Chinese enterprises are relying on AI to turn the tables in the high-end hearing aid market previously monopolized by foreign companies.

Text | Hu Xiangyun

Editor | Hai Ruojing

Currently, the upgrade of some medical products is reaping the benefits of DeepSeek.

The hearing aid is one of them.

Although it has been in use for decades, it is still not easy for the elderly to hear their family members' soft whispers in a noisy vegetable market or for those with hearing impairment to hear the birdsong in a corner of the park through a hearing aid.

As a product highly dependent on algorithms and circuits, hearing aids have a long history of applying algorithms such as machine learning. This time, open - source large models like DeepSeek have given relevant manufacturers "a new opportunity to solve the problems in the past R & D process of hearing aids by accessing world - class large models at low cost and leveraging the unique technical perspective of cloud - based large models," said Wang Fa, President of the Hearing Aid Division of United Imaging Microelectronics.

Globally, the competitive landscape of hearing aids has always been quite stable. Five established hearing aid manufacturers such as Sonova from Switzerland, Demant from Denmark, and Starkey from the United States monopolize more than 90% of the sales market. In recent years, although there are many emerging technology companies and medical enterprises focusing on hearing aid R & D, they generally face technical challenges such as difficulty in breaking through in core algorithms and chip monopoly.

So, riding on the wave of large models and endowing hearing aids with stronger artificial intelligence attributes, will it open up a breakthrough path for latecomers?

AI Enables "Directional Listening" in Complex Environments

Compared with other consumer - medical products, the overall penetration rate of hearing aids and the actual benefits for patients wearing them are not very high. A previous survey by the Population Research Institute of Peking University showed that 90% of elderly people with hearing loss were reluctant to have their hearing checked voluntarily even when they knew their hearing was declining; among those who wore hearing aids, only 6.5% were satisfied with the results.

Why are people reluctant to wear hearing aids? And why do those who buy hearing aids not want to wear them?

Liang You currently works in a top - tier hospital in Beijing and has more than 20 years of experience in the hearing aid field. In his experience, apart from the fact that many hospitals neglected the construction of otology departments in the early years and the public awareness was low, the insufficient effectiveness of traditional hearing aid devices was also a problem.

Objectively speaking, the current products on the market are much better than those in the 1990s. However, in a noisy environment, many hearing aid products still lack the ability to recognize sounds, which makes many elderly people feel that wearing a hearing aid is still useless and they can't hear clearly, so they only wear it a few times after buying it," Liang You said.

Some young people who want to or have bought hearing aids for their elders also told us that as soon as they proposed the idea, the elders' first reaction was often "this thing doesn't work," or they felt there was a lot of noise when using it and threw it aside after wearing it a few times.

In essence, this "poor usability" experience points to the classic "cocktail party effect" problem in the field of speech recognition technology during the R & D process of hearing aids, that is, how to distinguish the voice of a specific person when there are many different sound sources in an environment.

Wang Fa introduced that in the past few decades, hearing aid products have always used digital signal processing algorithms. These algorithms make many "assumptions" when dealing with noise reduction problems. For example, they assume that human voices need to be retained and background sounds need to be removed in a certain environment. In a noisy communication environment, the so - called background sounds may also include human voices, which become part of what needs to be retained in traditional algorithms. So, when two competing voices reach the wearer's ears, the problem of "not being able to hear clearly" naturally occurs.

"The industry has always been making difficult trade - offs between 'noise elimination' and'speech detail preservation,'" Wang Fa explained. "This is also the core technical focus of current global hearing aid manufacturers."

It was not until the past two years that a clear direction recognized by both the five established overseas hearing aid manufacturers (hereinafter referred to as "the Big Five") and the new domestic hearing aid forces to solve this technical problem is to introduce deep neural networks (DNN, a machine - learning technology commonly used to process complex data relationships that can automatically learn and extract patterns from data by simulating the structure and function of human brain neurons), enabling hearing aids to break away from the original assumption programs, learn and fine - tune from the massive real sounds heard every day, and thus improve their hearing recognition ability in different environments.

This is also the part that can bring the "greatest benefits" to people with hearing loss.

"Although the functional designs such as product volume, power, and battery life are also points of technical competition, they are not as direct as solving the problem of directional listening in a noisy environment," Wang Fa believes.

In addition, as another "algorithm in neural networks that can truly be implemented on hearing aids," after the deployment of LLM represented by DeepSeek, it can conduct adaptive learning and adaptation according to the user's environment, and complete self - fitting, intelligent fitting and other tasks.

And the product upgrade experience brought by this AI algorithm has also become the content that major hearing aid manufacturers have begun to emphasize frequently recently.

At the end of February this year, Phonak, a brand under Sonova, officially launched in China an AI super - computing hearing aid product equipped with a deep neural network processing algorithm, achieving a noise reduction effect of 10 decibels in a complex environment. Compared with the industry average, the noise reduction of hearing aids is generally around 4 - 6 decibels; a little earlier, the core product of Starkey, the Genesis AI series of hearing aids, also announced an update, equipped with a DNN processor that can automatically detect the surrounding environment by simulating the central auditory system of the brain.

Among domestic brands, United Imaging's uOrigin uses a self - developed heterogeneous six - core chip and an edge - side neural network algorithm, which can process sound signals 750 times per second and has functions such as intelligent noise reduction and directional sound pickup. In addition, some emerging hearing aid brands favored by capital, such as Tingxiang Technology, also take deep neural network technology as the basic theory for product R & D.

Schematic diagram of hearing aid products

It is worth mentioning that in addition to medical companies, tech giants have also entered the field: Apple has released AirPods Pro 2 with hearing test functions, and Tencent, iFlytek, etc. have entered the mid - end hearing aid market at the thousand - yuan level.

"In wearable health products, tech companies have almost exhausted the entry - level functions they can offer, such as heart rate, blood sugar monitoring, and sleep quality monitoring. Hearing aids have certain similarities with traditional true wireless Bluetooth earphones in technology, so it is a natural choice for hardware giants to enter the market," said an engineer who once worked for a leading overseas tech company.

Popularization of Consumer - Grade Hearing Aids

Can Latecomers "Overtake on a Curve"?

When AI slightly narrows the starting line of technological competition, will the "new generation" forces in the hearing aid field get closer to the goal of catching up with "the Big Five"?

First of all, an objective advantage may be that after multiple rounds of market education by predecessors, the public's understanding of hearing aid products is relatively more mature today. For example, some e - commerce insiders have observed that the number of consumers buying low - cost sound - amplifying analog machines is decreasing, and more and more people are willing to choose professional hearing aid products that require fitting.

While emerging hearing aid manufacturers provide more choices for the market, they also promote the popularization of the concept and products of consumer - grade hearing aids (OTC). By lowering the threshold of hearing aid fitting services, the overall cost and terminal price of products are reduced, and then greater sales are achieved to realize "overtaking on a curve".

The reason why this logic holds is that we need to first return to the issue of the revenue model of hearing aid products. The difference between medical - grade and consumer - grade hearing aids lies in whether professional assistance is required for fitting. The former usually needs to be purchased at offline hearing aid stores, and the fitter will adjust the product multiple times according to the user's personal situation, and the process can last from three months to half a year; while the latter can be directly purchased and used online by oneself.

In terms of price, based on the experience of practitioners, 4000 yuan can be roughly regarded as the dividing line. For example, the terminal price of medical - grade hearing aids within the "Big Five" system is often between 5000 - 40000 yuan (there may be discounts on actual transactions), and the price range of leading domestic brands is roughly 5000 - 20000 yuan; while the price of consumer - grade hearing aids is generally between 1500 - 4000 yuan.

The reason for such a large price difference is partly due to the cost of manual services. "Compared with consumer - grade hearing aids, medical - grade products have higher costs because they match more advanced algorithms and applications. But to be honest, the cost difference will not lead to a ten - fold difference in the terminal price. It is mainly to reflect the added value of services. Although there are voices in the industry suggesting that products and services can be charged separately, it is difficult for manufacturers because the domestic acceptance of medical service fees is not high," the aforementioned engineer believes.

In the manufacturers' vision, consumer - grade hearing aids that save this part of the "cost" may be able to seize a part of the market through a low - price strategy at the thousand - yuan level.

For example, iFlytek Medical once mentioned in its prospectus that in the first half of 2022 and the first half of 2023, the revenue from intelligent hardware was about 6.2 million yuan and 29 million yuan respectively. One of the variables during this period was that the company launched hearing aid products in May 2022. As of the beginning of this year, the registered users of iFlytek's intelligent hearing aids have exceeded 110,000.

However, this does not mean that taking the "consumer - grade" route is foolproof. Liang You said that the ability to purchase and fit hearing aids independently does not mean that hearing aid manufacturers can directly dispense with the configuration of professionals. "Since it is a hearing aid product, corresponding after - sales services need to be provided to ensure that the product will not cause hearing damage to patients during use; at the same time, when users have needs, merchants should be able to provide online remote support."

Wang Fa also believes that in the future, consumer - grade hearing aids will "definitely become an important part of the hearing aid market." However, from the existing product cases, although they can undertake some functions from hearing testing, fitting to actual use, "there is still a long way to go before they can become virtual fitters to assist in completing all relevant services."

After all, although hearing aids have a "medical" attribute, purchase decisions highly depend on consumers' personal preferences and needs. In the short term, the market pattern dominated by "the Big Five" may be difficult to shake. But from a more long - term perspective, "soft" conditions such as product price, the development team's understanding of the Chinese context, and the quality of after - sales service are also the keys to competition.

For domestic latecomers in the hearing aid industry, these may all become opportunities for them to turn the tables.

(Liang You is a pseudonym)