来源丨创业邦(en:ichuangyebang)
Author丨Barry
Editor丨Guan Ju
Source丨Hong Kong Stock Exchange
On April 24, Mobvoi (2438.HK), an "AI unicorn" that has undergone two submissions, was successfully listed on the Hong Kong Stock Exchange. Mobvoi's IPO was priced at HK$3.8 per share, and as of press time, Mobvoi was trading at HK$3.1 per share, with a total market capitalization of HK$4.624 billion.
As the world's first profitable large-scale model company, the Hong Kong stock IPO of Askou also means that it has become the "first AIGC stock" in 2024.
Mobvoi is a large-scale model company with AIGC and voice interaction technology as the core, mainly with three major business segments: AIGC solutions, AI enterprise solutions, smart devices and accessories, and has launched a multi-modal large-scale model "Sequence Monkey", which is applied to its Moyin Workshop, Wonder Yuan and other products.
In 2012, Mobvoi was founded by Zhifei Li, a former Google scientist and world-class expert in natural language processing and artificial intelligence. According to Smartbeast's analysis, the company has completed a total of 7 rounds of financing, with investors including Sequoia Capital, Zhen Fund, SIG Haina Asia, Google, Volkswagen Group, etc., with a cumulative financing amount of more than 255 million US dollars. In 2017, the company received $180 million in financing, breaking through the $1 billion valuation and becoming a unicorn.
According to the CIC report, in terms of revenue in 2022, the market share of AI speech technology and natural language processing (NLP) software in China is 1.4%, ranking third.
In 2022, the company's AIGC solutions accounted for approximately 0.9% of the global AIGC market share in terms of revenue, ranking 8th in the global market and 1st in Asia in terms of revenue from AIGC products and services in 2022.
Former Google scientist
From "debut is the peak" to "shift acceleration"
Li Zhifei, who was born in an ordinary family in Hunan, has had a lofty dream of becoming a scientist since he was a child.
Since then, he has also achieved his goals as a young man. He has an excellent resume:
In 1999, Li Zhifei left Nanjing University of Science and Technology to study abroad, and in 2004, he entered the Department of Computer Science of Johns Hopkins University in the United States, and entered the world's top speech recognition and natural language processing laboratory of the university.
After graduating, Li Zhifei received offers from Internet giants such as IBM, Microsoft, Google, Yahoo, and Facebook, and finally chose to join Google Translate Lab. During his time at Google, Li Zhifei presided over the development of Google's offline translation system for mobile phones.
With the background of a former Google headquarters scientist and the main developer of Joshua, the world's mainstream machine translation open source software, Li Zhifei was a very scarce talent at that time.
Mobvoi founder and CEO Li Zhifei Source: Mobvoi
In 2012, after receiving angel investment from Sequoia Capital and Zhen Fund, Li Zhifei resigned from Google and returned to China to found Shanghai Yufanzhi Information Technology Co., Ltd., and was determined to create the next generation of mobile voice search products - "Mobvoi".
Li Zhifei once said that early-stage investment institutions such as Zhenge, Sequoia, and SIG "invested in my personal background and my team, as well as some demos we made in technology." ”
Li Zhifei, who advocates Google's engineering culture, makes no secret of his dream - to make a Chinese version of Google Now. After starting his business, Li Zhifei gathered a group of engineers from Stanford, Maryland, Nanjing University, Zhejiang University, Fudan University, Shanghai Jiaotong University and other famous universities in China and the United States, and formed the core founding team of "Mobvoi".
At that time, among the entrepreneurs, Li Zhifei could be described as "the peak of his debut".
For To C consumer-level scenarios, Mobvoi has successively launched Mobvoi mobile app, AI smart watch TicWatch series, AI true wireless smart earphones TicPods series, smart rearview mirror TicMirror, smart speaker TicKasa series and other artificial intelligence software and hardware combination products.
In 2018, the company launched a two-wheel drive strategy of ToC and ToB. In the face of the ToB enterprise market, Mobvoi released the first mass-produced AI voice chip module "Wenxin". In addition, the company announced that it will open 20 new offline smart experience stores nationwide in 2018.
By 2019, the company's scale has approached 1,000 people, including six or seven product lines such as smart watches, smart speakers, smart bracelets, smart wireless headphones, and smart cars. In order to achieve rapid growth, the company's overseas channels are divided into three regions: North America, Europe and Southeast Asia, which are responsible for three complete independent teams.
After the crazy expansion in 2018, in 2019, Mobvoi is facing a big adjustment of "gear shift acceleration".
In the smart speaker track, Mobvoi directly confronts giants such as Xiaomi, Baidu, and Huawei, and has no power to fight back in the face of the giants' subsidy behavior. This is also the first time that Li Zhifei realized that the competition with the giants is very terrible. "This kind of growth is not sustainable, and the more you grow, the worse the loss. ”
First of all, the business line was streamlined, only two product lines of smart watches and smart cars were retained, and all 20 offline stores were closed. Secondly, the number of employees in the company has been reduced from 1,000 to 500.
"Investors only look at the growth in revenue, not at efficiency, and not at how much money they lose. If the company's foundation is not good, once it runs out of money and cannot continue to raise funds, this growth will not be sustainable. ”
In Li Zhifei's view, the company had a technological halo in the early days, and was given unrealistic expectations by investors and partners. By the middle of development, it was found that these cutting-edge technologies, both in terms of technology and commercialization, were very immature. The latter is the embarrassing situation faced by Chinese AI unicorn companies.
Mobvoi, which persevered, finally completed the adjustment in 2021 and embarked on the upward stage of the "U-shaped" curve.
In 2021, smart watches in the To C business accounted for more than 80% of the company's revenue, and to B businesses such as smart cars accounted for less than 20%. Li Zhifei said, "We can basically break even. The shipment of smart watches can reach hundreds of thousands of units, and the demand for the entire overseas market has skyrocketed. ”
Source: Mobvoi
In 2023, Mobvoi will launch a general large model "Sequence Monkey". The model has multimodal generation capabilities, can understand and generate text, audio, images, videos, and 3D content, and supports different tasks such as language generation and speech recognition, providing a solid technical foundation for AI solutions.
In addition to the sequence monkey, Mobvoi has also launched the AI dubbing platform "Magic Sound Workshop", the one-stop digital human production platform "Wonderful Yuan", the enterprise AI interactive digital employee generation platform "Wonderful Question", and the overseas version of the "Magic Sound Workshop" "DupDub".
China's earliest profitable large-scale model company
AIGC has a compound annual growth rate of over 300%
According to the analysis of Smartbeast, since its establishment, Mobvoi has completed a total of 7 rounds of financing, with a cumulative financing amount of more than 255 million US dollars, and investors include Sequoia Capital, Zhen Fund, Google, Yuanmei Optoelectronics, SIG, Goertek, etc.
Before the IPO, Li Zhifei held 26.72% of the shares, co-founder Li Yuanyuan held 3.02% of the shares, and Lei Xin held 3.00% of the shares. Li Zhifei and the other three are acting in concert, holding a total of 32.74% of the shares.
In terms of institutional investors, SIG holds 17.03%, Google holds 13.26%, Sequoia Capital holds 10.92%, Goertek holds 10.03%, Zhen Fund holds 3.07%, and Yu Minhong's Jinmeng Holdings holds 2.47%.
In addition, for the entire FY2023 period, Li Zhifei's total salary (including equity-settled share-based expenses, discretionary bonuses, allowances and benefits in kind, etc.) is as high as RMB6.543 million.
Although the company is back on an upward trajectory, Mobvoi's path to listing has not been smooth.
According to public information, as early as 2017, Li Zhifei said that he planned to seek an IPO within two years, perhaps in the United States or Hong Kong, in an interview in 2018, he said that he planned to list at the end of 2019 or the first half of 2020, and in 2019, it was reported that the company was seeking to raise $100 million in funds and planned to be listed on the Shanghai Stock Exchange's Science and Technology Innovation Board.
But Mobvoi said it would not comment on the financing news. Regarding the listing on the Science and Technology Innovation Board, the company said that "we are interested and have a plan". In the following four years, Mobvoi's listing seemed to be out of place.
Finally in May 2023, Mobvoi submitted its first application to the Hong Kong Stock Exchange, but it ended in the invalidation of the prospectus. The IPO was put on the agenda as a result of the second submission in December last year.
According to the prospectus, from 2021 to 2023, Mobvoi's revenue will be 397 million yuan, 502 million yuan and 507 million yuan respectively, of which the revenue from AI software will increase from 60 million yuan in 2021 to 343 million yuan in 2023, with a compound annual growth rate of 140%, and the corresponding revenue proportion will increase from 15% in 2021 to 67.7% in 2023.
By business, Mobvoi's revenue is mainly composed of AI software (including AIGC solutions and AI enterprise solutions) and AIoT hardware. Among them, Mobvoi AIGC solution accounts for the highest proportion of revenue, and has grown rapidly in recent years.
Among them, the revenue of AIGC solutions from 2021 to 2023 will be 6.822 million yuan, 39.857 million yuan and 118 million yuan respectively, with a compound annual growth rate of more than 300%.
At present, Mobvoi has more than 15 million users around the world, including content creators, enterprise-level customers and consumers. Among them, since 2020, the AIGC solution for content creators has more than 10 million users and about 865,000 paying users.
According to the CIC report, in terms of revenue in 2022, the market share of China's AI speech technology and natural language processing (NLP) software is 1.4%, ranking third. At the same time, Braintalk is also one of the earliest and largest technology companies focusing on generative AI in Asia, one of the few companies in Asia with general-purpose large model capabilities, and a leader and pioneer in the application of AICopilot technology in Asia.
From 2021 to 2023, Mobvoi's gross profit margin will be 37.5%, 67.2% and 64.3% respectively. Although the revenue growth rate has decreased from 2022 to 2023, the gross profit and gross margin in the corresponding years have remained at a high level.
Among them, from 2021 to 2023, the gross profit margin of AI software solutions will always be above 60%, reaching a maximum of 93.8%. Among them, in 2023, the gross profit margin of AIGC solutions will be 92.2%, and that of AI enterprise solutions will be 80.9%. In contrast, the gross profit margin of smart devices and other accessories peaked at 33.1% in 2021 and fell to 21.5% in 2023.
It is worth mentioning that Mobvoi has turned losses into profits in 2022, with adjusted net profits of -73 million, 109 million and 18 million yuan from 2021 to 2023, respectively. In other words, Mobvoi has been, and probably is, the only large-scale startup that has achieved profitability.
Although net profit declined slightly, this was mainly due to Mobvoi's continuous investment in R&D.
From 2021 to 2023, Mobvoi's R&D expenses will be 92 million yuan, 119 million yuan and 155 million yuan respectively. The company emphasized that due to the continuous technological advancement of the AI industry in which it operates, significant resources, including financial and human resources, are required to keep abreast of the latest developments in the AI industry, so that the company can expand its supply range and enhance the competitiveness of its solutions in the market. As a result, the company expects to continue to invest heavily in R&D.
According to the prospectus, the total funds raised in the Mobvoi IPO will be mainly used for the development and expansion of the scale of the multi-modal large model "sequence monkey" in the next three years, solution development and marketing, seeking potential strategic alliances, investments and acquisitions, as well as working capital and general corporate purposes.
Six reflections that AI entrepreneurs have to pay attention to
Regarding the ups and downs of the entrepreneurial experience, Li Zhifei had an exclusive conversation with Entrepreneur in 2021, attaching several of his reflections, hoping to inspire AI entrepreneurs:
1. Entrepreneurs need to be masters of problem solving
Entrepreneurship is like a problem, you need to define what the topic is, what kind of solutions to provide for which users, what pain points to solve, and what the final business model is. Technology entrepreneurs have a better understanding of the boundaries of technology, and finding a problem can form a barrier in the process of solving the problem.
The vast majority of technology entrepreneurs are not masters of problem solving, but problem solving. Technology is a process from left to right or from beginning to end, and the problem is a business-driven process that starts with the end in reverse, which is naturally conflicting.
2. The biggest problem of AI companies is fat, and the higher the valuation, the more painful it is
At present, AI companies generally enter the second half, and everyone may be able to brag in the first half, but in the second half, because the endgame is about to be seen, the result will be very cruel.
Today's AI companies are inflated in terms of valuation, team size, and external commitments. This is an abnormal development path, lacking down-to-earth development. AI companies need to have a suitable cost structure, strong project implementation capabilities, and scenario closed-loop capabilities. Many well-known AI companies have not experienced such a normal development path.
3. Technology companies need to create a business model with barriers to survive independently
Technology companies need to find sustainable rigid needs and use technology to create a sustainable and scalable business model with barriers. Unless it can always rely on financing or being acquired, in order to become a company that can survive independently, it needs to form a self-closed loop.
4. You can't make a technology company a money-losing outsourcing company
A lot of tech companies have turned themselves into money-losing outsourcing companies.
Pure software outsourcing companies do not lose money, but in the past, many AI companies had a big problem that the cost of 100 yuan was only 60 yuan because of competition, which AI companies must avoid. Instead of being a money-losing outsourcing company, you should be a product company where R&D costs can be marginalized.
5. There is bound to be a lot of trial and error in technology companies
If 10 out of 100 technologies can succeed, the contribution to human science and technology is already very good.
The problem of technology companies is to find nails with hammers, and when doing technological innovation, they can't just make hammers, and they don't care where the nails are, what is needed is some thinking that starts with the end in mind. But if you just stare at the nail and then look for the hammer, that's not the tech company's model either. Technology companies must have a lot of trial and error, otherwise they will not be doing cutting-edge technology, but becoming operating companies.
6. Financing for the sake of financing should not be done
There is no point in financing for the sake of financing, on the contrary it will be painful.
When capital comes in, it will magnify the true character of the company. If the company is good and the business model is clear, capital will play a positive role.
If the company does not have a sustainable business model, the cost of blind growth will eventually be paid. The input of technology does not necessarily have output, and in many cases, there is no return.