Transferring trained AI models into analog circuits is the vision of Styx, a founding team from Hamburg that is being advised and supported by Startup Port in its founding process. Digital AI systems often require a lot of energy, are dependent on the available network and are also susceptible to security risks. Styx wants to meet this challenge with a new type of software and won 2nd place at CODE 2025 for this project. <\/p>
The young start-up is developing software that converts AI models into chip designs, which enables Kl inference directly in the device, completely offline, with less energy consumption and without delays between data retrieval and response. This means that data processing works without a network, with a longer service life and greater functional reliability. <\/p>
In the pitch at this year’s CODE 2025 innovation conference, the jury awarded Styx second place and prize money of 10,000 euros for their idea. CODE 2025 would like to use the prize money to support the development of innovative ideas, demonstrators and prototypes. A great success for the interdisciplinary team around <\/p>