Tesla completes AI5 chip tape-out; TSMC (台积电) and Samsung Electronics (三星电子) to manufacture

What Tesla announced

Tesla (特斯拉) said it has completed the tape-out of its next‑generation AI chip, AI5. CEO Elon Musk posted congratulations to the design team, saying, “Congratulations to Tesla’s AI chip design team on completing the tape‑out of the AI5 chip.” It has been reported that the chip will be manufactured by Taiwan Semiconductor Manufacturing Company (TSMC, 台积电) and Samsung Electronics (三星电子), a rare public acknowledgment of a dual‑foundry production plan.

Technical context and timing

Tape‑out marks the transition from design to manufacturing — the design files are finalized and sent to a foundry for wafer fabrication. For Tesla that matters because the company has been building in‑house silicon for vehicle autonomy and data‑center training for years. What the AI5 is optimized for — in‑car inferencing, Dojo training, or both — has not been fully disclosed and it has been reported that technical details such as process node and die size remain unconfirmed.

Supply‑chain and geopolitical angle

Why use two foundries? Splitting manufacturing between TSMC and Samsung can hedge capacity constraints and speed ramping, but it also highlights the geopolitics of advanced chipmaking. TSMC and Samsung together control the most advanced production nodes worldwide, concentrated in Taiwan and South Korea — locations that sit at the intersection of U.S. export controls, cross‑strait tensions, and global demand for AI silicon. Reportedly, specifics of the production split and timelines are still being worked out.

What this means for China and the broader market

For Western readers less familiar with China’s tech scene: Chinese EV and AI firms are racing to close the gap on domestic chip capabilities, even as international suppliers remain central to high‑end production. Tesla’s move underscores how globalized, yet geopolitically fraught, AI chip supply chains have become. Next steps will be silicon bring‑up, validation and a production ramp — milestones to watch for because they will determine when AI5 actually starts powering Tesla’s vehicles and compute clusters.