Internet giants are pouring into AI drug discovery: ByteDance, Baidu, Tencent, Alibaba and JD have made different choices
On June 10th, a piece of news sent shockwaves through both the technology and pharmaceutical industries simultaneously: ByteDance officially initiated the process of splitting and independent financing for its AI pharmaceutical business line. The core team, core algorithms, technology platform, and existing pipeline assets were all integrated into a new entity, with ByteDance still holding a controlling stake. The new company will be led by ByteDance's AI pharmaceutical team.
Previously, Alibaba Health officially launched its strategic medical AI product, "Hydrogen Ion"; JD Health announced that the number of users of its AI doctor, "Dawei", had exceeded 500,000; BioMap, strongly supported by Baidu, secretly submitted an application for listing on the Hong Kong Stock Exchange; Tencent publicly disclosed a patent for an AI - designed GLP - 1 weight - loss drug.
Five Chinese Internet giants have each chosen very different paths at the intersection of AI and the pharmaceutical industry.
01
ByteDance: From Models to Pipelines, the Heaviest Bet on AI4S
Among all the five giants, ByteDance has the most comprehensive layout.
It didn't take the lightweight platform licensing route or the scattered investment route. Instead, it started from underlying model research, gradually moved towards self - research of drug pipelines, and finally accelerated industrialization in the form of independent financing for a new company.
ByteDance's AI pharmaceutical team was established in 2021 and is led by Liu Kai. The core members, about 50 in number, consist of AI4S algorithm experts and senior pharmaceutical experts. Previously, the relevant team within ByteDance responsible for protein structure prediction models has been fully incorporated, and the algorithm model team has completed integration.
This configuration means that ByteDance is not a casual participant "temporarily mobilizing resources", but is promoting AI pharmaceutical as a strategic business with an institutionalized team in the long term.
After the split, the new company will be led by ByteDance's AI pharmaceutical team and will continue to receive computing power support from Volcengine. From a business structure perspective, this means that ByteDance retains the future flexibility of its core assets while supplementing resources through external financing to advance R & D to the more critical clinical stage.
ByteDance's layout in the pharmaceutical field is not limited to AI pharmaceuticals. In July 2025, ByteDance invested 6 billion yuan to build the Beijing Airui International Medical Complex, planning for 800 beds. The goal is to create China's first "AI - native" physical hospital, integrating clinical data collection, AI - assisted diagnosis, and cell therapy transformation into its own closed - loop system.
AI pharmaceuticals research and develop drugs, and the AI - native hospital verifies the efficacy of drugs. ByteDance's ambition is to have full - link autonomous control from algorithms to clinical applications and establish a complete data flywheel.
02
Baidu: Platform First, Incubating BioMap to Compete with DeepMind
Baidu's choice forms an interesting mirror image compared to ByteDance.
If ByteDance is "starting from models and making drugs on its own", Baidu is "centering on the platform, incubating an independent company, and letting it serve the entire industry".
BioMap was founded with the support of Baidu and personally led by Robin Li. In August 2020, Robin Li and Liu Wei, the former CEO of Baidu Ventures, jointly founded BioMap, with Robin Li serving as the chairman.
The company doesn't make drugs but develops underlying large models for the industry, providing infrastructure - level capabilities for global pharmaceutical companies and research institutions. Its core products include the life - science large model xTrimoV4 with 268 billion parameters and the industry solution BioMapOS.
Currently, the platform has been verified in more than 60 proof - of - concept projects, with application scenarios covering target discovery, antibody R & D, and innovative drug R & D, serving more than 800 institutions, including more than 30 leading enterprises.
The most representative cooperation case is the $1 billion cooperation with Sanofi: BioMap provides the life - science large model (the general AI foundation), and Sanofi provides data and drug development experience. The cooperation goal is to design antibodies and protein drugs.
This model is highly similar to the routes of overseas companies such as NVIDIA BioNeMo and Google DeepMind AlphaFold. The core logic is "to make AI a toolbox for pharmaceutical companies".
Recognition from the capital market for BioMap's model has also arrived. In March 2026, BioMap secretly submitted an application for listing on the Hong Kong Stock Exchange, planning to raise hundreds of millions of dollars. China International Capital Corporation, Morgan Stanley, and UBS are assisting in the preparations.
This means that once the IPO is successful, Baidu's strategic layout will receive continuous capital injection from the independent capital market and enter a self - driven growth track.
03
Tencent: Innovative Drugs Become the New Hunt for the Internet Empire
Among the five companies, Tencent has the most unique layout path.
It doesn't develop underlying models, doesn't conduct self - research on drugs, and doesn't even directly operate medical services. Instead, it uses its huge capital as a lever to conduct a "carpet - style" investment in the pharmaceutical field.
Tencent's investment in the healthcare field has undergone three shifts. In 2014, it invested in Dingxiangyuan, marking the start of the "Internet healthcare" investment stage; after the "930" transformation in 2018, it shifted to the industrial Internet, expanding its investment direction to fields such as medical big data and AI - assisted diagnosis; in recent years, it has made a third shift, extending its investment reach to the most high - barrier field of innovative drug R & D.
In 2024, Tencent made only 22 investment moves throughout the year, but 8 of them were in the pharmaceutical field, covering tracks such as innovative drug R & D, ultrasound imaging, and early - screening detection.
From 2025 to 2026, Tencent continued to increase its investment: it invested in Libang Pharmaceutical and became its second - largest shareholder; it invested in several biotech companies such as Weilizhibo, Minwei Biology, Fanli Biology, Hongxin Biology, and T - Therapeutics. In the cell therapy track, Tencent bet on Xinjing Zhiyuan, the most technology - oriented upstream company, whose core pipeline NW - 101C is the first domestic TCR - T candidate drug targeting PRAME to enter clinical trials.
Most notably, Tencent is no longer satisfied with being just a financial investor. In February 2026, Tencent publicly disclosed a patent for a new short - sequence polypeptide GLP - 1 weight - loss drug completely designed by AI: for the first time, it has reached the forefront of molecular design, moving from "investing in pharmaceutical companies" to the uncharted territory of "developing new drugs".
Tencent has quietly built a huge "pharmaceutical empire" over the past decade, and its layout has shifted from "connecting everything" to "investing in all companies that may change the pharmaceutical landscape".
04
Alibaba and JD: When Pharmaceutical E - commerce Giants Embrace AI Large Models
If ByteDance, Baidu, and Tencent are cutting into "pharmaceutical manufacturing" in different ways, then the choices of Alibaba and JD are closer to their respective core genes: centering on pharmaceutical distribution and services, using AI to upgrade platform capabilities, and secretly cutting into cutting - edge tracks such as cell therapy.
Alibaba's path is a typical "platform empowerment" approach.
In January 2026, Alibaba Health launched the medical - specific large model "Hydrogen Ion", targeting the group of clinical doctors and highlighting its "low hallucination, high evidence - based" capabilities. All answers have authoritative sources and support one - click traceability. To fulfill the promise of "the AI assistant with the lowest hallucination rate in the medical field", "Hydrogen Ion" has built a four - layer evidence - based AI architecture, forming a closed - loop from evidence understanding, retrieval enhancement, model fine - tuning to expert review, and established a "Medical AI Expert Committee" composed of more than 300 Chinese clinical experts.
In the field of cell therapy, Alibaba chose the distribution interface closest to transactions. In April 2026, Alibaba Health reached a cooperation with the promotion party of the first approved stem - cell drug in China, "Aimeimaituosai Injection", and realized full - process traceability based on blockchain. Without building physical facilities or engaging in upstream technology, it cut into the distribution link of stem - cell drugs with e - commerce - level supply - chain management capabilities and traceability technology.
JD Health's path is the most practical.
It has built an AI health - service matrix. The "Jingyi Qianxun 2.0" large model has formed a mature technical architecture. The "AI Jingyi" system has launched more than 1,500 intelligent agents of expert doctors. The AI doctor "Dawei" provides services close to the public. The industry's first hospital - full - scenario large - model product, "JD Zhuoyi", has been implemented in multiple hospitals, serving more than 5 million patients in total.
In addition, its evidence - based medicine AI "Zhiyi", specially developed for doctors, deeply integrates tens of millions of global authoritative medical literatures and guidelines, aiming to become the "intelligent external brain" of doctors.
In the field of cell therapy, JD relies on its pharmaceutical cold - chain logistics system to build a health - asset management system starting from cell storage. Different from Alibaba's "distribution empowerment", JD chooses to cut in from the storage and service end to build user stickiness centered on health assets.
05
Collective Moves in the Cell Therapy Track
An interesting phenomenon is that in 2026, four Internet giants, Alibaba, Tencent, JD, and ByteDance, rarely stepped in unison and collectively made heavy investments in the cell therapy track.
The driving force behind this collective shift is the "Regulations on the Clinical Research and Clinical Transformation and Application of New Biomedical Technologies" issued by the State Council as Decree No. 818, which officially came into effect on May 1st, 2026. For the first time, it legally opened up the path for cell therapy from "clinical research" to "transformation and application" through administrative regulations.
When regulations inject certainty into the industry, the four Internet giants cut in with very different logics: ByteDance builds a closed - loop clinical trial capacity through its self - built AI - native hospital; Tencent seizes the upstream technology high - ground by investing in TCR - T technology companies; Alibaba cuts into the distribution of stem - cell drugs through supply - chain management; JD builds a health - asset management platform through cell storage.
This collective shift reveals a deeper fact: regardless of how the five giants diverge in their AI pharmaceutical paths, they all regard "medical and health" as one of the most core strategic high - grounds in the second half of the Internet era. And AI is their common weapon to capture this high - ground.
06
Conclusion
The different choices of the five Internet giants in the field of AI pharmaceuticals are essentially the projection of their respective core genes in the pharmaceutical track:
ByteDance's gene of "emphasizing content and closed - loop" is projected as the "full - link control" model of self - research drug pipelines and AI - native hospitals; Baidu's "AI - first" gene is projected as the "platform - based empowerment" model of BioMap's life - science large model;
Tencent's gene of "connecting everything and investing as a layout" is projected as extensive investments in AI pharmaceutical and biotech companies; Alibaba's gene of "platform economy and empowering merchants" is projected as the "empowerer" model of AI doctor assistants and cell - therapy distribution systems; JD's gene of "supply - chain and serving entities" is projected as the "service provider" model of AI health services and pharmaceutical cold - chain logistics.
However, beneath the surface of the diverging paths, two common trends are emerging.
Firstly, the upgrade from "investing in tracks" to "building moats". None of the five companies is satisfied with staying at the digital level but is penetrating deeper into the industrial chain: ByteDance conducts self - research on drug pipelines, Baidu enters the underlying large models, Tencent cuts into molecular design, and Alibaba and JD upgrade their supply - chain and service systems.
Secondly, AI is just a means, and data is the ultimate goal. ByteDance needs clinical data to verify its AI pharmaceutical models; Baidu needs pharmaceutical company data to train large models; Tencent needs data feedback from its investment assets; Alibaba and JD need prescription and health data to optimize recommendations and supply chains. Whoever masters more and higher - quality life - science data will be able to build real barriers in the next round of competition.
This track is changing from a playground for startups to a main battlefield for technology giants. And the five Chinese Internet giants' different choices in this field are exactly the biggest variables of this era.
This article is written based on public information and is for information exchange only, not constituting any investment advice.
This article is from the WeChat official account "Medical Shine", author: Yan Song. Republished by 36Kr with permission.