AI / LLMChina
← Sticker Wall
JOURNEY · DETAIL
森亿智能: From Medical Record AI to 800 Million Yuan Loss in Three Years, and a New Attempt at a Hong Kong IPO
Founded: Zhang Shaodian · Shanghai Synyi Medical Technology Co., Ltd.
JOURNEY
Key Fields
FIELD STAMPSIndustryAI / LLM
RegionChina
ScaleMid-size
ChannelOther
Origin
Zhang Shaodian, who previously conducted research on medical Natural Language Processing (NLP) in the United States, returned to China in 2016 to found Shanghai Synyi Medical Technology Co., Ltd. Recognizing that the structuring of hospital medical record data in China was severely lagging—with non-standardized records hindering clinical research—he decided to use AI for post-structuring and data governance, entering the hospital market. Early product validation was achieved by deploying NLP systems in Grade-A tertiary hospitals to extract medical record text.
Milestones
2016
Startup PMF
Zhang Shaodian returned from the US to found Shanghai Synyi Medical Technology Co., Ltd. He targeted the pain point of electronic medical record structuring, deploying NLP systems in Grade-A tertiary hospitals to extract text data and achieve initial commercial fit. The team, consisting of several dozen people, relied primarily on early angel funding to support business exploration between 2016 and 2017.
2018
Series A Financing and Expansion Growth
Secured Series A funding led by Xianghe Capital and others. Synyi began aggressively expanding its hospital client base, covering dozens of Grade-A tertiary hospitals. The business scope expanded from medical record structuring to research platforms and intelligent quality control. While contract values reached hundreds of thousands of yuan, the heavy reliance on customized delivery planted the seeds for future low gross margins.
2021
Large-scale Expansion Turning Point
Capitalizing on the medical AI trend, the company completed Series C and D financing, raising over 500 million yuan in total, and rapidly expanded its team to nearly 500 people. However, the company became heavily dependent on customized projects, which severely dragged down gross margins. With hospital payment cycles lasting six months to a year, accounts receivable piled up, leading to the first cash flow crisis.
2023
Continuous Losses Failure
Prospectuses and public reports revealed cumulative losses exceeding 800 million yuan over three years. Due to tightening hospital budgets, low willingness to pay, and high costs of customized delivery, the company faced a severe financial crisis. It was forced to slash R&D salaries by half, leading to the loss of core talent and stalling commercial implementation.
2026
Transformation and IPO Re-attempt Inflection Point
Riding the wave of medical LLMs, Synyi upgraded its original medical record structuring business into a one-stop clinical research and assistant tool in 2026, aiming to significantly reduce local delivery costs. It submitted a new Hong Kong IPO prospectus in 2026, attempting to secure capital market funding through the narrative of general-purpose large model technology to overcome its difficulties.
Turning Points
- Blindly expanding the team to nearly 500 people; heavy reliance on customized delivery severely eroded gross margins.
- Low willingness to pay among hospitals and long payment cycles led to an accumulation of accounts receivable, crushing the company's cash flow.
- Following the explosion of LLMs, the company quickly pivoted to a one-stop medical research assistant, attempting to lower delivery costs and re-enter the Hong Kong stock market.
Failures & Pitfalls
- Cumulative losses of over 800 million yuan in three years; heavy reliance on customized projects crippled the capital chain.
- Forced to cut R&D salaries by half under financial pressure, leading to the loss of core talent and severely damaging R&D capabilities.
- Tightening hospital budgets; medical AI systems in Grade-A tertiary hospitals were viewed as 'nice-to-have' rather than essential.
关键成功要素
- Targeted the critical bottleneck of post-structuring medical records, accurately hitting the real pain points of Grade-A tertiary hospitals early on.
- Long payment cycles and a heavy-delivery model were severely countered by low hospital budgets and poor collection efficiency.
- Salary cuts accelerated the loss of core talent, leading to a vicious cycle in product delivery and R&D.
- The transition to general-purpose large models offers a path to cost reduction, but the core commercial monetization remains in a difficult transition phase.
Lessons
- The inflection point for paid adoption of AI in healthcare is far from reached; heavy-delivery models consume enterprise survival space.
- Businesses with low gross margins and long payment cycles cannot support the blind expansion of large teams.
- Salary cuts are often the start of a vicious cycle; decisively cutting low-margin, non-viable projects is the fundamental way to save the company.
- Packaging a company with a single new medical technology only provides temporary funding and cannot replace the need for sustainable revenue generation and profitability.
Core Data
- Three-year cumulative loss：Over 800 million RMB (based on public data, independent verification not performed)
- Peak team size：Nearly 500 people (based on public data, independent verification not performed)
- Cumulative historical financing：Over 500 million RMB (based on public data, independent verification not performed)
- R&D salary reduction：Cut by half (based on public data, independent verification not performed)
- Affected core client base：Dozens of Grade-A tertiary hospitals (based on public data, independent verification not performed)
- Year of re-submitting prospectus：2026 (based on public data)
Competitors / Peers
Domestic peers include Airdoc (focused on fundus imaging, already listed), Infervision (deeply involved in AI screening for lung nodules), Shukun Technology (leader in cardiovascular AI imaging), and SenseTime Medical (entering clinical specialties via large models). These companies all face common challenges: long compliance and certification cycles for medical devices, difficulty in hospital adoption, weak willingness to pay, and low gross margins due to heavy delivery. By comparison, while Synyi's focus on medical record data governance and research AI avoids the direct competition of imaging AI, it has fallen into the trap of non-standardized data and low-margin customization. The entire sector is still waiting for a commercial profitability inflection point.