News · 2026-11-12
Nine South Korean startups vie for a place at Disrupt 2027
Dataist
Nine South Korean startups are competing for a place in the Startup Battlefield 200 at Disrupt 2027
Nine early-stage South Korean startups presented their projects at the DETO 2026 technology conference in Seoul. The winner, Voinosis, will have a chance to enter the Startup Battlefield 200, TechCrunch’s competition for companies from around the world that pitch to investors and conference attendees at Disrupt. The startup is developing AI to detect changes in brain health through voice. The other participants work in healthcare, financial technology, energy, semiconductors and AI. All nine plan to enter the US market.
Nine South Korean startups vie for a place in the Startup Battlefield 200
Winner: assessing health through voice
Voinosis is developing a tool that analyzes pitch, pauses, speech rhythm and response time to detect changes associated with brain health. A check takes less than five minutes.
The company said the technology had been tested in Greek, Japanese and English, with no need for separate training for each language. The startup says that in a clinical study involving 146 people, the system identified possible cognitive problems with 88% accuracy.
Voinosis’s target customers include hospitals, eldercare facilities, insurance companies and individual users. The startup also plans to connect voice checks with cognitive skills training, health assessments and treatment by medical professionals.
The company was founded by Grace Shin, who worked as an otosurgeon for more than 25 years. Voinosis is now raising a Series A round; a South Korean pharmaceutical company has already become a strategic investor.
146study participants
88%screening accuracy
Less than 5 minutesper check
Other participants
Awesome Lab from Seoul makes ThermoFlux, a heating system that generates heat through water electrolysis and ionization. It is designed for heating buildings, water supply and industrial processes. The company says the technology reduces energy consumption and carbon emissions. In the US, Awesome Lab plans to supply core modules to manufacturers of boilers, water heaters and heating, ventilation and air conditioning systems, as well as co-develop products.
Crosshub helps travelers and international visitors verify their identities and set up payment methods abroad. Its IDBlock platform combines identity verification and payment services using self-sovereign identity (SSI) and zero-knowledge proofs. The company says users can verify their identity across countries without disclosing unnecessary personal data. Crosshub also connects this technology to payment systems to make financial services easier for travelers to use.
DeepFusion AI has developed an imaging system based on four-dimensional radar and additional sensors. It is designed for autonomous transport, robotaxis, maritime surveillance and defense applications. The system combines radio-frequency sensors, radar and lidar to detect objects, movement and distance even without a direct line of sight. The company says all the sensors are housed in a single configurable module.
EFLOW is developing specialized motors that deliver greater torque in a smaller size. They can be used in robots, drones, satellites and other mobile devices. The company also makes actuators that control the movement of robotic joints.
Otiton Medical makes two medical devices: one measures body temperature, while the other captures images of the ear canal and analyzes the eardrum. According to the company, its AI was trained on about 55,000 eardrum images, and the device has received medical certification in South Korea.
Riibotics says it can deploy robots in a warehouse in one day. To reduce deployment costs, the startup simplifies infrastructure setup, warehouse mapping, system integration through custom middleware, task programming and safety checks.
Ternell, a semiconductor startup from Ulsan, is developing a new type of memory intended to reduce the power consumption of AI chips. Unlike conventional memory with two states, the T-SRAM system adds a third state to store data more efficiently. The company was founded in 2019 by Kyung Rok Kim, a former Samsung Electronics engineer. According to Ternell, tests of chips made using a 28-nanometer process showed a 33–102% increase in memory capacity per unit of area and a 55–71% reduction in data-access energy compared with conventional SRAM.
Testbank is developing Solve, an app that turns printed learning materials, including workbooks and exam-preparation books, into interactive digital versions. AI identifies spaces for exercises and answers in publishers’ PDF files while preserving the original layout. Users can write answers directly on the page, and the app checks them automatically.
Dataist’s take: Some participants report measurable results and certification, while others mainly describe plans and potential use cases. All nine companies are targeting the US market, but the source does not specify the terms or timing of their entry.
Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.
