A key executive who led Tesla's AI hardware design has moved to a startup founded by the former Dojo team. Tesla is accelerating development of next-generation AI chips, and continued departures of key hardware staff are drawing attention to its in-house semiconductor strategy.

Electrek, an electric vehicle outlet, reported on Aug. 24 that Sishuai Sun (시슈앙 선) has been in charge of packaging and system hardware at DencityAI since July.

Sun spent more than five years at Tesla working on Dojo and Autopilot computer development. Before leaving, he served as senior director of AI hardware design. He played a key role in turning Tesla's in-house AI chip designs into systems that can be mass-produced.

His career is closely linked to Tesla's in-house semiconductor development strategy. Sun joined the Autopilot hardware group when Jim Keller was at Tesla and took part in developing the first Full Self-Driving computer. At the time, Tesla was building an in-house chip design organization to reduce its reliance on Nvidia.

He later moved to Waymo, where he served as a hardware engineering manager. He led teams covering design validation, signal and power integrity, electromagnetic compatibility and printed circuit board-related work.

Sun returned to Tesla in January 2021. He worked as a senior AI hardware engineer for Dojo and Autopilot and was promoted to senior director of AI hardware design in April 2025. His remit was broad. It included integrated circuit packaging and vertical power modules, as well as signal and power integrity, PCB design and system thermal and mechanical engineering. He oversaw the broader hardware system needed to turn chips into products, not just chip design itself.

DencityAI, which Sun joined, is a startup formed after Tesla halted Dojo. After Tesla stopped Dojo development in August 2025, Ganesh Venkataramanan (가네시 벤카타라만안), who led the supercomputer team at the time, founded the company. About 20 people from Tesla's supercomputer team moved with him, and another former Tesla employee, Bill Chang (빌 창), serves as chief technology officer.

DencityAI is developing AI data center chips, hardware and software for automotive, robotics and industrial customers. It also touches on a space left as Tesla shifted AI training computing infrastructure toward Nvidia after shelving Dojo.

A particular focus in this personnel move is Sun's area of expertise. In the AI semiconductor industry, advanced packaging and system integration, as well as transistor design, are emerging as key competitive factors.

High-performance AI chips are not made viable by boosting computing performance alone. They must be integrated into a system while delivering large amounts of power stably and controlling generated heat. That is why Sun's experience at Tesla in packaging, power delivery and thermal design can directly connect to DencityAI's AI hardware development.

Attention is also on Tesla's next-generation chip development timetable. Tesla completed tape-out of the AI5 chip in April, but that was about two years later than the schedule previously presented by CEO Elon Musk. Mass production is not expected to be easy until mid-2027. AI6 is set to use Samsung Electronics' 2-nanometer process, but that process is known to be facing yield issues. Musk has also mentioned a separate concept for a next-generation AI computing system called 'Dojo 3'.

Manufacturing infrastructure is also being developed. Tesla and SpaceX are pursuing a so-called TeraFab chip plant in Austin, Texas, of more than $20 billion.

But the success or failure of the in-house chip business ultimately depends on the people and technical capability to turn designed chips into products that can be shipped. As Tesla reorganises its semiconductor strategy around AI5 and AI6, the move of key hardware staff to a rival startup could be a burden. After the breakup of Dojo led many former leaders and engineers to leave Tesla, a serving senior director-level figure has also joined DencityAI, drawing greater attention to staffing at both companies.

Tesla is positioning AI and robotics as future growth engines and is strengthening its in-house chip competitiveness. DencityAI, for its part, is also challenging the market for AI hardware for vehicles and robots based on former Tesla personnel.

Going forward, attention is expected to focus on whether Tesla's outflow of key semiconductor staff will continue and whether DencityAI can grow into a real competitor in AI chips and systems based on technology and personnel accumulated at Tesla.