Insights

Amid slowing internet healthcare traffic and the limits of single-transaction models, Fangzhou Pharmaceutical is repositioning itself as a “lifelong health service partner.” In H1 2026, revenue reached RMB 1.8245 billion, up 22.2% YoY, while monthly active users exceeded 14.7 million, validating the new strategy. As online consultations struggle to generate long-term value and pharmaceutical e-commerce faces margin pressure from price competition, chronic disease management offers a promising growth path due to high user stickiness, strong demand, and willingness to pay. Focusing on “AI + chronic disease services,” Fangzhou has built an integrated system covering consultation, medication guidance, health interventions, follow-ups, and data monitoring, creating service premiums and increasing user lifetime value. The model provides a potential path for internet healthcare platforms to overcome growth bottlenecks and unlock long-term user value.

BCIs have long faced neural decoding latency, limiting patients’ sense of “embodiment” and prosthetic integration. Recently, the Shanghai Institute of Microsystem and Information Technology of the Chinese Academy of Sciences and QBOSON achieved microsecond-level neural decoding on a 1,000-qubit quantum computer, offering a potential solution to computational bottlenecks. Following the breakthrough, China’s first BCI-specific fund, the “Heying Brain-Computer Interface Fund,” partnered with QBOSON. The two sides plan to leverage quantum computing to process massive noisy neural signals and accelerate deep-brain intention decoding via cloud computing. Within three to five years, they aim to achieve hardware-accelerated decoding for language and motor applications, potentially overcoming BCI latency and advancing the technology toward becoming an extension of the human body.

BD deals between US and European pharma are growing more complex and volatile, signaling new shifts across the innovative drug industry. "BD Deep Dive" focuses on how IIT clinical research is reshaping FDA approvals, capital markets, and cross-border collaboration models. Industry insiders said a US rare-disease company is preparing to partner with a Chinese drugmaker on an siRNA IIT study, using Chinese data to accelerate FDA approval and cut early-stage clinical costs. Beacon Capital's founder noted that many overseas companies frequently sought IIT resource connections in China over the past six months, covering small nucleic acids and gene therapy. Although the new Regulation No. 818 brings uncertainty, IIT cases unlocking M&A and capital-market value still attract global biotechs. Overseas companies hope to use China's vast patient pools and efficient clinical system as a "fast track" to the FDA, compressing time and capital costs.

Inspired by Avatar, Tianjin University graduate Zhang Xingzhi became convinced that brain-computer interfaces (BCI) would reach an industrialization tipping point. Over a decade, he progressed from human-computer interaction research to an executive role at a listed company. In 2025, amid policy and technological breakthroughs, he resigned to found WABO Technology. The company completed three financing rounds within three months. Rather than focusing on medical applications, WABO Technology centers on the “Human Neural Intent Model,” integrating proprietary hardware with software to collect neural signals and build a large EEG model. Zhang believes BCI will have a “GPT moment,” driven by massive real-world neural data. By end-2025, he had built a core team in Chengdu and adopted a lightweight, AI-native model.

In collaboration with Central South University, Tencent's AI for Life Sciences laboratory has built a new AI framework, UniPert-G2CP, with its paper accepted by Cell. The model integrates multimodal molecular perturbation factor representations, enabling transfer learning from genetic to chemical perturbation phenotypes, connecting genetic and chemical screening for more accurate perturbation prediction, mechanism discovery, and efficient computer-assisted drug screening. Its breakthrough: answering why the same compound responds differently across genetic and cellular contexts. Previous AI achievements in drug development shared a blind spot — ignoring downstream effects after compounds enter real cells, making results irreproducible and costly for pharma. UniPert-G2CP, a two-stage deep learning framework, predicts phenotypic responses to compounds in specific cellular contexts from gene-level perturbation information, offering a solution to this dilemma.

Over the past six months, more than $8 billion has flowed into the oral cyclic peptide sector through financing and BD deals, as Big Pharma and leading investors compete for platforms and pipelines. Recent FDA approvals of MSD’s Enlicitide and J&J’s Icotrokinra have validated the modality, but most assets remain preclinical or in early clinical development, with low oral bioavailability still a key technical hurdle. AI-driven platforms from Unnatural Products, Syneron Bio, and Pinnacle Medicines are attracting substantial capital, signaling a shift toward platform capabilities that can iteratively address these challenges. While validated extracellular targets offer near-term commercial certainty, the greater long-term potential lies in drugging intracellular targets inaccessible to small molecules or antibodies. Given the high scientific and development risks, platform resilience and a staged target strategy will be critical to success.