Listings，歷史上未必是同一套系統的三個連續階段。
它們更像是：
三個部分相互連結、但又各自運作的世界。
這可能正是為什麼台灣的新創 IPO pipeline，和整體 IPO 市場看起來如此不同。
資料來源：AVA Angels。
4. 44 家公司的悖論
但接下來，出現了一個非常意外的結果。
AVA 從 520 家 IPO 公司中，進一步辨識出 44 家真正可以被視為「獨立新創一路成長至 IPO」的企業。
接著，他們把這 44 家公司的產業分布與創投投資組合進行比較。
結果相關係數突然變成：
+0.88
這與創投投資和全部 IPO 公司之間只有 +0.06 的相關性，形成巨大反差。
這可能是整份研究中最重要的一個數字。
因為它意味著：
台灣創投投資的方向，未必真的錯了。
當我們把傳統企業、成熟中小企業、集團分拆等公司從 IPO population 中移除，只觀察真正從 Startup 一路走到 IPO 的企業時，它們的產業分布，竟然與創投資金配置高度一致。
換句話說：
Venture pipeline 本身可能比我們原本以為的合理得多。
真正的問題或許不是：
「台灣 VC 投錯公司。」
而是：
能夠完整走完這條路的公司太少。
當然，這裡必須非常謹慎。
產業分類之間的相關係數，不能證明創投資金「造成」這 44 家公司成功 IPO。
它也無法告訴我們究竟是哪一輪融資、哪一種管理能力，或哪一個市場因素讓這些公司成功。
但它至少證明了一件比較窄、卻非常重要的事情：
台灣創投的產業配置，與最後成功進入國內公開市場的新創企業，其產業結構高度一致。
因此，問題可能不是方向錯了。
而是：
路太窄了。
5. 消失的中間層：台灣是否存在 Scale-Up Gap？
那麼，台灣的新創 IPO pipeline 到底在哪裡開始變窄？
目前的資料還不足以證明單一因果關係。
但它提出了一個非常值得進一步驗證的假設：
台灣可能存在 Scale-Up Gap。
因為台灣並不缺新創。
我們有大學、育成中心、Accelerator、政府計畫、天使投資人、企業創新計畫與創投基金。
台灣也不缺資本市場。
TWSE 與 TPEx 背後有成熟的證券商、機構投資人、分析師、會計師、投資銀行，以及大量上市櫃企業。
真正困難的地方，可能在中間：
Startup
↓
Product-Market Fit
↓
Early Revenue
↓
Scale-Up
↓
Growth Capital
↓
International Expansion
↓
IPO Readiness
↓
Public Market
從一間「不錯的新創公司」成長為「可以進入公開市場的企業」，需要的遠遠不只是找到 Product-Market Fit。
一家公司的營收可能必須從新台幣 1 億元，往 10 億元，甚至數十億元前進。
它可能必須建立海外業務團隊、延攬更成熟的管理人才、建立內控制度、強化公司治理、建立可預測的財務報告流程，經歷多輪融資，並且讓機構投資人相信公司能夠持續擴張。
硬體新創還可能面臨龐大的製造資本需求。
軟體公司則可能面臨完全不同的問題：
國際客戶取得、Recurring Revenue 規模，以及能否找到願意在短期獲利之前先相信成長的投資人。
台灣非常擅長透過：
供應鏈＋客戶關係
幫助企業成長。
但 Venture model 還要求另一種能力：
Risk Capital＋Rapid Market Expansion
兩者有所重疊，但並不完全相同。
這也是為什麼我們不能直接拿：
44 ÷ 13,227
然後說：
「台灣新創 IPO 成功率只有 0.33%。」
這樣在方法論上是不成立的。
13,227 家新創來自不同成立年份。
其中很多公司還太年輕，根本還沒有走到 Exit 階段。
另外一些可能被併購、繼續維持私人企業、停止營運，或者選擇海外上市。
因此，更嚴謹的描述是：
2016 至 2026 年間，在台灣掛牌的 520 家企業中，只有 44 家可以明確認定為獨立新創出身。
這告訴我們的是：
台灣 IPO pipeline 的組成。
而不是：
一家台灣新創終其一生成功 IPO 的機率。
真正值得研究的下一個問題是：
Startup Formation 與 IPO Readiness 之間，到底發生了什麼？
註：由於兩組資料屬於不同 cohort，因此 44 / 13,227 不應被解讀為新創 IPO 成功率。
6. 台灣新創 IPO，其實有一張全球地圖
還有另一個容易被忽略的問題。
如果我們只看台灣證券市場，就無法捕捉所有成功走向公開市場的台灣新創。
因為一些最知名的台灣科技新創，最後選擇了不同市場。
例如 91APP。
91APP 創立於 2013 年，並在 2021 年 5 月 25 日於台灣上櫃，成為台灣第一家掛牌的 SaaS 公司。
這是一個非常重要的反例。
它證明：
軟體公司並非無法在台灣資本市場掛牌。
但 Appier 選擇了另一條路。
Appier 於 2012 年在台北創立，之後快速國際化，也獲得多輪機構投資，其中包括 8,000 萬美元 Series D。
2021 年 3 月，Appier 登上東京證券交易所 Mothers 市場。
2022 年 12 月，再轉至東京證交所 Prime Market。
Gogoro 又是另一條路。
在建立台灣電池交換生態系並拓展海外市場後，Gogoro 於 2022 年完成與 Poema Global 的企業合併，並於同年 4 月在 Nasdaq Global Select Market 掛牌，股票代號 GGR。
而台灣創立的 AI／AR 美妝科技公司 Perfect Corp. 玩美移動，則在完成與 Provident Acquisition Corp. 的企業合併後，於 2022 年 10 月 31 日正式登上紐約證券交易所，股票代號 PERF。
這些案例揭露了一個非常重要的 distinction。
我們其實在問兩個不同的問題：
有多少台灣新創在「台灣」IPO？
以及：
有多少台灣新創最後在「全球任何一個市場」成為上市公司？
這不是同一個問題。
AVA 的研究非常適合回答第一個問題。
但如果要完整回答第二個問題，就必須建立更大的資料庫，追蹤台灣新創在海外交易所、M&A，以及其他 Exit route 的發展。
即使如此，上面幾個案例已經說明為什麼第二個問題值得研究。
對於全球化程度較高的軟體、AI、平台型或消費科技公司而言：
選擇在哪一個市場 IPO，本身可能就是企業策略的一部分。
不同市場可能擁有不同的：
投資人結構、
估值框架、
分析師生態系、
流動性，
以及對不同 Business Model 的理解程度。
一家台灣半導體設備公司，很容易被已經研究半導體供應鏈數十年的台灣投資人理解。
但一家客戶遍布日本、東南亞、歐洲與美國的 AI SaaS 公司，面對的可能就是完全不同的考量。
四家公司當然不足以證明「台灣軟體新創應該去哪裡上市」。
但它們至少證明：
Going public 與 Going public in Taiwan，是兩個不同的決策。
台灣新創 IPO 的地理範圍，其實可以超出台灣本身。
7. 台灣創新板，能不能補上 Startup IPO 的缺口？
台灣資本市場其實也逐漸意識到這個 mismatch。
**台灣創新板（Taiwan Innovation Board，TIB）**於 2021 年成立，目的就是替擁有關鍵技術、創新能力或創新商業模式的公司，建立另一條進入公開市場的道路。
它的制度設計與傳統上市制度有明顯不同。
按照目前創新板第一類上市條件，公司可以在符合以下要求時申請：
市值至少 新台幣 10 億元；
最近四季營業收入合計超過 新台幣 1 億元；
以及具備足以支應上市後十二個月營運所需的營運資金。
另外也有針對生技醫療公司，以及較大型創新企業的其他上市條件。
最重要的是：
創新板以市值為核心，而且 IPO eligibility 並不要求公司必須已經獲利。
這對台灣的新創 IPO 市場非常重要。
傳統資本市場制度，非常適合台灣過去最擅長培養的企業：
成熟、
有營收、
有獲利、
有產能，
而且擁有長期營運紀錄。
但 Venture-backed growth company 往往完全不同。
它可能比較年輕。
持續把資金投入成長。
優先追求市場擴張，而不是當期獲利。
公司的價值可能來自：
IP、
Software、
Data、
Platform，
或者還沒有完全商業化的技術。
如果完全使用為成熟工業企業設計的框架來評估這些公司，自然容易產生結構性 mismatch。
因此，台灣也持續調整創新板制度。
2025 年 1 月 6 日，原本限制只有合格投資人才能交易創新板股票的規定正式取消，市場向一般投資人開放，但仍必須遵循相關風險揭露要求。
到了 2025 年，創新板平均每日成交值已經從 2024 年約 新台幣 1.03 億元增加至約 2.77 億元。
同一年共有 15 家公司申請創新板上市，創下創新板成立以來新高。
政策目標也愈來愈明確。
2025 年 10 月，台灣啟動更大的 **亞洲創新籌資平台（Asia Innovation Capital）**計畫。
TWSE 更直接把創新板定位為打造具有台灣特色 **「亞洲那斯達克」**的重要支柱，希望透過以市值為核心的上市制度，讓創新企業更早取得公開市場資金。
到了 2026 年 6 月，證交所再次把創新板描述為連結創新企業與資本市場的重要橋梁，也是台灣希望成為全球科技資本樞紐的一部分。
這讓我們再次回到 AVA 資料中一組非常值得注意的數字：
TWSE 一般板：21.5 年
TPEx：16.8 年
TIB：10.1 年
創新板的樣本仍然很小。
比較年輕的上市年齡，並不能證明創新板一定會成為「台灣版 Nasdaq」。
成交量增加，也不能保證它最後一定會形成自我強化的 Growth Company Market。
一個真正成功的公開市場，除了上市門檻之外，還需要：
流動性、
機構投資人、
分析師 Coverage、
公司治理、
高品質 Issuer，
以及足夠多成功案例形成正向循環。
但是目前的早期證據至少支持一個比較保守、卻很重要的結論：
創新板已經開始吸引一群明顯比台灣傳統公開市場更年輕的企業。
所以創新板不只是另一個上市板塊。
它更像是一場實驗。
台灣能不能建立一個真正適合成長型企業的本土資本市場，讓那些過去可能選擇繼續維持 Private，或者前往東京、紐約上市的公司，也願意留在台灣？
現在還不知道答案。
但這場實驗已經開始。
8. 台灣不缺拼圖，真正缺的是一座更寬的橋
過去，我們經常使用一些很容易量化的數字衡量新創政策。
一年成立多少新創？
創投投資金額多少？
一年有多少家公司 IPO？
這些都是有價值的指標。
但 AVA 的資料提出了一個令人不太舒服、卻非常重要的可能性：
這些數字可能正在衡量三套不同的系統。
台灣已經有新創。
台灣已經有創投。
台灣也已經有非常活躍的公開資本市場。
過去真正可能缺少的，是：
一座足夠寬、能夠把三者連接起來的橋。
這也意味著，我們衡量台灣新創生態系的方式可能需要改變。
目標不能只是：
「創造更多 Startup。」
如果一家公司通過早期階段後，卻無法跨越 Scale-Up，那麼再增加數千家 Early-Stage Startup，也未必能解決真正的問題。
目標也不能只是：
「增加 VC 投資金額。」
更多資金並不會自動解決國際市場拓展、管理人才、公司治理、客戶取得，以及 IPO readiness 等問題。
同樣地：
「一年有多少家公司 IPO」
也不一定能告訴我們 Venture ecosystem 是否真的成功培養出可規模化的新創企業。
或許真正值得追蹤的 KPI 應該是：
到底有多少家公司，成功走完整座橋？
Startup Formation。
Venture Financing。
Product-Market Fit。
Scale-Up。
Growth Capital。
International Expansion。
IPO Readiness。
Exit。
如果從這個角度來看，台灣真正的挑戰可能不是缺少更多拼圖。
而是：
如何強化這些已經存在的拼圖之間的連結。
而且，我們其實有理由保持審慎的樂觀。
AVA 研究中最值得注意的數字，或許不是：
13,227 家新創，
不是 520 家 IPO，
甚至也不是只有 44 家真正的新創 IPO。
而是：
+0.88。
因為那些真正成功從 Startup 一路走到台灣公開市場的公司，其產業分布，與創投資金配置的產業分布高度一致。
這不能證明整套 Venture system 已經運作得非常有效率。
但至少代表：
創投生態系與最後成功進入公開市場的新創，並沒有朝完全相反的方向前進。
台灣第一套企業製造機器，花了數十年才形成。
科學園區。
工程教育。
供應鏈。
銀行。
出口市場。
跨國客戶。
製造能力。
以及一代又一代的企業家。
最後，它建立出全球最強大的工業生態系之一。
第二套 Venture-backed machine 則年輕得多。
它的制度仍在發展。
Growth Capital Network 還在形成。
與公開市場之間的關係，也仍在重新設計。
而台灣創新板，正是試圖把這些世界連接起來的其中一項實驗。
所以，台灣新創故事的下一個篇章，可能並不取決於：
再成立一個 Accelerator，
再推出一個政府新創計畫，
或者再創造一個「今年成立多少家新創」的漂亮數字。
真正重要的，可能是一件沒那麼性感、卻更關鍵的事情：
建立一套能讓有潛力的新創，成功從 Startup 活著走到 Scale-Up 的基礎設施。
台灣早已證明自己知道如何打造世界級企業。
接下來真正的問題是：
我們能不能為下一代企業，蓋出一座更寬的橋？
如果這篇分析對你有幫助，
我會在 LinkedIn Newsletter 《Taiwan Tech Dispatch》
分享更多沒有寫在文章裡的觀察與判斷。
→ 追蹤並訂閱（LinkedIn）
延伸閱讀
- 台灣的新創、創投與 IPO，其實是三個不同的世界
https://www.angeltoventure.com/post/startups-vc-ipo - 台灣能打造世界級技術，為什麼卻很少打造出全球級科技公司？
https://whitehsu.blog/2026/09/09/taiwan-technology-companies-global-tech/ - FunNow 啟動 IPO：台灣軟體產業正在進入一個新的時代
https://whitehsu.blog/2026/08/31/funnow-ipo-taiwan-software-industry/ - 台灣 AI 新創正在改寫全球創業版圖：AI 如何重寫創業的地理位置？
https://whitehsu.blog/2026/08/17/taiwan-ai-startups/
