Company Reports

We are providing an update on no-moat Yidu after it rolled out a new large language model, or LLM, on its platform. This should bolster its artificial intelligence and Big Data services that could enable further clinical trials and improve hospital management. We are unsure if this will translate into an incremental number of clients immediately, but we believe that the model should enhance Yidu’s product portfolio as it competes to be China’s top healthcare Big Data platform. The company shared that its model, Yidu HH, can perform different functions that should help pharmaceutical clients speed up their clinical trials and its other hospital service businesses. For now, we maintain our fair value estimate at HKD 5.20 per share for Yidu until there are signs of significant adoption of the new model. The shares are currently trading in 4-star territory.

Yidu Tech provides AI solutions to hospitals, policymakers, and life sciences companies using its algorithm, Yiducore, which works by aggregating and converting raw data sets that lie within the IT systems of hospitals in China into structured and standardized data. Its algorithm is embedded in 2.6 billion full-cycle medical records covering 600 million patients from hospitals and provides models from real-world evidence. As more hospitals and disease registries partner with Yidu, Yiducore can process AI solutions more efficiently and accurately as it adds incremental data to its algorithm. As a result, hospitals can use real world-based evidence from the records to determine the best course of action for treatment and drugs for their patients. We think that the rich amount of data from medical records is a barrier entry, and Yidu has built a client base comprising 88 of the top 150 hospitals. We also believe that it is gaining traction commercially with life sciences companies but this has not yet translated to 40%-50% year-on-year revenue growth as previously forecast. Life sciences companies can use the medical database, insights, and disease models to increase chances of clinical trial success while shortening the cost and time. There are switching costs as life sciences companies are unlikely to switch solution provider mid-clinical trial, and should a trial be successful through all four phases, the recurring life span of a contract could potentially last nine years.

We maintain our HKD 5.20 fair value estimate for no-moat Yidu Tech after mixed fiscal 2024 first-half results. Operating losses narrowed to CNY 130 million from CNY 292 million a year ago, but revenue declined 25% year on year to CNY 356 million. The revenue decline was driven by a 71% drop in sales from the health management platform, which sells mostly health insurance and disease management services, as Yidu is restructuring products that have low margins. The life science solutions segment, which focuses on clinical development and commercialization of drugs, saw a 16% increase but did not disclose the incremental customers. We believe that this business may still see headwinds in the short term, given the recent crackdown on graft in the healthcare industry, which is stalling the development of new drugs. The Big Data solutions segment, Yidu’s other core analytics business, only increased 3% year on year, which may signal lukewarm demand from the government and public research sectors. Nevertheless, Yidu was still able to add 6 hospitals (to 94 total) and 4 regulatory bodies (to 38) to its client list; we think this reflects modest demand for its artificial intelligence technology and services. While we are encouraged that Yidu is streamlining loss-making businesses and better allocating its capital, we believe management’s previous forecast of 40%-50% annual revenue growth for the next one to three years may be unlikely now. We want to see stable double-digit revenue growth coupled with margin expansion before shifting to a more positive outlook.

We recently spoke with the management teams of several healthcare companies to get an update on the progress of the anticorruption movement in China’s healthcare sector to assess how this could affect companies under our coverage for the long term. We believe that this will likely become a long-term driver for companies that provide data-driven studies to facilitate research and drug promotion, such as Medlive and Yidu, as they provide doctors a legitimate medium for them to prescribe medicines that could otherwise be perceived as taking kickbacks from pharmaceutical companies. We believe that the anticorruption movement is likely a positive for the long-term clinical development of drugs and this should provide transparency for hospital policies. As for healthcare e-commerce companies such as JD Health and Alibaba Health, we believe the initiative is likely to have less of an impact on their revenue growth in the long term, as they rely on greater healthcare retail demand. However, we think the policies could provide greater consumer confidence for online retail—and we estimate penetration of healthcare e-commerce is low at 5%-10% in China.

Yidu Tech provides AI solutions to hospitals, policymakers, and life science companies using its algorithm, Yiducore, which works by aggregating and converting raw data sets that lie within the IT systems of hospitals in China into structured and standardized data. Its algorithm is embedded in 2.6 billion full-cycle medical records covering 600 million patients from hospitals and provides models from real-world evidence. As more hospitals and disease registries partner with Yidu, Yiducore can process AI solutions more efficiently and accurately as it adds incremental data to its algorithm. As a result, hospitals can use real world-based evidence from the records to determine the best course of action for treatment and drugs for their patients. We think that the rich amount of data from medical records is a barrier entry, and Yidu has built a client base comprising 88 of the top 150 hospitals. In addition, we believe that it is gaining traction commercially with life science companies and expect to see 80% revenue growth year over year in fiscal years 2023 and 2024 (ending March) that will be the basis of its long-term growth. Life science companies can use the medical database, insights, and disease models to increase chances of clinical trial success while shortening the cost and time. There are switching costs as life science companies are unlikely to switch solution provider mid-clinical trial, and should a trial be successful through all four phases, the recurring life span of a contract could potentially last nine years.

We keep our fair value estimate at HKD 5.20 for Yidu after its lackluster fiscal second-half 2023 revenue of CNY 330 million, down 55% year over year. While the company increased the number of clients and users for its main Big Data and platform solutions segment, the growth was overshadowed by a 56% decline in average revenue per client to CNY 1.8 million. Yidu attributed the revenue decline to continuing headwinds from the pandemic, as hospitals were just beginning to focus on other services than COVID-related in March and April, and a lag in resuming normalized operations. Management remains optimistic that revenue growth can return to its previous forecast of 40%-50% per year starting in 2025, but its fiscal 2024 forecast is likely to be conservative as hospitals slowly transition into their clinical trials and services again. We forecast for overall 20% revenue growth year over year in fiscal 2024.

Yidu Tech provides AI solutions to hospitals, policymakers, and life science companies using its algorithm, Yiducore, which works by aggregating and converting raw data sets that lie within the IT systems of hospitals in China into structured and standardized data. Its algorithm is embedded in 2.6 billion full-cycle medical records covering 600 million patients from hospitals and provides models from real-world evidence. As more hospitals and disease registries partner with Yidu, Yiducore can process AI solutions more efficiently and accurately as it adds incremental data to its algorithm. As a result, hospitals can use real world-based evidence from the records to determine the best course of action for treatment and drugs for their patients. We think that the rich amount of data from medical records is a barrier entry, and Yidu has built a client base comprising 88 of the top 150 hospitals. In addition, we believe that it is gaining traction commercially with life science companies and expect to see 80% revenue growth year over year in fiscal years 2023 and 2024 (ending March) that will be the basis of its long-term growth. Life science companies can use the medical database, insights, and disease models to increase chances of clinical trial success while shortening the cost and time. There are switching costs as life science companies are unlikely to switch solution provider mid-clinical trial, and should a trial be successful through all four phases, the recurring life span of a contract could potentially last nine years.

Yidu Tech provides AI solutions to hospitals, policymakers, and life science companies using its algorithm, Yiducore, which works by aggregating and converting raw data sets that lie within the IT systems of hospitals in China into structured and standardized data. Its algorithm is embedded in 2.6 billion full-cycle medical records covering 600 million patients from hospitals and provides models from real-world evidence. As more hospitals and disease registries partner with Yidu, Yiducore can process AI solutions more efficiently and accurately as it adds incremental data to its algorithm. As a result, hospitals can use real world-based evidence from the records to determine the best course of action for treatment and drugs for their patients. We think that the rich amount of data from medical records is a barrier entry, and Yidu has built a client base comprising 82 of the top 150 hospitals. In addition, we believe that it is gaining traction commercially with life science companies and expect to see 80% revenue growth year over year in fiscal years 2023 and 2024 (ending March) that will be the basis of its long-term growth. Life science companies can use the medical database, insights, and disease models to increase chances of clinical trial success while shortening the cost and time. There are switching costs as life science companies are unlikely to switch solution provider mid-clinical trial, and should a trial be successful through all four phases, the recurring life span of a contract could potentially last nine years.

We are lowering our fair value estimate for Yidu Tech to HKD 5.20 from HKD 11.00 after the company significantly missed its revenue forecasts after reporting first-half fiscal 2023 results. Yidu reported revenue of CNY 474 million, a 5.5% decline year on year. The company previously forecast for a 47% growth on a two-year CAGR and that it will reach breakeven in 2024; but given the lackluster results, we believe that it is unlikely that Yidu can come close to its targets. The company reported operating loss margins of 62%, which is an improvement from 92% loss a year ago, but losses widened sequentially from 36% in fiscal second-half 2022. In addition, gross margins also declined year on year to 26% from 29% a year ago. As a result, we are revising our revenue assumptions for Yidu significantly to a 3% decline in 2023 year on year and for a hopeful return to modest growth in 2024. We are also pushing back Yidu’s breakeven target date to 2026 rather than 2024 as originally forecast.

We initiate coverage on no-moat Yidu Tech with a fair value estimate of HKD 11. Our moderate upside represents positive key trends, but also significant risks, given a lack of company history within a nascent industry. The stock declined significantly since its IPO but Yidu has key highlights that justify the moderate upside in our valuation. The company has robust growth prospects but is generating significant net losses, and we see a wide range of investment outcomes for the business. If the company can improve on its margins in the next year, we could see higher forecasts and a better long-term outlook.

Yidu Tech provides AI solutions to hospitals, policymakers, and life science companies using its algorithm, Yiducore, which works by aggregating and converting raw data sets that lie within the IT systems of hospitals in China into structured and standardized data. Its algorithm is embedded in 2.6 billion full-cycle medical records covering 600 million patients from hospitals and provides models from real-world evidence. As more hospitals and disease registries partner with Yidu, Yiducore can process AI solutions more efficiently and accurately as it adds incremental data to its algorithm. As a result, hospitals can use real world-based evidence from the records to determine the best course of action for treatment and drugs for their patients. We think that the rich amount of data from medical records is a barrier entry, and Yidu has built a client base comprising 82 of the top 150 hospitals. In addition, we believe that it is gaining traction commercially with life science companies and expect to see 80% revenue growth year over year in fiscal years 2023 and 2024 (ending March) that will be the basis of its long-term growth. Life science companies can use the medical database, insights, and disease models to increase chances of clinical trial success while shortening the cost and time. There are switching costs as life science companies are unlikely to switch solution provider mid-clinical trial, and should a trial be successful through all four phases, the recurring life span of a contract could potentially last nine years.