Trace Intelligence, established by Pan Hongyang, is dedicated to the edge AI inference market. The company posits that as AI models progress towards architectural convergence, the development of customized AI chips, tailored to customer requirements and striving for ultimate (extreme) PPA (Performance, Power, Area) efficiency, will emerge as the preferred solution. Trace Intelligence has brought together a team of experts in both global AI chip R&D and design automation, and has developed a chip design process that is inherently AI-driven. This approach initiates from customer specifications, facilitating the collaborative optimization of software and hardware architectures. By harnessing AI technology, it reduces development costs and accelerates the iteration cycles for customized chips. At present, the company has successfully moved from the method validation stage to the commercial delivery stage, completing its incubation round and securing a new round of funding exceeding 100 million yuan. Looking ahead, Trace Intelligence will partner with customers to co-define, co-design, and tape out a 6nm dedicated inference chip, with an initial focus on edge-side speech recognition applications. The company's core strength lies in its innovative design methodology, which effectively bridges the divide between customer needs, application scenarios, and the integration of software and hardware. Trace Intelligence aspires to disrupt the negative cycle prevalent in edge inference scenarios, bring AI computing capabilities closer to end-users, and deliver a seamless intelligent experience through edge-cloud synergy.