Norman Chi, Analyst (紀泳瑜 / 分析師)

負責投資與基金管理。在加入 AppWorks 之前，曾於勤業眾信審計部門服務近兩年，參與過的財務報表查核案件橫跨多個產業。畢業於台大會計系，學生時期曾擔任台大學生會活動部部長與會計系系學會公關長。喜愛品嚐美食、棒球與推理懸疑類影集。期許自己成為理性，但不失熱血與溫度的創投分析師。

不論是與 AppWorks 創業者們閒談，或是觀察他們與 AppWorks Partners 們 Office Hour，總有個老生常談，卻又不可不迴避的議題：現金。

畢竟現金就好比新創戰士的血液，就算擁有再英勇無畏的心，在血液流乾前，如果沒有天使帶著血袋前來補血，也只有倒下成為戰場另一具枯骨的命運。

創業很難不先經歷一段虧損期，尤其在急速擴張之際，很自然會產生大量費用、資本支出，但新創業者在現金流管控的平衡與拿捏上，總是跌跌撞撞。

此外，在找到真正的 PMF (Product Market Fit) 前，創業者需要管理現金，但並不需要做真正的現金流管理。因為這時候的創業者，要全力投入從用戶行為與數據中挖掘 Insight，作為打造 PMF 的依據，最需要的，是相關的知識與能力，而不是想靠著資金來解決問題。這樣的模式與習慣，很容易在找到 PMF 後延續下去，但此時營收開始快速成長，現金流入與支出都變得更為複雜，加上創業者若沒相關的經驗或背景，往往缺乏警覺，很容易就忽略現金流管理的重要性，一旦問題浮現，通常都已是現金水位嚴重不足的危險情況。 

因此，理解如何做好現金流管理與預測，其實是創業者找到 PMF 後，邁向下一個階段應該直接面對的課題。為此，我採訪了 AppWorks Accelerator 校友中曾走過這段路的創業者，與他們交流聊了創業歷程中有關財務、現金流管理的經驗，以及與曾面臨的困難，進而整理出三個重要的 Takeaways，希望對即將或正在這條路上的創業者們，有所幫助。

詳實紀錄，鑑往知來

在創業初期，光是忙著打造產品，與產品 TA 溝通獲取回饋並苦苦尋找 PMF，往往就已耗去創業者大半的心力。然而成功找到 PMF 後，營運的挑戰才正要開始。儘管擁有可規模化的商業模式，進行產品或服務擴展，但這時邁向公司化、制度化，要能踏實地活下來，是否擁有安全、健康的現金流，就需要創業者認真思考且管理。

AW#8 校友、提供房地資訊的 Foundi 房地 創辦人 Andy 鄭依桓就分享，對創業者來說，搞清楚每筆支出的流向，並時刻提醒自己確認 Burn rate (資金消耗率)，是找到 PMF 以後經營成功與否的關鍵。很多創業者會因初期現金流入支出動向相對簡單，還在自己管理能力所及，就忽略了追蹤記錄流水帳的必要，但隨著放大商業模式後，當創業者想探究現金為什麼總是不夠用，就為時已晚了。因此，Andy 分享到，當公司步上軌道，他便開始記錄流水帳，弄清楚每一分錢花去哪裡。只有知道手上有多少資源，才能妥善地安排運用，這是為往後分析與規劃打下良好基礎的關鍵。

另外，因為 Andy 自己是軟體工程師背景出身，一開始對財務似懂非懂。除了買財務分析相關的書籍閱讀外，他嘗試先用試算表，用自己理解的方式打造紀錄工具，再求精進呈現的方式。如果實在有不懂，或不知道怎麼認列的，AppWorks 財會輔導長 Jasmine 蔡金鳳，便是尋求協助最好的對象。

他也分享到，除了詳實記錄現金進出外，更重要的是每月定期回顧現金流入與支出，是否與事前的規劃差很多？每年的十二月，Andy 便會著手下一年度的規劃，根據今年的狀況調整並嘗試預測下一年度的現金流。在每月固定的檢視過程中，除了能即時知道手上究竟還有多少資源可以分配，更可以知道哪裡之前預測疏忽了，並得以對資源配置做即時調整。對創業者來說，因為已經有了創業初期都是在賠錢的覺悟，所以更要好好掌控自己的支出，把資源花在當下最有價值的地方。

瞭解自身的變動與固定支出，有多少本事花多少錢

找到 PMF 並加速規模化的過程中，評估資金充足與否變得更加重要。因為市場變化總是非常快，可能忽然黑天鵝振翅飛過，抑或競爭對手進行大手筆補貼進行削價競爭，甚至其他市場巨頭開始做和你一樣的產品等，使新創成長不如預期。這時如果一不小心，自己辛苦打造的事業就可能陷入不堪虧損的狀態。

AW#4 校友、專門協助企業進行 email 與電子報行銷的 NewsLeopard 電子豹創辦人 Louis 張國基就分享，當創業處在順境時，通常不會發現現金流的問題。根據他的親身經驗，創業初期也是覺得營收數字都在增加，且已經大概估算帳上現金可以維持 18 個月，便對自己的現金流感到放心，但很快地，現金開始少於預設水位，而且持續探低，Louis 便警覺不對勁。而更不利的，是產品面臨改版優化，要重新上線、測試並培養客戶忠誠度皆需要時間，但每個月的支出還是要付，最糟糕時曾經只剩不到半年的現金水位。在最黑暗的時刻，只好靠自己信貸咬牙撐過。

正是因為走過被金流壓到喘不過氣的歲月，Louis 才決定痛定思痛，開始分析公司現金流到底出了什麼問題？也是在這時，「固定與變動成本」的觀念進入了他視線內。什麼是固定成本？舉凡每月繳交的辦公室租金、員工薪資等，不隨著營業額高低改變的支出皆屬固定成本；相對地，變動成本即是會隨著提供產品或服務改變的項目，例如隨營業額增加的業績獎金。

有了這個觀念，才能理解每月營運的成本結構。再搭配公司每月現金流入分析，拆分什麼是經常性流入的收入，什麼是一次性的收入 (例如專案收入、政府針對創業的一次性補助、融資借款等)，進而透過各種情境，模擬推算出手上現金究竟還能撐多久，從而及早準備應急措施，或是檢視什麼支出是該省下來的。

以電子豹為例， AWS (Amazon Web Services) 的費用有部分會歸類為固定費用支出，另一部分則是依據主要業務而變動的成本，例如，每月固定支付 AWS 雲端資料存儲的費用 (DynamoDB, S3)，就會被歸類為固定支出；依據郵件發送量計價的 SES (Simple Email Service) 服務，則是列為變動成本。Louis 以前並不會特別去關心這件事，但開始對錢斤斤計較後，就會促使自己嘗試分開計算，並得到更精準的現金流出數字，Louis 笑著說：「量力而為、量入而出，以前太常衝動行事，總認為只要業績成長，現金流就不太會出問題，所以得到慘痛的教訓。」

掌握事業的收入與支出 Pattern (模型)

有了確實的記錄，並逐步了解事業的成本、收入結構後，創業者最終必須很清楚自己選擇的營運模式，究竟是屬於 B2B、B2C、專案型還是 SaaS 模式？ 每一種都有不同的 Cash Cycle（現金循環) 及潛在的營運風險。AW#1 校友、台灣線上餐廳訂位及餐券預購龍頭 EZTABLE 創辦人 Alex 陳翰林即分享，剛開始創業時，金流單純，頂多只需煩惱應收帳款多久可以收回來。但自從 EZTABLE 開始提供預購餐券後，他發現損益表逐漸失真：營運模式開始變得有點像電商，要先準備一筆錢付給餐廳賣出的預售餐券 (類似電商的進貨成本)，時限內若沒銷售出去，便須自行吸收損失 (類似電商的庫存損失)。那時公司又處於高速擴張期間，錢收進來後，很快就要急著去買下一批貨，造成手頭非常拮据，無法負擔其他支出。從那個時候，他便開始要求內部週會的第一件事，就是先看現金流狀況，並嘗試打造 EZTABLE 的財務與現金流模型。

隨著營運模式改變，Alex 決定開始觀察並計算過去現金流變化。因為現金水位處於枯水期，通常與付薪水和購買商品 (餐券) 的時間息息相關。透過統計收現、付現的週期，並根據自身的採購議價能力與對市場、客戶的認知，他會先試著抓出一個最高與最低現金水位的範圍，並隨時調整，幾個月後，基本上就能對現金如何進出有一定的感覺，進而找出收支平衡的 Pattern，架設屬於最適合自己的模型。儘管每位創業者的方法不同，但一定要知道自己公司現金水位最高與最低的日子與金額數字。

當然，在掌握現金水位的過程中，撞牆期是無可避免的。Alex 提到，EZTABLE 最初的模型歷經一年九次改版，最終才完成較為精準的版本。在一路的修改過程中，從早期的基本加減乘除，後來運用指數、逼近法，甚至現在能直接預測消費者行為。因為累積足夠多的數據與失敗經驗，並加上一些經濟學指標做為計算參數，讓 EZTABLE 現在可預測下個月現金水位最低點時，需要有多少錢在手上才算安全。

「撞牆一定要去撞，」Alex 說。他認為，很少創業者沒撞過現金流壓力的大牆，常看到很多的新創或企業面臨現金周轉不靈，但他們都是第一次撞牆嗎？肯定不是。「一定是有人之前靠借錢急救或是運氣撐過去，如果最後仍是因週轉不靈而倒下，基本上一定是玩太大或是已經把借錢信用耗盡了。」建立現金流模型，重點不在於如何預防撞牆，而是每一次撞牆後，有沒有化為往後的經驗，得到更精準的模型，才是對創業者最重要的事情。

結論

現金流之於創業者的重要性，可用 Alex 的心得總結：「創業的過程中，如何花錢絕對不是賭注，你可以 Gamble (賭博) 其他任何事情，但創業談到錢、談到現金流，絕對不能像在賭場一樣恣意揮霍。生意是做長久的，現金流長久健不健康，與花錢的紀律是正相關的，不能憑感覺花錢。」

把現金流盡可能預測準確，避免讓公司陷入危險的境地，是身為創業者最重要的責任之一，也是新創得以持續發展的必要條件。透過以上三位成長期創業者的分享，希望能幫助在類似階段的創業者，更瞭解如何掌控現金流。也希望明明賺錢但卻因現金而倒下這樣令人惋惜的結局，再也不要在創業圈發生。

【歡迎所有 AI / IoT、Blockchain / DeFi、佈局東南亞的創業者，加入專為你們服務的 AppWorks Accelerator】

Photo by Breitling on Pixabay

Jun Wakabayashi, Analyst (若林純 / 分析師)

Jun is an Analyst covering both AppWorks Accelerator and Greater Southeast Asia. Born and bred in America, Jun brings a wealth of international experience to AppWorks. He spent the last several years before joining AppWorks working for Focus Reports, where he conducted sector-based market research and interviewed high-level government leaders and industry executives across the globe. He’s now lived in 7 countries outside US and Taiwan, while traveling to upwards of 50 for leisure, collectively highlighting his unique propensity for cross-cultural immersion and international business. Jun received his Bachelors in Finance from New York University’s Stern School of Business.

As the world came to a standstill from COVID19, the first half of 2020 proved to be a challenging time for industries across the board. With all its promises and potential, even the nascent blockchain industry was not immune to worldwide lockdown and social distancing measures. A handful of established players certainly benefited from the accelerated shift to online consumption and digital living. But, those projects without solid foundations, compelling use cases, and sustainable business models quickly found themselves at the end of their runways.

With only 447 confirmed cases and 0 local transmissions in the last two months, Taiwan has garnered worldwide adoration for its handling of the novel coronavirus. The local blockchain industry, however, did not go completely unscathed. Compared to 19H2, 21 projects have since ceased operations and been taken off the map. This is the largest number of removals since AppWorks Accelerator started producing the ecosystem map two years ago. With only 16 additions, there were also a fewer number of new organizations entering the ecosystem compared to any other half. Collectively, there was a total of 112 projects noted on the map, down from 115 in the last update.

2020 is clearly shaping up to be a year of reckoning. Financial markets are in disarray, global supply chains have been crippled, and economies across the globe are struggling with unprecedented productivity loss. That said, if 2008 was any indication, heroes are often born in the midst of a crisis. The pandemic has highlighted glaring efficiency and reliability gaps in our current processes and systems, forcing upon us new ways of working and thinking. “The new normal” may very well serve as the necessary accelerant to convert blockchain’s perception of cautious skepticism into widespread pragmatism, which we’re already starting to see early signals of.

Enterprise adoption picks up

There’s been a lot of talk about how and when to potentially use blockchain. Now, it seems like enterprise adoption is starting to pick up steam. As many of them lack the capabilities in house, we’ve seen several corporates in the last six months announce their latest initiatives in partnership with local blockchain startups.  

Taiwan’s leading financial conglomerate with over US$335B AUM Cathay Financial Holdings announced that it’s developing an electric vehicle IoT financing platform in conjunction with EV charging service provide ChargeSmith (AW#16) and supply chain-focused blockchain startup BSOS. The new platform collects and stores EV owners’ driving behavior, which will then be used to provide real-time personalized financial services ranging from insurance to banking products. In another example of enterprise partnerships, longstanding email security provider Openfind unveiled its collaboration with BlockChain Security to launch an intellectual property protection solution, which aims to help enterprises store and transmit confidential information more securely. 

It’s no secret that banks have been some of the earliest adopters of blockchain technology in Taiwan, as seen by Cathay and CTBC, the latter of which runs its own blockchain incubator and is a part of international blockchain consortium R3. Taiwan Business Bank was the latest to hop on the bandwagon. They recently unveiled their intentions to implement blockchain technology into the process of issuing confirmation letters in order to lower the risk of falsification.

Meanwhile, crypto asset management platform Bincentive recently disclosed its partnership with Japan-based Rakuten. Users of the international e-commerce site will now be able to purchase crypto investments using their loyalty points. Along the same lines, BitoEx (AW#6) which operates one of Taiwan’s largest crypto exchanges now allows FamilyMart members to redeem their loyalty points for bitcoin purchases.

Thus far, Tawianese corporations are clearly displaying a positive embrace of blockchain technology. However, due to their traditional management structure, implementation of the technology is still veered towards a centralized model, inhibiting some of the touted benefits of blockchain such as autonomy, no single point of failure, and overall the democratization of data and authority. Moving forward, it’ll be worth observing how enterprises might adjust their cultures and governance models as decentralization increasingly transitions from fringe to mainstream adoption. 

The new normal

Blockchain has been in development for well over a decade now, and the crypto market has correspondingly undergone its fair share of cycles. In the process, investors have become increasingly more sophisticated in terms of their understanding and approach to the market, as well as the instruments and structures they use to finance deals. Nevertheless, crippled economies across the globe have led investors to reel back their money for the time being, allocating their attention to existing portfolios or waiting for brighter horizons.

With the current circumstances, the development of early stage startups has become polarized. The market will be unrelenting for those relying on projects to generate ad hoc cash flow, as the downturn has caused a tightening of the belt for customers across the board. On the other hand, those teams that have found product-market fit and solve a real world pain point for clients at scale will have greater commercial prospects, ultimately contributing to a stronger and more sustainable ecosystem that’s driven by a project’s actual value rather than market hype and speculation.

There were a few bold blockchain entrepreneurs, for example, with innovative technology and solid business models that were highly favored by investors. CoolBitX successfully closed a US$16.75M series B round of financing, led by SBI Holdings. The newly raised funds will be used to continue development and rollout of its messaging-based KYC/AML solution for virtual asset providers. In addition, BSOS also completed a million-dollar seed round in the first half of 2020—including NT$20M (US$678K) from Taiwan’s National Development Fund—to advance its blockchain-based supply chain financing solution for enterprises. 

Digitalization vs decentralization

Looking back in the first half of 2020, it was a rather harrowing time for the blockchain industry. Right at the onset of COVID19 in March, cryptocurrency prices dropped by 40%. A few months later in May, Telegram—which in 2018 raised the second largest ICO in history at US$1.7B— announced the suspension of its blockchain project the Telegram Open Network (TON) shortly after the SEC won a preliminary injunction, citing a plausible case that Telegram had sold unregistered securities in its coin offering.

Crypto prices have since more or less recovered to pre-COVID19 levels, especially as people increasingly sought out alternative stores of value like bitcoin as a hedge against inflationary currencies. The halving of bitcoin this past May and the added supply scarcity will likely continue to drive up its value, as it’s done so consistently in the past. Since TON was officially abandoned, there have been many communities around the globe that have been rallying to keep it alive, including the Chinese TON community which claimed that it would launch its own version of the blockchain. 

At this point, what are the next growth opportunities for blockchain? Everyone has different interpretations and expectations. However, currently, conversations are mostly centered around the pandemic, with criticisms mainly targeted at the limitations and failures of centralization. Whether it’s in terms of tracing vaccination records, tracking medical supplies, eliminating fake news, or disbursing economic aid, decentralization has been thrown in as a prominent solution to many use cases. However, as impactful as these use cases may be, the reality is that boilerplate digitalization is still higher on the pecking order for most people. The age of remote work catapulted people into a new normal, forcing everyone to figure out how to do everything over the worldwide web, from basic necessities like ordering food or groceries to lifestyle needs like exercise and socialization to work needs like conducting meetings or business presentations. 

Digitalization and decentralization are not mutually exclusive, however. If anything they are complementary. Novel thinking often calls for novel solutions. Decentralization can provide the perfect combination of efficiency, transparency, and autonomy that people didn’t even know they wanted or needed. For now, aside from decentralized finance (DeFi), games and entertainment-related services will likely dominate the majority of use cases, both of which  benefited tremendously from the pandemic in terms of adoption and usage. 

Existing entrepreneurs in this space would be wise to continue building their experience in this area, acquiring users, and strengthening their positioning. For aspiring entrepreneurs, realize that the world has hit reset and undergone a rude awakening; consequently, there’s never been a better time to explore and introduce radical new ideas.

Taiwan’s Blockchain Ecosystem Map First Half 2020 is produced by AppWorks and updated every six months. If you have any comments or suggestions, please send us an email at [email protected].

【If you are a founder working on a startup in SEA, or working with AI / Blockchain, apply to AppWorks Accelerator to join the largest founder community in Greater Southeast Asia.】

Photo by Robert Pastryk on Pixabay

Natalie Lin, Analyst (林楓 / 分析師)

Natalie is an Analyst covering AppWorks Accelerator and Greater Southeast Asia. Before joining the team, she worked in the search engine marketing and email marketing teams at Zappos, America’s leading shoes and fashion online retailer, where she primarily focused on KPI management, campaign optimization, and project management. Born in Canada and raised in the Middle East, Natalie returned to Taiwan for high school before moving to the US for college and work. She received her Bachelors of Marketing at Case Western Reserve University in Cleveland, Ohio. Outside of work she likes to read, travel, and play video games.

As a founder thinking about Taiwan, engineering talents and a sizable and well developed digital ecosystem are normally at the top of mind. But increasingly entering the mix now is Taiwan’s growing AI capabilities. It is a result of both industry trends and government policies. In her inaugural speech given in May, re-elected President Tsai Ing-wen mentioned six core strategic industries for Taiwan that included an acceleration of AI and IoT while highlighting the future of digital and data. 

In order for an AI ecosystem to flourish, both the government and the private sector must embrace and work to build it. In October 2019, the National Center for High-Performance Computing co-worked with Quanta Computer, ASUS, and Taiwan Mobile to build TAIWANIA 2, a supercomputer built for AI, which offers its cloud service through Taiwan Computing Cloud (TWCC). Since its commercialization, TAIWANIA 2 has driven over 300 projects from industries, academia, and research organizations. 10 million GPU computing hours have been utilized, AI training time has been reduced by 90%, and the efficiency of deep learning has been raised 498 times. The added capacity far exceeds the current computing demands of Taiwan’s AI industry, seamlessly aligning with the country’s national strategy. 

As Taiwan gradually becomes a major AI center for the development of Greater Southeast Asia (ASEAN + Taiwan), tech companies are starting to pay attention. Microsoft has already promoted their AI Infinity program to recruit talents and set up R&D centers. In addition to covering commercial AI solutions, think tanks are being set up and Microsoft is also cooperating with National Chengchi University to open AI business courses for enterprises to understand how to implement AI to accelerate transformation within their company.

Although the new virus that broke out in the first half of 2020 halted some commercial activities and significantly impacted many existing industries, the epidemic did not stop the development of Taiwan’s AI ecosystem. In the process of updating the 2020 H1 Taiwan AI Ecosystem Map by AppWorks Accelerator, we found that in the last 6 months, there have been several trends worth noting:

The retail industry was forced to go online during the pandemic, giving digital marketing a platform to shine

During the pandemic, consumer behavior has changed. Their shopping habits accelerated from going to physical stores to switching to online channels instead. How companies can stand out in the online world has become a new learning curve for retailers. This epidemic will not only accelerate the speed of digital transformation of the traditional retail industry, but also give rise to smart retail, which integrates online and offline experiences as their focus.

In the past few years, we’ve noticed that AI innovation focusing on digital marketing operations have sprung up. By using AI technology to accurately do ad placements, track user journey and behavior, and analyze the collected data, companies are able to outline consumer profiles and thus more accurately predict consumer shopping preferences and price orientation, creating personalized shopping experiences for each visitor.

Even visitors that don’t end up converting or placing an order provides a gem of data analysis. Companies can use AI and big data to target this group of potential customers who have yet to make a transaction by carrying out remarketing campaigns to improve conversion rates and overall marketing performance.

The opportunities are endless under this trend, especially for startups. In February 2020, MarTech startup Accuhit, which just raised a NT$ 70 million pre-A round, provides both subscription-based use and project-based consulting to analyze customers’ historical purchasing behavior and make predictions. Omnichat (AW#16), an alumni of AppWorks Accelerator, specializes in a marketing automation subscription service that provides multi-channel conversational commerce focusing on ecommerce clients. The platform has an average conversion rate 3x to 7x higher than the overall trend in ecommerce. They recently completed a seed round earlier this year, led by AppWorks Funds.

An explosive introduction of AI applications in healthcare

Alongside digital marketing, AI is also thriving within the medical field. According to Global Market Insights, the global healthcare AI market will reach US$ 13 billion by 2025, with an average annual compound growth rate (CAGR) of 40.6% from 2019 to 2025. Over the past six months, particularly during this pandemic, AI usage in improving the efficiency of medical consultation and disease screening is at an all-time high.

Taiwan has accumulated years of experience in the development of AI in healthcare. Since medical AI algorithms require a large amount of labeling data in order to train the model, as early as 2018 the Ministry of Science and Technology (MOST) co-worked with National Taiwan University, Taipei Veterans General Hospital, and Taipei Medical University to open the first pathology database for cross-institutional annotated medical images in Taiwan. These kinds of developments not only help physicians accelerate medical image interpretation and improve the consistency and accuracy of diagnoses, but also shorten the time for patients to seek medical treatment and reduce invasive examinations.

In the startup ecosystem, we also see a rise in the development of AI in healthcare startups. We see applications like developing AI technology for over a billion pixel image analysis and high-performance computing (AI HPC) with AI Explore; medical imaging AI platform services with aetherAI; medical staff assistance to interpret Computer Tomography (CT) scans for cerebral hemorrhaging with Deep01; and during this epidemic, Heroic-Faith developed a smart respiratory monitoring system.

A friendly ecosystem for AI startups is gradually taking shape

As more AI startups plan on expanding to other markets, Taiwan is considered a hotbed in terms of startup accelerators, education, and research for AI founders to leverage the ecosystem. On the accelerator side, AppWorks Accelerator was established in 2010 and focuses on recruiting AI / IoT and Blockchain founders only since August 2018 (AW#17). So far, it has recruited 89 of these startups, with 44 of them being AI and 16 of them being IoT. Together these startups are well positioned to further accelerate Taiwan’s burgeoning AI ecosystem.

In addition to AppWorks, there are many other accelerators supporting the AI ecosystem in Taiwan. Microsoft for Startups, Sparklabs Taipei, Taiwan AI x Robotics Accelerator, and more are all community builders that recruit AI startup teams. Under this AI wave, other accelerators that aren’t focusing on AI have also found success in recruiting AI startups, such as TAcc+ and BE Accelerator.

In VC, investment in AI companies have also grown year-over-year, providing financial assistance to AI startups looking for growth or expansion opportunities. For example, AppWorks Funds is a VC focusing on the three themes of AI, Blockchain, and Southeast Asia. Other well-known VCs in Taiwan include ACE Capital, Cherubic Ventures, TransLink Capital, and more have also greatly invested in AI in the past 2 years. Related Taiwanese investments to AI startups include CloudMile, Umbo CV (AW#9), Cubo AI (AW#16), and MoBagel (AW#16).

In Taiwan’s thriving AI ecosystem, new AI developments can be seen in various well-known technology exhibitions and forums. At the annual COMPUTEX conference, there’s a new startup-themed exhibition called InnoVEX that has attracted many talented AI founders to sign up to build bridges with investors and enthusiasts from all over the world. This year due to the pandemic, COMPUTEX was cancelled for 2020 and InnoVEX changed to #InnoVEXOnlineDemo. NVIDIA also held their famous AI conference NVIDIA GTC (GPU Technology Conference) in Taipei in 2018, providing opportunities for all stakeholders to meet and each other. GTC Taiwan 2020 is expected to be held this year in Taipei.

Taiwan AI Academy and Taiwan AI Labs are still Taiwan’s representative institutions in the field of AI education and research, respectively. They continue to inject talent and innovation to Taiwan’s AI development. There was a tragic loss for the community of Taiwan’s AI ecosystem as Sheng-Wei Chen passed away earlier this year. He was the CEO of Taiwan AI Academy and founded the organization that has cultivated more than 6,000 talents for Taiwan in the span of 2 years. His contributions include progressing the mission of transforming industries with AI, making data science known as his north star, and growing AI talent.  He and his work will forever be remembered by Taiwan’s AI growing community.

Taiwan’s AI Ecosystem Map First Half 2020 is produced by AppWorks and updated every six months. If you have any comments or suggestions, please send us an email at [email protected].

【If you are a founder working on a startup in SEA, or working with AI / Blockchain, apply to AppWorks Accelerator to join the largest founder community in Greater Southeast Asia.】

Photo by tingyaoh on Pixabay

Ching Tseng, Associate (曾意晴 / 投資經理)

負責區塊鏈投資，尤其專注東南亞市場。學生時期曾於 AppWorks 實習一年半，2015 年政大企管系畢業後正式加入擔任分析師，主要參與投資案相關業務，最得意的案例是協助 CHOCO TV 從 A 輪一路到被 LINE 併入。是我們的年輕人趨勢專家，2019 正式升任經理。平常熱愛嚐鮮、美食以及旅遊。

對多數人來說，2020 上半年，過得並不輕鬆。由於 COVID-19 在全球爆發，改變了人們許多的工作與思考方式。對尚屬於發展初期、還仰賴很多實體活動來獲取用戶、投資人的 Blockchain 產業來說，不少新創因此面臨生存危機，更有不少服務的推進計畫，也因疫情延宕了，募資過程受到干擾的團隊更不在少數。

在 AppWorks Accelerator 每半年更新一次「2020 H1 台灣 Blockchain 生態系地圖」(Taiwan’s Blockchain Ecosystem Map First Half 2020) 的過程中，就可清楚感受到大環境帶來的衝擊。在這一版的更新中，總計納入 112 家相關的新創、機構與單位，第一次列入的共 16 家，是過去三次以來新納入數量最少的一次，而半年前還在地圖上的，則有 21 家已從這一版移出，也是歷來數量最多的一次，兩個數字此消彼長之間，代表過去半年間，在台灣 Blockchain 生態系中新生的新創變少、折損的新創變多。

然而，歷史上亂世時，卻也是英雄出世的時刻。即便有不少 Blockchain 創業者暫時退出賽道，卻也有許多的團隊，在這樣的情況下，繳出傲人的成績單。

機構用戶導入加速，創業技術 Support 

Blockchain 發展至今，不時會有大企業高呼要揮軍 Blockchain 領域的消息，近半年來，過去累積的底蘊發酵，大型企業與機構導入 Blockchain 技術的增速比過去更快；乘著 2019 年下半年，以輸出技術成為服務的趨勢，許多 Blockchain 創業團隊，也抓緊了這波機會，與各種機構用戶結盟。

例如，國泰金控數位數據科技研發中心近期發佈，將在 2020 年底推出「電動車車聯網區塊鏈金融平台」，除了能將數據透過加密存證上鏈外，還結合國泰產險、國泰世華銀行，提供從汽車融資、保險理賠到商品推薦等服務。在這背後，國泰也與 AppWorks Accelerator #16 (AW#16)、充電站服務平台 ChargeSmith 宅電以及 Blockchain 新創團隊 BSOS 共同合作，透過電動車的行車資料上鏈，來協助國泰金控提供更豐富的銀行及保險業務場景。

其他大型企業與 Blockchain 新創合作的案例，還包括樂天市場與 Bincentive 合作，讓會員能將樂天購物點數轉為加密貨幣投資。OpenFind 網擎資訊則與 BlockChain Security 區塊科技合作，結合區塊鏈存證應用，協助企業在處理機密資訊時能夠更安全的儲存與傳輸。另一較親民的例子，是全家便利商店與 AW#6 BitoEX 幣託合作的會員回饋系統，民眾在全家進行消費並累積會員點數，就可以免費兌換成比特幣。此外，台灣企銀也推出金融區塊鏈函證，減低函證作業流程中，容易發生資料遺失或者遭竄改的風險。

大企業加速導入 Blockchain 技術顯然是一樁好事。但也因為大企業的管理型態，在考量 Blockchain 技術時，還是會偏向較中心化的商業模式，結合 Blockchain 技術去中心化的本質，未來將帶來哪些新發展、新演化，值得持續關注。

炒作行情結束，發展各憑真本事

Blockchain 發展至今已有一段時間，投資加密貨幣也經歷過幾次上沖下洗的行情。在這過程中，投資人的型態與工具也逐步調整、日趨成熟，已很少追捧熱門的短線議題，而受到這一波疫情影響，投資步調甚至有些放緩。

在這樣的外在環境下，使得新創發展呈現兩極化，由市場啟動了淘汰機制。新創的商業模式，若偏重靠接專案帶來營收，隨著經濟前景保守，開發新客戶與專案的難度提高，直接造成現金流面臨斷炊的問題；另一方面，持續累積自有客戶、挖掘商業護城河的新創團隊，則能靠著業績持續成長、逐步累積 Traction 向投資人募資。

儘管時局艱困，但過去半年也不乏體質健全、前景展望佳的 Blockchain 新創，獲得投資人青睞。例如，CoolBitx 庫幣科技在持續推廣旗下產品 Sygna 的過程中，也完成由日本 SBI 集團領投的 B 輪募資，金額高達 1,675 萬美元。除此之外，BSOS 也在這段期間，完成百萬美金的種子輪募資，其中包含國發基金挹注的 2,000 萬新台幣。

數位化優先於去中心化

回顧 2020 上半年全球的 Blockchain 發展，無論是三月主流加密貨幣閃崩，24 小時內多數跌幅超過 40%；又或者是五月 Telegram 因不敵政府監管的壓力，宣佈暫停旗下區塊鏈項目 TON 的開發。

此時，探討 Blockchain 下一個成長機會是什麼？每人各有不同解讀與期待。然而，在疫情盛行的當下，許多人詬病中心化組織，在傳遞正確訊息的速度及可靠性，紛紛討論起去中心化的優點，提出不少想法。但回頭反觀現實世界的發展，最多數人的痛點，都是圍繞在遠端工作情境，對網路創業者來說，數位化早已如空氣一樣存在已久，而對多數大眾的需求與使用情境而言，數位化的優先順序，還是遠高於去中心化。

整體觀察下來，Blockchain 產業依舊在等待下一個浪潮，而這道浪，得為產業注入下一群從未使用過 Blockchain 的新用戶。根據過往多次數位典範轉移的經驗，遊戲、娛樂相關的服務，最有可能帶來大量願意嘗試新技術的用戶，想要加入的創業者，若有這方面的優勢，不妨從中思考可以立足的機會。

已經在耕耘的創業者，在此逆風時刻在確保資金不斷鏈的前提下，不妨趁此練兵、多探索各種可能性，持續建立自己的獨特性，等下一波浪潮席捲而來時，抓準時機獲取有質量的用戶。

造浪者、衝浪者，只要還在海上，永遠都有下一波乘風破浪的機會。

Taiwan’s Blockchain Ecosystem Map First Half 2020 由 AppWorks，以及 Blockchain 媒體 Blockcast 區塊客 (AW#14)、BlockTempo 動區 (AW#16) 聯合製作，每半年更新一次，有任何指教與建議，請 email 至 [email protected]。

【歡迎所有 AI / IoT、Blockchain 與佈局東南亞的創業者，加入專為你們服務的 AppWorks Accelerator】

Photo by Robert Pastryk on Pixabay

Norman Chi, Analyst (紀泳瑜 / 分析師)

負責投資與基金管理。在加入 AppWorks 之前，曾於勤業眾信審計部門服務近兩年，參與過的財務報表查核案件橫跨多個產業。畢業於台大會計系，學生時期曾擔任台大學生會活動部部長與會計系系學會公關長。喜愛品嚐美食、棒球與推理懸疑類影集。期許自己成為理性，但不失熱血與溫度的創投分析師。

AI 在台灣不管是業界還是政府，都已是重點發展的項目。甫連任的總統蔡英文在就職演說中，談及未來台灣六大核心戰略產業，第一項就是資訊及數位相關產業，並宣示將全力促進 AI 與 IoT 的發展。

在政策支持、民間企業響應下，AI 的發展，正在台灣快速開花結果。例如，2019 年十月，由國研院國網中心攜手廣達、華碩、台灣大哥大共同打造的人工智慧超級電腦「台灣杉二號」與「台灣 AI 雲」( Taiwan Computing Cloud，簡稱 TWCC ) 正式商轉後，迄今已協助約 300 組產學研界專案計畫， 提供了上千萬的 GPU 運算小時使用量，協助縮短 AI 訓練時間達 90％，並提高深度學習效率 498 倍 ，大大滿足了產業界的 AI 運算需求，也是政府前瞻計畫中非常重要的一環。

當台灣逐漸成為大東南亞 (東協＋台灣) 發展的 AI 重鎮之時，科技巨擘 Microsoft 除早已大舉來台徵才、設立研發中心外，更推動台灣的 AI Infinity 計畫，除網羅各業界經由 Microsoft AI 技術驅動的商用 AI 解決方案，以此建立台灣 AI 智庫外，也攜手政治大學，為企業開設 AI 商學院，協助企業整體升級，運用 AI 加速轉型。

2020 年上半年爆發的新冠肺炎，雖然使部分商業活動停滯，甚至顛覆了不少既有產業，但疫情並沒有使台灣 AI 生態系的發展停下腳步。在 AppWorks Accelerator 每半年更新一次「2020 H1 台灣 AI 生態系地圖」(Taiwan’s AI Ecosystem Map First Half 2020) 的過程中，我們發現在過去半年中，出現了幾個值得注意的重要趨勢：

零售業重新佈局邁向全通路經營，給了 Digital Marketing 最好的年代

在後疫情時代，當消費者購物行為從實體商店加速轉往線上通路，企業要如何讓自家品牌在網路世界脫穎而出，成了新零售顯學。這次的疫情，不僅將促使傳統零售業數位轉型速度加快，也讓線上線下融合體驗的智慧零售成為焦點。

在過去幾年，我們可以發覺，專注於數位行銷領域營運的 AI 新創，如雨後春筍般冒出。透過將 AI 的技術，運用在精準投放數位廣告、追蹤用戶旅程 (User Journey)，並分析蒐集來的數據，讓企業更容易描繪出消費者輪廓，進而更準確地預測消費者的購物偏好與價格取向，並為每一位造訪的消費者打造個人化、客製化的購物體驗。

此外，那些雖然沒下單就離站的用戶數據，也成為數據分析的瑰寶。企業可透過 AI 與大數據的應用，針對這群尚未成交的潛在消費者，進行精準的再行銷推廣，進而提升轉換率與行銷績效。

在這個趨勢下，可看到許多潛力新創。例如，剛於 2020 年二月獲得 7,000 萬新台幣 Pre-A 輪募資的 MarTech 新創 Accuhit 愛酷智能，即是提供科技行銷系統，來協助企業客戶掌握用戶旅程，並提供行銷的數據整合，解決方案可運用的範圍之一，就是協助藥局整合線上會員的資料；此外，AppWorks Accelerator #16 (AW#16) 校友、專攻行銷機器人訂閱服務的 Omnichat，提供的全通路對話式商務，可協助電商客戶進行跨通訊平台的行銷方案，轉換率比電商整體平均高 3 到 7 倍，也在 2020 年三月完成 2,400 萬新台幣的種子輪募資。

醫療領域導入 AI 應用蓬勃發展

除了數位行銷，醫療領域是台灣在 AI 運用上另一個重要領域。依據 Global Market Insights 的預測，至 2025 年，全球 AI 醫療應用市場規模，將達 130 億美元，2019 至 2025 年的平均年複合成長率 (CAGR) 達 40.6%。過去半年，特別是疫情影響之下，如何利用 AI 提升看診與疾病篩檢效率，更是把這股 AI 應用於醫療產業的浪潮推向了高峰。

事實上，台灣在 AI 醫療的發展已累積多年經驗。由於開發醫療 AI 演算法，需要大量疾病標註數據，以此來訓練 AI Model，因此，早在 2018 年底，科技部就與台灣大學、台北榮民總醫院、台北醫學大學合作，建置了台灣首座本土化跨醫療院所的「AI 醫療影像資料庫」，並將成果與其他研究團隊及資料建置團隊合作，進行醫療衛生目的相關的學術研究，以提升台灣醫療技術。透過 AI 辨識的技術，不僅可以協助醫師加速醫療影像判讀及提高診斷的一致性與精準度，也可以縮短病人的就醫時間及減少侵入式檢查，降低醫療支出。

在新創圈，也可看到越來越多團隊加速開發 AI 在醫療上的應用。例如，開發即時 10 億級像素影像分析高效能運算 (AI HPC) 平台的 Ai Explore 影豹科技 ；提供醫療影像 AI 開發平台服務的 aetherAI 雲象科技；協助醫療人員判讀電腦斷層掃描（Computer Tomography，CT）並獲得台灣與美國 FDA 認證、提供腦出血 AI 影像判讀系統的 Deep01 ；以及在此次疫情下，開發出智慧呼吸監測系統的 Heroic-Faith 聿信醫療等等，展現出台灣 AI 應用在醫療領域上呈現欣欣向榮之姿。

AI Startup Friendly 的生態系逐漸成型

隨著越來越多 AI 新創冒出頭並打算出海，台灣在創業加速器、教育、研究等面向，也依舊蓬勃發展。在創業加速器部分，2010 年成立、自 2018 年 8 月 (AW#17) 起限定招募 AI / IoT 與 Blockchain 新創的 AppWorks Accelerator，目前為止已經招收四期 AI / Blockchain 共 89 組新創，其中有  44 組 AI、16 組 IoT 新創，持續為台灣 AI 生態系挹注新能量。

AppWorks 之外，包括微軟新創加速器 (Microsoft for Startups)、Sparklabs Taipei、Taiwan AI x Robotics Accelerator 等，皆是以 AI 新創團隊為主要招募對象的創業加速器。而在 AI 浪潮下，其他非限定召募 AI 新創的加速器，包括 TAcc+、BE Accelerator，也吸引不少 AI 新創團隊進駐。

在創投領域，AI 新創的投資案也逐年增加，提供新創擴張甚至出海的資金協助。例如，AppWorks Funds 為專注於 AI、Blockchain、Southeast Asia 這三個主題的創投基金，此外，包括 ACE Capital、Cherubic Ventures、TransLink Capital 等國內外知名創投，也都在近兩年投資不少 AI 相關的台灣新創，例如 CloudMile 萬里雲、Umbo CV 盾心科技 (AW#9)、Cubo AI (AW#16)、MoBagel 行動貝果 (AW#16) 等。

隨著台灣 AI 生態系蓬勃發展，在各個知名的科技展覽與論壇中，也可看到不少 AI 新創的身影。例如，歷史悠久的 COMPUTEX，旗下以新創為主題的特展 InnoVEX 即吸引不少 AI 團隊報名參展，並為台灣 AI 創業者搭橋，有機會與來自全世界的新創圈人士與投資人進行深度交流 (受疫情影響，2020 年 COMPUTEX 確定取消，InnoVEX 則改為線上 #InnoVEXOnlineDemo)；NVIDIA 2019 年也在台北舉辦了 AI 盛會 NVDIA GTC (GPU Technology Conference) 全球巡迴海外首站，提供各界互相交流的契機，而 2020 年的 GTC Taiwan 2020，預計也將回到台灣舉辦，屆時又將是一場 AI 的饗宴。

台灣人工智慧學校、台灣人工智慧實驗室，則分別仍是台灣在 AI 教育與研究領域的代表性機構，持續為台灣 AI 發展挹注人才與創新技術的能量中。對台灣 AI 生態系來說，2020 上半年最令人遺憾的，莫過於台灣人工智慧學校執行長陳昇瑋的英年早逝，他與台灣人工智慧學校，在二年內，為台灣培育超過 6,000 位 AI 人才，全力達成產業 AI 化及轉型升級的使命，充滿熱情、實踐理想的典範，不僅讓人無限懷念，更為台灣 AI 發展留下不可抹滅的貢獻。

Taiwan’s AI Ecosystem Map First Half 2020 由 AppWorks 製作，每半年更新一次，有任何指教與建議，請 email 至 [email protected]。

【歡迎所有 AI / IoT、Blockchain、佈局東南亞的創業者，加入專為你們服務的 AppWorks Accelerator】

Photo by tingyaoh on Pixabay