Elon Musk稍早於X上透露，Tesla的下一代自動駕駛晶片AI5已經「接近」設計完成，而AI6晶片目前則是處於早期研發階段。他甚至發下豪語，目標是在9個月內完成後續晶片 (包含AI7、AI8、AI9)的設計週期。
AI5 was promised last year, and this year it's "nearly complete."
This post immediately attracted the attention of tech media and analysts because it clearly contradicted Elon Musk's previous statements, especially since Musk had publicly stated back in July of last year that AI5 chips...The design is already "complete".However, six months later, the statement changed to "nearly completed design".
In the semiconductor industry, the process from chip design finalization (tape-out) to receiving samples, followed by a lengthy verification and testing cycle, often takes several months or even longer before mass production can begin. This discrepancy in the stated figures inevitably raises suspicions that Elon Musk may have exaggerated, or that Tesla may have encountered a bottleneck at some verification stage of chip development.
Now that the AI5 chip design is in good shape, Tesla will restart work on Dojo3.
If you're interested in working on what will be the highest volume chips in the world, send a note to AI_Chips@Tesla.com with 3 bullet points on the toughest technical problems you've solved.
- Elon Musk (@elonmusk) January 18, 2026
Cybercab may be forced to continue using AI4.
The delays in AI5's development will most directly impact Tesla's highly anticipated Cybercab. If AI5 chips really won't be in mass production until mid-2027, then the original launch date of 2026 will be affected.Cybercab self-driving taxi fleetIt is likely that they will be forced to continue using the current generation of AI4 chips, instead of the originally expected next-generation powerful computing core.
A new chip every 9 months? Pushing the limits of semiconductor physics.
In addition to AI5, Elon Musk has put forward an even more astonishing goal: to compress the chip design cycle to 9 months, and touted that he has already planned for AI7, AI8, and even AI9.
Such a pace of development is almost unheard of in the semiconductor industry. Take Apple, which has top-notch chip design capabilities, for example. Although its A-series chips maintain a pace of one update per year, the underlying architecture planning and design are often finalized several years in advance and developed in parallel. Therefore, if Tesla really wants to achieve one update every nine months, the investment and pressure on its R&D resources will be astronomical.
