The AI medical track is crowded with major tech giants.

In the quarter that is about to pass, ByteDance has made two significant moves in AI healthcare.

From establishing its own hospital on Douyin to initiating the spin - off and independent financing of its AI pharmaceutical business line, this leading tech giant's strategic investments in the healthcare and wellness sector have sent shockwaves through the entire internet industry.

In fact, internet giants have never stopped investing in healthcare. Especially in the context of AI, a new round of equipment competition has been launched.

In the past six months, JD Health has upgraded its AI doctor agent cluster, Tencent Health has released an AI full - stack solution, Ant Aifu has successively launched new services, and Alibaba has released the medical AI "Hydrogen Ion" for doctors...

It's just that the footholds of these giants in AI healthcare differ. They usually choose to bet on one or more of the user side, doctor side, hospital, and institutional side.

Behind the differences in route selection are the different genes and aces of each company. Most of them already had a foundation in internet healthcare before.

Compared with internet healthcare a decade ago, the changes that AI brings to healthcare may be far more profound than we imagined.

A Hundred Flowers in Bloom

Investing 6 billion yuan to build its own hospital is ByteDance's largest single investment in offline healthcare.

Different from traditional hospitals, this hospital is positioned as a "smart AI hospital". Relying on the core advantage of "AI + digitalization", it will achieve full - scenario intelligence from pre - examination, consultation, inspection, treatment to follow - up visits through medical services such as AI - assisted diagnosis, intelligent diagnosis and treatment decision - making, and digital medical record management.

Moreover, the hospital focuses on international healthcare. It plans to introduce 1,600 international talents and lay out key investments in cutting - edge medical fields such as high - end comprehensive diagnosis and treatment, specialty departments, and cell and gene therapy. The industry generally believes that this will help fill the supply gap of high - end international comprehensive healthcare in Beijing.

Also focusing on offline healthcare, the Xiaohe Clinic established by ByteDance in Shanghai in February this year focuses more on convenient basic diagnosis and treatment services. That is, catering to the daily medical needs of the general public, it enables the mature Xiaohe AI healthcare technologies, such as intelligent consultation and AI disease analysis, to be applied in real offline medical scenarios.

The online layout has also continued.

As early as July 2025, ByteDance launched the AI healthcare APP "Xiaohe AI Doctor", which is positioned as an all - weather health butler for C - end users. When users search for health issues on Douyin or click on the consultation entry under medical science popularization videos, they will be guided to the Xiaohe AI Doctor for consultation.

In addition to physical healthcare and online consultations, ByteDance is also betting on AI drug R & D.

The recently spun - off business line originated from the AI pharmaceutical team established in 2021, which has been responsible for the core functions from basic model research to industrialization since its establishment. According to industry insiders, this spin - off is ByteDance's first attempt at the industrialization of AI4S.

In contrast, Ant's Aifu has a less intensive deployment in AI healthcare. Currently, it mainly serves users through the Aifu App. However, since its name change at the end of last year, it has been very active in the past six months.

In May, Aifu started offering free interpretations of physical examination reports to 100 million people. On June 15th, Aifu launched a new service called "Doctor Review". After users use AI to inquire about skin problems, they can choose to invite doctors from top - tier hospitals to review the results and provide additional advice. One week later, it got involved in national weight - loss management.

Another important player in the AI healthcare track is JD Health. On the basis of its existing businesses in drug retail and online consultations, JD Health is trying to integrate AI into all scenarios of its health business.

At the beginning of this year, JD Health launched two products: "Zhiyi" for practicing physicians and "Zhuoyi 2.0" for hospitals and institutions.

In April, for the medical device industry, it signed the JoyInside access agreement with 10 medical device brands and plans to connect 1 million devices within a year. JoyInside is an intelligent brand launched by JD, which can equip traditional medical devices with an "AI brain".

Two months later, for the user side, JD upgraded its "AI Jingyi" agent cluster, including AI doctors, AI nutritionists, AI pharmacists, and AI psychological counselors.

Tencent has taken a different path.

On June 5th, Tencent Health officially released an AI full - stack solution for users, medical institutions, pharmaceutical companies, and research institutions.

For example, for hospitals, the Tencent Health open platform provides six AI capabilities such as intelligent guidance and intelligent Q&A, covering the entire process before, during, and after diagnosis. Doctors can flexibly combine these capabilities to quickly build their own medical agent services. For users, the mini - program uses products such as the health management assistant and AI health inquiries to make AI a daily available personal health assistant for users.

The AI healthcare track is crowded with internet giants.

Different Paths

Internet giants are competing to gain a foothold in the AI healthcare track, but their layout paths are different. Behind this, it reflects the differences in their past advantages.

ByteDance is taking a relatively radical heavy - asset path: building a closed - loop of "AI + healthcare" from content to diagnosis and treatment.

Algorithms, traffic, and content ecosystems have always been ByteDance's core advantages. Specifically in the healthcare field, according to a set of data released by Hong Xiaowu, the director of ByteDance's healthcare and wellness department, 90.5% of the 934 million Douyin users have behavior of obtaining medical information. Among them, 300 million people have watched medical science popularization content, there are more than 50,000 certified doctor accounts on the platform, and the annual number of likes for medical - related content exceeds 3 billion.

This means that when users search for health issues on Douyin or click on the consultation entry under medical science popularization videos, they will be guided to the Xiaohe AI Doctor for consultation.

But ByteDance's ambition is not limited to the online world.

As early as 2021, it opened its first Xiaohe Clinic in Beijing. In 2022, it spent 10 billion yuan to acquire the high - end private women's and children's hospital Mei Zhong Yi He in full and gained over 90% control of Beijing Mei Zhong Ai Rui Cancer Hospital. In the first half of 2026, it reopened the Xiaohe Clinic in Shanghai and invested 6 billion yuan to build an international AI hospital in Beijing.

Obviously, ByteDance wants to replicate a Douyin - style healthcare closed - loop. Start with content, use AI services for conversion, and finally complete the delivery of healthcare services offline.

While ByteDance tends to open up the entire industrial chain through heavy - asset layout, Tencent has taken a completely opposite path.

Wu Wenda, the president of Tencent Health and the head of Tencent's Life Science Laboratory, mentioned in a recent interview that WeChat has 1.4 billion active users and 900 million medical insurance users, and most of the top - tier and third - level hospitals in the country have settled in. Users are already used to registering, checking reports, and handling medical insurance business on WeChat. Such infrastructure allows Tencent to naturally stand in the position of "connection" and no longer needs to build hospitals, pharmacies, or internet hospitals with heavy assets.

Wu Wenda said, "If Tencent were to create a complete closed - loop by itself, it would stand against all industry partners, which goes against our original intention of building an open healthcare ecosystem."

Based on this gene, Tencent has chosen a path of open connection and lightweight empowerment. The core is to help various medical and pharmaceutical institutions make digital connections and improve work efficiency.

Take the compliance system of pharmaceutical companies as an example. Pharmaceutical companies hold tens of thousands of academic conferences every year, and it is impossible to fully cover them with manual post - event spot checks alone. With the help of AI, Tencent can help pharmaceutical companies achieve full - process intelligent control before, during, and after the events.

Therefore, even though Tencent's smart healthcare plan is very ambitious, covering many aspects such as medicine, medical institutions, payment systems, and individuals, it still firmly adheres to its boundaries and only acts as a digital assistant. In Wu Wenda's view, choosing what not to do is more important than choosing what to do.

Pharmaceutical e - commerce retail is the starting point and foundation of JD's healthcare business.

A set of data can confirm this: In the fiscal year 2025, JD Health's total revenue was 73.441 billion yuan, a year - on - year increase of 26.3%. Among them, the product sales revenue was 60.885 billion yuan, accounting for 82.9%.

Starting from drug retail, JD Health has gradually extended to health services such as online consultations and home - based rapid tests, and finally opened up a closed - loop of the entire scenario of "medical examination, diagnosis, and medicine".

With the support of AI, on the premise of keeping its original core business unchanged, JD Health wants to make every link on the supply side more intelligent and efficient. For example, "Zhiyi", which helps doctors quickly obtain evidence - based evidence, and the AI doctor "Dawei", which provides all - weather health consultations for patients.

From Internet Healthcare to AI Healthcare

The competition among giants in AI healthcare is a real - world trend, and a more important question has also emerged: What are the differences between the internet healthcare that everyone has been talking about in the past and this round of AI healthcare?

Internet healthcare emerged more than a decade ago. Up to now, the industry pattern has basically stabilized:

WeChat and Alipay, with their payment portals, have become important infrastructure for connecting hospitals and patients in services such as registration, medical insurance, and payment. The foundation of Alibaba Health and JD Health is online drug retail, and on this basis, services such as online consultations have been extended.

However, the progress of internet healthcare always seems to be lacking. The real problems faced by the healthcare industry, such as the structural imbalance and insufficient supply of medical resources, have not been given particularly ideal solutions by the internet.

Internet platforms require high frequency and high retention, but seeing a doctor is a low - frequency demand. After 12 years of entrepreneurship, Wang Shirui, the founder and CEO of Future Doctor, said in an interview with "LatePost" that the real bottleneck in healthcare is the insufficient supply, not the connection. The number of high - level doctors is limited, and one doctor can only see 20 patients a day. "Matching and connecting do not solve the problem. The platform does not treat patients, but treating patients is what matters."

A set of data released by the National Health Commission also confirms this: In 2024, tertiary hospitals in China accounted for 3.8% of the total number of medical and health institutions, but they undertook 28.3% of the national diagnosis and treatment volume.

Compared with internet healthcare, the biggest change in today's AI healthcare is that it aims to solve the problem of insufficient supply.

Take ByteDance's Xiaohe AI Doctor and Ant's Aifu for the C - end as examples. Although they do not try to replace doctors, they create an incremental supply by meeting a large number of users' health needs between the hospital and daily life, such as micro - symptom inquiries and physical examination report interpretations. For the D - end, medical vertical large models such as Hydrogen Ion and Zhiyi can reduce the time cost for doctors to obtain evidence - based evidence and help doctors improve efficiency in a single aspect.

There are also further explorations, such as the Future Doctor aiming to achieve medical AGI.

The envisioned model is as follows: First, AI conducts a structured consultation with the patient and provides a diagnosis plan, and then the doctor determines whether it is correct and signs for confirmation. Wang Shirui said that by putting AI in the front, "when seeing 100 patients, the workload for experts may be 1% of the previous level, and the supply will be 100 times that of before."

In this direction, AI may bring greater imagination to the healthcare and wellness field.

Nevertheless, according to the views recently expressed by Wu Wenda, the current application of medical AI is still in the initial exploration stage, and there is still a long way to go before it can be deeply integrated into the core business process.

As "LatePost" mentioned in an article, the healthcare industry is in a contradictory state. The number of AI projects launched in domestic hospitals has increased rapidly, but the AI that can be used smoothly, promoted widely, and continuously evolved is still rare. "Chimney - style" single - point AI models, such as the lung nodule recognition system in the radiology department and the AI review of ultrasound reports in the physical examination center, can effectively solve individual small problems, but they cannot solve the underlying problems across departments, hospitals, or even the industry.

The entire industry is also constantly discussing two core controversies. One is whether AI can truly be implemented to solve clinical pain points. The other is whether the development of medical technology should take the path of a large - scale and all - inclusive industrial closed - loop or a small - scale and refined vertical deep - diving route.

In response to this, Wu Wenda gave an answer that may inspire everyone. "The infinite possibilities of artificial intelligence do not lie in replacing humans, but in clarifying the core boundaries between the technical tool attribute and the medical human - centered attribute through human - machine collaboration and reconstructing the underlying logic of healthcare services."

This article is from the WeChat public account “New Daybreak” (ID: new - daybreak), written by Lu Yuqing and published by 36Kr with authorization.