Tesla is hemorrhaging the specialized engineers it needs to turn its custom AI chips into working products, with another senior hardware architect departing for a competitor founded by former Tesla insiders. Shishuang Sun, who spent more than five years as a Tesla AI hardware architect designing the Dojo supercomputer and Autopilot computers, recently joined DensityAI as a packaging and systems hardware specialist, starting in July. His departure marks the latest blow to Tesla's in-house silicon strategy, which the company has staked its future valuation on.

Who Is Shishuang Sun and Why Does His Departure Matter?

Sun was not a junior engineer or a mid-level contributor. He held the title of Senior Director of AI Hardware Design at Tesla, a role he was promoted into in April 2025, just months before his departure. His expertise covers the unglamorous but critical work that separates chip designs from shipping products: integrated circuit packaging, vertical power modules, signal and power integrity, printed circuit board design, and system thermal and mechanical engineering. In simpler terms, Sun knew how to get power into a chip and heat out of it at scale, the exact bottleneck that determines whether high-performance AI silicon succeeds or fails.

Sun's career traces the entire arc of Tesla's custom silicon ambition. He first joined Tesla's Autopilot hardware group during the Jim Keller era, when the company was assembling the team to design its first Full Self-Driving computer and reduce dependence on Nvidia. After a two-year stint at Waymo managing hardware engineering teams, he rejoined Tesla in January 2021 as a principal engineer on AI hardware for both Dojo and Autopilot. His decision to leave now, after being promoted to a senior leadership role, signals deeper problems within Tesla's chip organization.

What Is DensityAI and Why Is It Poaching Tesla Talent?

DensityAI is not a random startup competing in a crowded market. It was founded by Ganesh Venkataramanan, Tesla's former Dojo chief, with roughly 20 members of Tesla's supercomputer team following him out the door after Tesla shut Dojo down in August 2025. Bill Chang, another Tesla veteran, serves as the company's Chief Technology Officer. The startup is building chips, hardware, and software for AI data centers aimed at automotive, robotics, and industrial customers, effectively the mission Tesla abandoned when it killed Dojo and shifted to Nvidia for training compute.

Sun's move represents a second wave of departures. DensityAI already pulled in Tesla's former Dojo leadership and roughly 20 of its engineers, and now, a full year later, it has attracted a sitting Senior Director of AI hardware who appeared to be promoted into an executive role after the initial talent drain. This pattern suggests the startup has become an attractive destination for Tesla's chip talent, offering either better opportunities or a clearer mission than Tesla's current strategy provides.

How to Understand Tesla's Chip Strategy and Its Risks

• AI5 Chip Timeline: Tesla taped out its long-delayed AI5 chip in April, nearly two years behind Elon Musk's original timeline, with volume production not expected until mid-2027. This delay already signals execution challenges.
• AI6 and Manufacturing Challenges: AI6 is slated for Samsung's 2nm process, which has hit yield problems, and Musk has floated a "Dojo 3" supercomputer on top of that. Manufacturing bottlenecks compound the talent shortage.
• Terafab Factory Bet: Tesla and SpaceX are pitching the $20 billion-plus "Terafab" chip factory in Austin, a massive capital commitment that depends on having the right people to execute.

Every one of these bets depends on people who can turn a tape-out into a shipping product, the packaging, power delivery, and thermal work that Sun did at Tesla for five years. Custom silicon is fundamentally a people business before it is a fabrication business, and right now the people are moving in one direction: away from Tesla.

Why This Talent Drain Threatens Tesla's Valuation Strategy

Tesla's entire investment thesis has shifted toward becoming an AI and robotics chip player. The company is no longer just an automaker; it is positioning itself as a semiconductor company that happens to make cars. That strategy only works if Tesla can actually design and manufacture competitive chips faster than Nvidia and other established players. Losing a Senior Director of AI Hardware Design, especially one with five years of institutional knowledge about Tesla's Dojo and Autopilot systems, is not a minor setback.

The irony is stark: Tesla killed Dojo, watched the team spin up a competitor, and is now watching that competitor attract the very engineers Tesla needs to succeed with AI5 and AI6. When a company pins its entire valuation on becoming an AI and robotics chip player, the steady loss of the people who actually build that hardware matters profoundly. Sun's departure suggests that even after Tesla supposedly regrouped around AI5 and AI6, the bleeding continues.

For investors and industry observers, the question is whether Tesla can retain enough chip talent to execute its ambitious roadmap, or whether DensityAI and other startups will continue to attract the specialized engineers Tesla needs to compete in the custom silicon space.