作者|James
How did a company that started as a voice assistant and C-end consumer goods transform into the first share of AIGC in China? and how did it get the investment of Volkswagen and finally lose Volkswagen?
That's the story revealed in Mobvoi Inc.'s prospectus.
It opened its public offering on April 16-19 and plans to be listed on the main board of the Hong Kong Stock Exchange on April 24 under the stock code "2438". This is after the failure of the submission on May 30, 2023, Mobvoi is once again trying to sprint to the "first share of AIGC", and the goal of financing is to improve its large-scale model capabilities.
The development of Mobvoi can be roughly divided into three stages: C-end hardware, B-end solutions, and large model-related services. Since its development stages are less related to each other, it is difficult to say that the follow-up business has effectively inherited the previous "legacy", and it can be said that the company's current operation and development expectations are also significantly different from when it received early strategic investments from Google, Goertek, Sequoia, SIG, etc.
However, there is a more important thing in the capital market, which is that through its IPO, it can be seen whether the Hong Kong stock market is like the A-share market, and it is fanatically buying the concept of "AIGC".
VOL
Consumer hardware: the strategic contraction of the "Internet entrance".
The first stage of development of Mobvoi was from 2012 to 2018, during which the company developed from the earliest voice assistant app to watches, speakers, and headphones, and gradually completed the layout in the field of consumer electronics and smart hardware. At the same time, the company is also expanding its offline retail channels. If the company continues to develop along this path, it is likely to become a similar enterprise to iFLYTEK or Xiaodu.
However, the asset-heavy investment was considered unsustainable, and the company was also experiencing obstacles in attracting strategic investments, so it decided to make strategic adjustments. In 2019, the company eliminated offline stores and reduced its business to only watches, headphones and car machines, and other product lines were canceled. In the process, some of the executives responsible for growth also left.
Although Mobvoi's watch seems to be a stock business at the moment, it still brings significant revenue to the company. With these watches, the company successfully introduced Google's Wear OS and helped Google solve some localization problems. However, due to the need for localization of Google services, the search uses the voice search of Aski, and the map is Sogou Map, in fact, there is not much "pro-son" component.
Now in the wearable watch market, Huawei and Xiaomi are trying to change the native Wear OS system to be in the same sequence as its mobile phone ROM in the same sequence of "Hongmeng" and "surging", to get rid of the restrictions on native apps and app stores. As a result, Google had to rely on independent third-party branded hardware to promote its native systems, which was a prerequisite for Mobvoi to grow its business back then.
The reward for Ask is to run the Ask Store, the largest independently operated app store for Wear OS. But of course, in fact, most users still mainly use Alipay and WeChat, which are the top watch apps. Some TicWatch users told Shizhi Future that at present, some "wrist XX" applications and watch faces are charged, but basically health detection and official apps such as Alipay and WeChat have met the needs of most people, so it is difficult for similar wishful thinking such as "sending hardware and selling software" to run through in the watch industry.
The strategic investment introduced in 2019 was the company's last round to date, and Mr. Li said in an interview that it was done to ensure the purity of access to financing and avoid accepting unaffordable strings attached.
VOL.2
B-end business: Rely on a single customer, Volkswagen
The second stage was during the epidemic prevention and control period in 2019 and beyond, when the company shifted from C-end to B-end business. Due to the combined impact of the epidemic and other factors, the number of offline physical dealers on Mobvoi has dropped significantly, from 141 in 2020 to 58 in 2022. In 2022, Mobvoi's consumer hardware division revenue fell sharply by 41.7% year-on-year, and the number of inventory turnover days also rose to 242.5 from 159.2 in 2021, with the explanation in the prospectus being that the launch of new flagship products was delayed, while old products entered the later stages of the product life cycle.
Mobvoi's on-board hardware "Ask the Magic Mirror" attracted Volkswagen's investment in 2017. In 2022 and 2023, Volkswagen will be Mobvoi's largest customer, with revenue from Volkswagen accounting for 42.6% of total revenue in 2022. However, on the eve of the company's first submission in 2023, Volkswagen emptied its stake in Mobvoi and split the joint venture into an independent supplier, increasing the uncertainty that Mobvoi will continue to generate revenue from Volkswagen in the future.
In-vehicle voice assistant is a relatively easy function to implement in in-car AI, and car companies can consider allowing the assistant to achieve a higher level of in-car function control, even involving safety, or it can be limited to Siri-like in-car entertainment. The former means that the AI technology of the whole vehicle must be a holistic solution, and it is unlikely that the voice will be taken apart and handed over to another company, while the latter means that the differentiation between different voice suppliers is not so high. Volkswagen and Mobvoi are not exclusive partnerships, and later expanded the range of service providers in some new models, such as the ID.7 in February this year, which uses iFLYTEK's technology.
Mobvoi's other B-end services also include intelligent voice customer service, CRM, offline voice, etc., but if it is an impressive big customer, there is only one Volkswagen at present.
VOL
Large model stage: Making differentiated perceptions is a major challenge
The third stage of Mobvoi, on the whole, is mainly driven by AI and large models. At present, it has stated that it is "a company focused on artificial intelligence technology, providing solutions with generative AI and voice interaction technology at its core".
Compared with the previous accumulation of AI technology, this part of the progress may not be so impressive. Judging from the public information, the launch of the "Sequence Monkey" text model does not particularly use the company's previous technical capabilities, which reflects similar problems to the "AI Four Tigers" such as SenseTime, YITU, Yuncong, and Megvii in the face of the era of large models.
In the future, if a company wants to grow through an asset-light approach, it must focus on maintaining its traditional strengths and finding new points of integration. The products that the company is focusing on promoting are making use of its technical accumulation in sound recognition and generation to a certain extent. It is no wonder that the company's main competitors include Agora and Yunzhisheng, rather than large-scale model companies as we understand them in the conventional sense.
However, Yunzhisheng has withdrawn its IPO application several times before, which may indicate the difficulty of focusing on the track of voice AI. For example, Microsoft just announced last month an update to the cloud TTS service carried by Azure, including realistic capabilities such as air sound and pause, which has been indistinguishable very early. New technologies such as Sovits, which is based on large models, provide an alternative and more affordable and open-source solution. Jianying's sound clone is now also in the stage of rapid user growth.
For the entertainment industry, there is indeed a high demand for audio post-processing and sound generation. At present, although there are several preliminary solutions on the market for technologies such as automatic dialogue generation and sound cloning, the level of technology and availability are not the same. It is true that any new product provides more choices for entertainment practitioners, but for Mobvoi, it has become particularly important to establish the user's mind and expand the number of users as soon as possible.
In the name of the "Sequence Monkey" model, the company has launched the "Wonderful Text" writing, "Words Painting" painting, "Magic Sound Workshop" dubbing and "Wonderful Yuan" digital human, and there is still a long way to go to establish a differentiated user mental cognition.
It must be pointed out that the strategic investment introduced by Mobvoi in the first stage of development, that is, the intelligent hardware stage, was mainly aimed at its effect as the next "Internet entrance", and the investment mentality for 91, Wandoupod, shared bicycles, etc. was similar to that at that time. Now, the company is faced with the proposition of how to make early investors exit reasonably, and the situation that the company is facing now and the selling point it emphasizes are completely different from the beginning. It is rapidly transforming into a company with a robust and predictable business model, shifting its focus to businesses that generate stable cash flows.
However, there is a more important thing in the capital market, which is that through its IPO, it can be seen whether the Hong Kong stock market is like the A-share market, and it is fanatically buying the concept of "AIGC".