The First Medical LLM Stock Takes the Lead Toward Medical AI's Profit Inflection Point

In a declining medical cycle, medical AI is advancing against the trend.

Recently, Xunfei Healthcare — the first listed medical LLM company in China — released its H1 2026 results, posting revenue of RMB 446 million, up 49.4% year over year, and gross profit of RMB 236 million, up 53.5%. Although it still recorded a net loss attributable to shareholders of RMB 59.3 million, the loss narrowed by 20.0%.

For a long time, the upfront investment in much of medical AI has been decoupled from its later returns: models built with advanced technology and high costs could only gain traction through low prices or by being bundled as giveaways.

Now, Xunfei Healthcare is maintaining rapid growth while its gross profit growth outpaces revenue, which means that LLM-driven software, product, and operations revenue is systematically improving income quality.

It is also proof that genuinely valuable medical AI still possesses ample pricing power.

Before 2026, Xunfei Healthcare had organized its business lines around primary care, regions, hospitals, and patients. The problem was that legacy IT businesses and the new AI businesses were measured under the same segment, so the value created by AI was diluted by traditional businesses.

During H1 this year, Xunfei Healthcare strategically focused on three segments — AI Diagnosis & Treatment Assistant, AI Health Assistant, and AI Digital Foundation — corresponding directly to the three payers: hospitals, patients, and government. Under the new structure, its valuation logic has moved away from that of a medical IT integrator and fully shifted toward a medical AI platform company.

Interim results show that all three segments performed strongly in the first half.

The AI Diagnosis & Treatment Assistant segment (covering B-end and G-end) generated revenue of RMB 202 million, up 47.5% year over year, accounting for 45.2% of total revenue and remaining the largest revenue base.

The AI Health Assistant segment (mainly C-end) posted revenue of RMB 182 million, up 74.9% year over year, representing 40.9% of total revenue. With the fastest growth, it has become the new core growth engine.

The AI Digital Foundation segment (mainly G-end) grew more slowly than the other businesses, contributing RMB 62 million in revenue, up 7.9% year over year. However, its gross margin jumped sharply from 25.9% to 48.6%, making a significant contribution to overall earnings quality.

It is worth noting that the three businesses are not mere expansions of existing operations — each has found new growth points and successfully achieved commercial delivery.

For example, AI Diagnosis & Treatment Assistant was previously deployed mainly at the primary-care level, where regional health commissions were the principal payers and Xunfei Healthcare holds more than 80% market share. This year, however, the segment's main growth contribution has come from hospitals.

The shift from G-end to B-end payment is highly significant for Xunfei Healthcare. The company already commands more than 80% of the primary-care market. Now, with rapid growth at the B end, it has found a second growth curve for the AI Diagnosis & Treatment Assistant business line.

The rise of AI Health Assistant is expected to become Xunfei Healthcare's most critical growth driver going forward, with the potential, ideally, to unlock a RMB 10-billion market at the C end.

The business line centers on the post-visit imaging cloud and post-visit patient management.

Starting with the imaging cloud, the segment's breakthrough lies in the scaled replication of benchmark cases and a stable ARR business model.

In Anhui, Xunfei Healthcare spent five years moving the imaging data of more than 2,300 public hospitals to the cloud, enabling cross-hospital access and AI quality control. Medical insurance pays per service encounter, at approximately RMB 5 per encounter. This model has been established by the National Healthcare Security Administration as a national reference standard, and has been replicated in Guangxi, where the first hospital went live within 20 days and all deliverable hospitals at Grade II and above went live within 90 days.

Xunfei is also building a pan-imaging cloud that includes laboratory data such as blood tests, urine tests, and pathology, which has already been deployed in some hospitals. This is a market at least as large as the imaging cloud. Billed at 10% of examination fees, the Anhui imaging cloud generates annual revenue of approximately RMB 150 million to RMB 200 million, and the pan-imaging cloud could lift the revenue scale of Xunfei's B-end business by another order of magnitude.

Then there is post-visit patient management. After discharge, the AI continuously tracks patients' recovery, reminds them about medication and follow-up visits, and intervenes promptly when problems are detected. The service is purchased out-of-pocket by patients, requiring no investment from hospitals.

Currently, more than 400 hospitals nationwide are target hospitals for AI post-visit management, and the data at hospitals where Xunfei Healthcare has deployed the service are striking: patient satisfaction rose from 80% to above 95%; the three-month unplanned readmission rate fell from 1.9% to 0.9%; the serious complication rate dropped from 3.8% to 0; medication adherence rose from 49% to 95%; and the planned follow-up visit rate rose from 39% to 67%.

At the 2026 interim results conference, Xunfei Healthcare indicated an overall patient willingness to pay of approximately 50%–60% — already a very high ratio. It means that a large number of patients have seen and experienced the incremental value of AI and are ultimately willing to pay for this additional service package.

The final segment, AI Digital Foundation, covers universal health records and a tiered diagnosis and treatment platform. Together, the three form a closed loop of GBC (Government–Business–Consumer) combinations. As resources tilt toward recurring-revenue businesses, the profit structure continues to improve. The AI Digital Foundation has been deployed in Shanghai, Shenzhen, Hefei, and other cities; in Shanghai's Jing'an District, "AI-Driven Resident Health Profiles" was selected into the case library of "Digital and Intelligent Empowerment of Primary Health Care — China in Practice" at the 79th World Health Assembly.

In addition, the AI disease control digital foundation has been deployed in four provinces — Shaanxi, Liaoning, Fujian, and Anhui — while the AI regional collaborative digital foundation is deeply involved in building compact county-level medical communities and strengthening primary care at the grassroots level.

The secondary market has clearly been satisfied with Xunfei Healthcare's new scorecard. After the results were released, the company's share price rose by as much as more than 30% cumulatively over three trading days; CMB International raised its 2026 revenue forecast for the company by 7.6% and lifted its target price to HK$99.58, maintaining a "Buy" rating.

Among brokerages, leading institutions such as Guosheng Securities and Northeast Securities gave positive assessments after the earnings release, believing that the strong showing was not a short-term earnings rebound but a structural improvement in the company's business and business model.

In VCBeat's view, the change in Xunfei Healthcare's share price stems not only from solid earnings but also from the fact that the company has addressed three long-standing challenges facing the entire medical AI industry: high training costs, low application benefits, and a lack of payers.

Before the earnings data landed, medical AI companies had limited credibility when discussing model rankings and technical metrics. Although Xunfei Healthcare became the world's first AI system to pass the comprehensive written examination of the National Medical Licensing Examination back in 2017, and has continued to iterate to a clinically practical level, what doctors truly value has never been benchmark scores but real-world usability.

Today, the AI Diagnosis & Treatment Assistant segment grew revenue 47.5% year over year in the first half, with accelerating adoption at tiered hospitals — the new-generation assistant has entered more than 700 hospitals, and the adoption rate of its outpatient medical record generation stands at 91%. That proves the product's practical value in clinical settings more than any technical metric.

Having secured a leading position, Xunfei Healthcare continues to leverage the data flywheel to further enhance its AI diagnosis and treatment capabilities. With 1.71 million daily feedback responses from doctors on AI assistance and 520,000 daily feedback responses from patient-side services, real-world data continuously feeds back into the model.

In the latest batch of data, the AI's medical hallucination rate fell from 3% to 1.9%, and its diagnostic recommendation accuracy reached 92% — on par with chief physicians at tiered hospitals.

AI Health Assistant follows a similar logic. When patients are willing to pay for the service, it shows that the benefits of AI are genuinely perceived. Willingness to pay, in turn, supports data accumulation and model optimization, creating a positive cycle. As efficiency improves, training costs decline while model capability improves significantly, ultimately forming a complete closed loop of "services for data, data for capability, and capability for revenue".

The ARR model solves the payer problem. The imaging cloud charges per service encounter, with medical insurance volume-based procurement, giving it a clear and sustainable payment source. The model also hosts more AI applications: AI quality control, multi-disease AI-assisted diagnosis, remote consultation, and more can all be embedded into the imaging cloud platform, gaining a channel for value conversion.

In other words, it embeds AI capabilities into the medical insurance payment system, giving AI applications a stable and continuous source of revenue.

For medical AI as a whole, every step Xunfei Healthcare has taken provides a replicable, referenceable path. But with Xunfei Healthcare's first-mover advantage, latecomers will find it difficult to build a data flywheel of comparable scale and replicate the growth trajectory of Xunfei Healthcare's large models.

After a decade of investment, the medical AI industry urgently needs a benchmark case to validate the possibility of commercial profitability. Given the current pace of business growth, there is a strong probability that Xunfei Healthcare will become that trailblazer and achieve profitability within this year.

But turning a profit is only a milestone. The industry as a whole still needs to answer: how to find more growth poles in medical AI, so that more fragmented single-point AI applications and solutions that serve doctors but have not yet closed deals can also achieve commercial delivery.

Looking back, Xunfei Healthcare actually still holds a large amount of data resources that have yet to be put to use.

Against the backdrop of accelerated overseas expansion in the medical AI industry, Xunfei Healthcare signed an AI healthcare strategic cooperation memorandum of understanding with Indonesia's Sinar Mas Group in June 2026, making Indonesia the company's first key overseas market. Product and operating models refined at home are now being replicated in the populous Southeast Asian country, marking a substantive breakthrough in its internationalization efforts.

Recently, the Brazilian government announced plans to invest approximately BRL 2.3 billion (about USD 444 million) to further strengthen AI infrastructure and industry ecosystem development. Of this, about BRL 1.3 billion will be used to build supercomputing infrastructure in Rio de Janeiro, in cooperation with Chinese companies including iFLYTEK and Huawei, focusing on the development of large language models and general-purpose and industry-specific AI applications.

This joint project for AI supercomputing and Portuguese-language large models is, in essence, an extension of the iFLYTEK ecosystem's comprehensive strengths in large-scale computing engineering and multilingual pretraining. It lays a solid technical foundation for extending Xunfei Healthcare's vertical capabilities and opens the door to an entirely new overseas market.

Second, while the world relies on the value of data flywheels to provide real-world evidence services to pharmaceutical and medical device companies, Xunfei Healthcare is also focused on product refinement. If it launches related businesses in the future, this could become another growth focus for the company.

All in all, the deeper the data accumulation, the higher the barriers and the more scenarios that can be monetized in the future. Standing at a new crossroads, how Xunfei Healthcare puts the data in its hands to good use will determine the results it achieves in this intelligent era.