Technological value ≠ commercial value. In AI healthcare, there are no 'myths,' only 'marathons.'

This year, AI healthcare is undoubtedly the hottest 'belief' in the capital markets. The theme has fully erupted, with concept stocks taking turns to shine. According to Frost & Sullivan forecasts, from 2023 to 2033, the scale of China's AI healthcare market will soar from 8.8 billion yuan to an astonishing 315.7 billion yuan, with an annual compound growth rate of 43.1%.

However, behind the spotlight and the trillion-dollar imagination lies the silent struggles of countless Ventures and a series of loss-making earnings forecasts. Many are asking:Beyond laboratory breakthroughs, who exactly is making real money?

I. Commercial Exploration: Which monetization models have succeeded?

The ruthless market has filtered out several proven survival paths, outlining the realistic landscape of AI healthcare commercialization.

Pay-per-use modelThe most straightforward example is represented by Tencent's AI medical imaging solution, Miying. Its AI-assisted diagnostics have been integrated intoover 1,800 hospitals, with more than 120 million diagnoses completed cumulatively. Each CT image analysis costs approximately 0.5 yuan, which is only one-tenth of the cost of traditional manual diagnosis, with a marginal profit rate of up to 60%. However, the bottleneck for this model in the domestic market lies in the incomplete integration with medical insurance payments, and settlements are mostly conducted internally within hospitals.

Hardware bundling modelis the current mainstream approach. AI algorithms are deeply integrated into high-end imaging equipment as a value-added feature to enhance premium pricing. The 'AI-CT' all-in-one machine from United Imaging Healthcarecan generate annual service revenue of up to 1.5 million US dollars per unit, with the Chinese market contributing 40% of the share. This model cleverly avoids separate budget approvals for 'pure software' in hospitals, turning AI into a mandatory scoring criterion in equipment bidding.

SaaS subscription modelreconstructs the profit logic. Tencent Miying's 'AI Diagnostic Service Cloud Platform' adopts a 'pay-per-use plus annual cap' model, where hospitals pay between 50 to 80 yuan for each CT analysis, and no additional fees are charged once annual consumption exceeds 500,000 yuan.This model has been contracted by over 1,200 hospitals, with a paid conversion rate reaching 82%, among which 65% are small and medium-sized medical institutions, significantly lowering the threshold for AI applications.

InIn pharmaceutical enterprises and insurance sectors, the value-based payment modelhas gone further. Insilico Medicine’s AI drug discovery platform can compress the new drug R&D cycle from 4.5 years to 12-18 months, and its collaboration with pharmaceutical companies adopts a combination model of "upfront payment + milestone payments + sales sharing". OpenEvidence, a US company, provides its clinical decision support platform to doctors free of charge but monetizes through offering precision advertising services to pharmaceutical firms. By 2025, it achieved50 million USD in advertising revenue, with a gross margin exceeding 90%。

The data licensing modelpoints towards the ultimate vision of healthcare AI—monetizing data itself. Tempus AI, a US precision medicine firm, delivered an impressive performance in 2025: annual revenue exceeded1.27 billion USD, surging 83% year-on-year, and for the first time achieved a positive adjusted EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization). Its core model is precisely the integration of massive clinical and molecular data to provide pharmaceutical companies and research institutions with data insights and AI tool licensing services.

2. Harsh Reality: Why is scaled profitability still distant?

Although the business model has taken shape, one harsh reality that cannot be ignored is that for the vast majority of AI healthcare companies, especially pure software and algorithm firms, commercialization remains slow, and the contribution cycle to financial performance is long.

Look at the shadows behind these accolades: Ambry Genetics, considered the leader in AI pathology, saw its revenue in the first three quarters of 2025 decline by nearly 30%, with net profit turning into losses; Bsoft, a giant in medical information technology undergoing full-scale transformation towards cloud and AI, faces short-term performance pressure; even high-profile companies like Insilico Medicine and iFlytek Healthcare remain mired in losses.

The core pain point lies in:Technical value ≠ commercial value, product deployment ≠ financial results materializing.From hospitals ‘trialing’ to ‘purchasing,’ from ‘projects’ to ‘stable income,’ there is a long decision-making chain, strict price approvals, complex health insurance integrations, and the reshaping of clinical habits in between.

The commercialization of AI healthcare products is a 'marathon' testing endurance, resource integration, and ecosystem construction, rather than a 'sprint' of technological innovation.

3. When skepticism reaches its peak, what are we doing?

When 'When will AI healthcare make money?' becomes the market’s existential question, Huajian Healthcare’s (01931.HK) strategic transition from IVD to an AI healthcare capital operation platform is undoubtedly under intense scrutiny. Market doubts are sharp and concentrated:When will the synergy value be reflected in the financial reports? Is there a timeline for commercial implementation?

We hear these voices. True strategic planning often goes against the patience of the short-term market. Our confidence lies not in a vision built on castles in the air, but on three solid 'ballasts':

– Channels and ecosystems are scarce tickets. Huajian Medical has an IVD distribution network with over 30 years of deep cultivation in China, covering more than 1,700 tertiary hospitals (K1), which is the 'expressway' for products to reach clinical applications directly. The acquisition of Venture further secured valuable data entry points and application scenarios, serving nearly 7,000 medical institutions and accumulating 300 million health records. Technology must be implemented, and we have the broadest testing ground and distribution network.

Second, it's the validation of the 'blitzkrieg' in ecosystem synergy.During the recent Nipah virus outbreak, Carbon Bio within the ecosystem quickly launched a testing solution, backed by early synergy with Venture’s data warning system. This validates the feasibility of our closed-loop model of 'AI warning - rapid diagnosis - ecosystem synergy'.

Third, innovate the 'reverse NewCo' model to define the rules of the game.Facing the dilemma of high valuations of top global AI technologies, we pioneered a new model: using China’s 'market entry ticket' and ecosystem empowerment for equity participation, in exchange for cutting-edge technologies that have been validated overseas. This elevates us from 'heavy capital buyers' to 'joint issuers of technological value', opening up limitless possibilities.

Market skepticism is the whip urging us to move faster. But we believe that before the commercial inflection point,staying power is more important than speed, and ecosystem barriers are more enduring than single technologies.。

Fourth, in 2026, the inflection point for AI healthcare commercialization has arrived.

Though the night is long, dawn is approaching. Multiple signals indicate that 2026 is likely to become a definitive turning point for AI healthcare commercialization.

Policy Turning Point:The national 'Implementation Opinions on Promoting and Regulating the Application Development of Artificial Intelligence + Healthcare' has clarified the roadmap, transitioning from pilot exploration to a new phase of large-scale promotion.

Payment inflection point:The National Healthcare Security Administration has successively issued guidelines, for the first time explicitly including 'artificial intelligence-assisted diagnosis' in pathology diagnosis service pricing projects, removing the biggest obstacle to AI healthcare billing. The clarification of the payment pathway means that demand will be converted into real, sustainable revenue.

Cognitive Turning Point:The demonstration effect of leading hospitals combined with healthcare security recognition will rapidly educate the entire market and accelerate clinical adoption.

Capital Consensus:As CITIC Securities stated: 'In 2026, the logic of AI healthcare fundamentally changed, with the core being that this year's payer for AI healthcare became clearer and its payment ability stronger. Therefore, the certainty of AI healthcare commercialization is expected to strengthen this year, opening up the space for AI healthcare commercialization.'

Amidst the current climate of capital frenzy and profitability challenges, the commercialization path for AI healthcare is transitioning from exploration to clarity. From pay-per-use models, integrated hardware and software solutions, to empowering pharmaceutical companies and data licensing, diverse monetization strategies have proven viable.

Despite persistent obstacles to achieving scaled profitability, three pivotal turning-point signals have emerged: clear policy direction, breakthroughs in payment models, and maturing market awareness, indicating that the industry is moving from value validation into a new phase of value expansion.

Ultimately, the trophy for this marathon will go to those who can deeply integrate cutting-edge technology, extensive clinical channels, and a complete industrial ecosystem over the long term.

9379 Views