Medical Imaging AI Startup Soombit AI Raises USD 9.1M Series A, Selected for Scale-up TIPS


Soombit AI, a generative AI startup that builds medical imaging reading solutions, raised USD 9.1 million (12 billion won) in a Series A round. The company was also selected for the Ministry of SMEs and Startups‘ Scale-up TIPS program.

SOOMBIT.jpg

Soombit AI was founded in July 2024 by the team that developed a chest X-ray reading-assistance AI business at Kakao Brain. CEO Woong Bae, a former chief healthcare officer (CHO) at Kakao Brain, co-founded the company with other Kakao Brain alumni.

SV Investment led the round. Altos Ventures, Kiwoom Investment and iM Investment Partners also participated. Altos Ventures was the sole investor in Soombit AI’s USD 3.8 million (5 billion won) seed round in 2024 and joined this follow-on round as well.

The ministry’s Scale-up TIPS program matches government R&D funding to companies that private investors have discovered and invested in first. Selected companies receive up to USD 2.3 million (3 billion won) in R&D funding over three years. Soombit AI’s selected project is the development and global commercialization of a generative AI solution for multi-sequence CT (computed tomography) and MRI (magnetic resonance imaging). The company plans to develop generative AI for three-dimensional multi-sequence imaging and to pursue regulatory approvals overseas.

Soombit AI’s main product today is AIRead-CXR, which automatically drafts chest X-ray reading reports. In April, South Korea’s Ministry of Food and Drug Safety granted it Class 3 digital medical device approval. It is the first generative AI-based software as a medical device (SaMD) in South Korea approved to analyze medical images directly and write preliminary findings reports. It is also the first case to which the ministry’s guidelines for approving and reviewing generative AI medical devices were applied. The ministry drew up the guidelines under the Digital Medical Products Act, which was implemented last year.

AIRead-CXR has been trained since 2021 on about 14 million chest X-ray images and reading reports from South Korea and abroad. It analyzes at least 57 types of findings, including pleural effusion, pneumothorax, pulmonary edema, lung nodules, cardiomegaly, active tuberculosis, and rib and clavicle fractures. It produces a preliminary report in a format similar to a specialist’s reading report within one to two seconds. Unlike existing reading-assistance AI that marks lesion locations, the product also supports the report-writing stage. Clinicians can finish a reading by reviewing and editing the AI-generated draft instead of starting from a blank screen. According to the company, it can be adopted without changing existing PACS (picture archiving and communication system) environments or reading workflows.

Six studies on AIRead-CXR have been published in radiology journals: three in Radiology, two in AJR and one in European Radiology. Radiology is described as a top-tier journal in the field.

Soombit AI will use the new funding to extend its product line beyond chest X-rays into more demanding three-dimensional multi-sequence imaging such as CT and MRI. It is developing AIRead-CT, a CT preliminary-findings solution covering the brain, chest, and abdomen and pelvis, with a target launch next year.

AIRead-CT will be shown for the first time at the Radiological Society of North America (RSNA) 2026 exhibition, held at McCormick Place in Chicago from Nov. 29 to Dec. 2. Soombit AI plans to use the showing as a springboard to pursue global approvals and commercialization, including from the US Food and Drug Administration (FDA) and for European CE marking.

“Building on the generative AI technology for drafting reading reports and the business know-how we gained in chest X-ray, we will expand quickly into CT and MRI and become essential infrastructure for radiologists around the world,” CEO Woong Bae said.

Share

Leave a Reply

Your email address will not be published. Required fields are marked *