iFLYTEK has long been considered the top Artificial Intelligence (AI) voice recognition company in China. Founded in 1999, it has been listed on the Shenzhen Stock Exchange since 2008. iFLYTEK has been a trailblazer in Chinese voice recognition for the last 20 years. In the recent AI hot, iFLYTEK was considered part of the Chinese “AI national team” along with Alibaba, Baidu and Tencent.

Compared with its peers, however, iFLYTEK is much more controversial among the public. In September 2018, it was accused of misleading people into believing its software was translating words automatically at a conference – something which in fact was being done by a human interpreter. A month later, news came out that iFLYTEK was developing buildings in a national nature reserve, thus violating environmental regulations– this, together with the first news mentioned, led to a more than 30% decrease in iFLYTEK’s stock price in late 2018.

The real dilemma of iFLYTEK, right now, is that it does not have an AI tech company image like Alibaba and Baidu. Pushed to the front of the public’s mind by the hot AI concept in recent years, iFLYTEK has shown the image of a company that relies on government support and yet does not have its own popular products or services. Equipped by cutting-edge AI voice recognition technology but not able to fully release it in the consumer market, iFLYTEK is struggling in realizing the potential that people wish it to have.

Riding on the wave of AI era, backed by Beijing 

The dramatic evolution of AI technology is the key to iFLYTEK’s recent rise. Concerns from the market and the Chinese government have brought iFLYTEK ample financial support, but also the suspicion of being overvalued in the stock market.

As a well-known A-share-listed company focusing on AI-related technology, the fast development of AI technology and its commercialization since 2015 have raised interest in iFLYTEK. Since 2015, iFLYTEK has raised more than RMB 5.5 billion (USD 0.8 billion) through private placements, almost once a year.

Meanwhile, developing AI technology is considered a national strategy of China. In recent years the Chinese government has come up with a series of policies to stimulate the AI industry, including various kinds of government purchases and subsidies in which iFLYTEK has partaken. 

iFLYTEK’s revenue from government purchases accounts for a large proportion of its total. In 2018, more than 60% of iFLYTEK’s sales came from these (mainly in education, smart city and the legal sector [Chinese Judiciary]). 

By successfully winning purchase contracts with local governments around China (two examples include contracts of CNY 0.86 billion (USD 0.12 billion) from Qingdao City Western Coast District’s AI+Education project and CNY 1.59 billion (USD 0.23 billion) from Bengbu smart school project, iFLYTEK has become an important e-government solution provider. The concentrated connection with the government has brought iFLYTEK stable income and persistent growth, but keeping its distance from most individual customers has also seen voices claim that iFLYTEK does not have the ‘genetics’ of a true internet company.

iFLYTEK’s high market value looks more doubtful when one considers the high proportion of government subsidy in its net income. The proportion was as high as 30% for years and iFLYTEK does not seem to have a remedy for this. It even rose to 50% in the year 2018 (the full report of 2019 is not out yet), which was seen as a negative signal, showing that iFLYTEK is not guaranteed to have an independent operational and business model any time soon.

B2C for iFLYTEK: weak, adjusting and necessary

iFLYTEK has long been questioned for its weak B2C (Business-to-Consumer) production line. The pain point is that iFLYTEK does not have a final product integrating its signature technologies. Its main product – the Easy Trans Portable Smart Electronic Voice Language Translator Device – lacks application scenarios and, attracting even harsher criticism, it can be replaced by translator apps on phones.

With more B2C products being release – including an AI translator, iFLYTEK input, recording pen, laptop, etc. –  the weakness of iFLYTEK’s consumer products reveals a deep problem: iFLYTEK does not have an applied core technology to support its B2C market.

 iFLYTEK claims to be a “speech intelligence” company on its website. The concept contains speech recognition, speech synthesis and translation. However, it remains quite vague and the connection between each concept is not clearly reflected by iFLYTEK’s product. As a result, all of iFLYTEK’s products have alternatives on the market, and iFLYTEK cannot provide a star product to dominate the market, unlike Apple or Tencent. 

The B2C market is the inevitable home territory for iFLYTEK. A pure solution provider for government departments and business companies can hardly become a real consumer AI tech giant, especially in a service like voice recognition that is not a daily prerequisite for individual customers.  In order to support its USD 12.58 billion market value, iFLYTEK needs a stronger consumer market performance to prove its profitability and potential.

On the positive side, iFLYTEK is making a remarkable effort in its B2C product line. Its B2C revenue has increased continuously since 2016. By expanding the proportion to near 40% of total sales revenue, iFLYTEK can diversify its sources of income and shorten the distance with individual customers, which provides the potential to reduce the significance of government purchase and subsidy at iFLYTEK.

Furthermore, as IT giants such as Google and Baidu have emphasized the importance of voice recognition to their overall strategies, they have also accelerated the research in related fields. The interconnected modern academic ecosystem ensure most research results are available to every participant in the AI voice recognition field, which weakens iFLYTEK’s technological advantages accumulated in the past 20 years. With even more sufficient funds supporting their R&D teams, IT giants that have entered the voice recognition field only recently compared to iFLYTEK are also able to develop mature technology in a short time. 

Recognized research ability, not fully released

Widely recognized as Chinese national champion in voice recognition, iFLYTEK’s research has been widely approved by academic circles. The new challenge for iFLYTEK is to thoroughly integrate them to serve the market, in order to win against the many IT giants who have joined the race.

In April 2019, a model from the joint iFLYTEK Research and HIT (Harbin Institute of Technology) laboratory came 1st at SQuAD2.0 (Stanford Question Answering Dataset), a widely recognized, top-level machine reading comprehension challenge in the field of cognitive intelligence. Meanwhile, researchers from iFLYTEK’s AI research institutes have published a considerable amount of research papers about reading comprehension and speech recognition in top conferences, including the AAAI and EMNLP.

However, being far away from the consumer market obstructs iFLYTEK from fully releasing its technological power. What iFLYTEK needs right now is a product or service that it can bet all its accumulated technology advantage on. iFLYTEK has to prove to the public, with a widely accepted product, that a tech company with voice intelligence as its single focus is able to change people’s lives, and so support its 110 PE ratio.

 With part of the public now believing that its stock price is overvalued, iFLYTEK should quicken it step towards developing B2C products that will boost the confidence of investors.