Robots are also vying for chips! Tesla lowers AI5/AI6 storage specifications to make way for Optimus mass production.

Tesla is adjusting the storage specifications of its next-generation AI chip, which is used in both Autopilot and Optimus, to meet the mass production needs of humanoid robots.

According to recent reports, Tesla CEO Elon Musk announced a significant reduction in the storage specifications of the AI5 and AI6 chips to ensure sufficient storage for the mass production of Optimus. Musk also stated that this adjustment will not affect the performance of the humanoid robot.

Specifically, Tesla halved the memory capacity of the AI5 chip to 72GB LP5 and reduced the memory capacity of the AI6 chip by one-third to 144GB LP6. Musk stated that this adjustment was the "only way" to obtain sufficient production capacity for Optimus manufacturing, while also significantly reducing costs .

Musk also stated that the impact of this adjustment on Optimus's performance should be "negligible" because memory bandwidth is the more significant limiting factor compared to total memory capacity.

More noteworthy is that this comes just over two months after Tesla thanked Micron for providing an "extremely substantial" storage quota during its July earnings call. Even with the quota, chip specifications still need to be adjusted, highlighting the tightness of storage supply and indicating that humanoid robots are becoming a new source of high-end storage demand.

The performance impact of AI5/AI6 storage downgrade is still unclear.

Tesla has not yet disclosed the specific storage specifications of the AI5 and AI6 chips after the adjustments. However, given the long-term tightness in storage supply, Tesla's decision to lower these specifications indicates that it is reallocating its limited storage resources: the storage needs for Optimus mass production are being prioritized.

During the July earnings call, Musk stated that AI5 was progressing well and would initially be used in Optimus, and that AI6 would become "the world's best edge computing chip." Now, the significant reduction in storage specifications for these two chips contrasts sharply with their high-end positioning. Adjustments to storage specifications typically involve metrics such as bandwidth, capacity, or latency, which can affect the chip's data processing efficiency in real-time perception and decision-making scenarios. Tesla has not quantified the actual impact of this adjustment on FSD performance.

Musk claims that the storage adjustments will not affect Optimus's performance, but this statement is currently unsupported by any technical parameters or test data.

Optimus relies on the collaboration of visual perception, motion control, and decision-making models, and storage specifications affect model inference speed and the efficiency of multi-sensor data fusion. If the adjustment does not affect performance, a possible explanation is that Optimus's current software stack has not yet reached the adjusted upper limit of storage bandwidth requirements, or Tesla has offset the hardware specification change through algorithm optimization. However, Tesla has not specified which scenario applies, and it is currently impossible for outsiders to independently verify.

Storage shortages spread from data centers to cars and robots

The significance of Tesla's adjustment lies in providing new evidence that storage shortages are spreading to the automotive and robotics sectors.

In July of this year, Musk admitted during an earnings call that the expansion of the AI system had led to a shortage and high prices of memory chips. He thanked Micron for providing Tesla with a "very substantial" storage quota on "very reasonable" terms, despite "extremely crazy" memory prices. He also mentioned that Karn Budhiraj, VP of Supply Chain, helped secure this quota. Even now, despite receiving the quota, Tesla still needs to downgrade its AI5/AI6 storage specifications, indicating that storage resources remain tight and reflecting competition for storage needs among Tesla's different business units.

Optimus' weekly production has increased from dozens of units per week in the second quarter to hundreds, and management aims to reach over a thousand units per week by the end of the year, with a final plan of approximately 20,000 units per week. As production ramps up, the storage demand from robots will rise accordingly, and humanoid robots are becoming another new source of demand for high-end storage capacity, following AI servers.

Optimus prioritizes over FSD chip

Tesla's trade-off between AI chips and robot mass production points to two possibilities: either Optimus's mass production progress exceeds market expectations, requiring the early locking of key components; or storage supply has become a new bottleneck for Optimus's ramp-up, forcing Tesla to "borrow" resources from other product lines.

Musk has repeatedly emphasized that Optimus is Tesla's most difficult product to mass-produce ever, with production ramp-up following an S-curve. Tesla's recent adjustment to the storage specifications of AI5 and AI6 suggests that, given current storage supply constraints, Optimus's mass production needs may receive higher resource priority.

If storage shortages persist, Tesla may need to make more similar trade-offs between Optimus and other businesses in the future.

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.