Tesla Shifts AI5 Focus to Optimus and Data Centres as AI4 “Plus” Upgrade Nears

Tesla CEO Elon Musk clarified the company’s silicon roadmap during the Q1 2026 earnings call, revealing that the upcoming AI5 chip is no longer a “pressing issue” for Tesla vehicles. Instead, the high-performance hardware will be prioritized for the Optimus humanoid robot and the company’s data centres.

The early completion of the AI5 tape-out was the result of extreme effort from the engineering team. Musk noted that “the team worked incredibly hard” and “had to work every weekend for six months straight, including every holiday.” He added that he was personally “there, of course, myself every weekend” to ensure the project stayed on track without major mistakes.

AI4 Performance Exceeds Expectations

The shift in strategy comes because Tesla believes its current AI4 hardware is already capable of achieving “unsupervised self driving with AI four that is far greater than human safety levels.” Because the current hardware is proving so capable, Musk explained that AI5 is “not certainly not immediately needed in the car.”

While he expects Tesla will eventually switch to AI5 in vehicles when AI4 becomes “so old” that the factories producing it are no longer efficient to keep open, there is no immediate rush to do so.

Introduction of AI4 “Plus” in 2027

To bridge the gap until an eventual AI5 vehicle integration, Tesla is planning an interim hardware update currently dubbed AI 4.1 or “AI4 Plus.” This upgraded version of the current hardware is expected to enter production around the “middle of next year,” though the timeline depends on Samsung’s progress with the necessary modifications.

The new AI4 Plus hardware will feature significant memory and performance boosts:

• Memory Increase: Moving from 16GB to 32GB per system-on-a-chip (SOC), totaling 64GB.

• Compute Power: A “ten percent increase in compute” and improved memory bandwidth.

Musk noted that this refreshed hardware will ensure the fleet remains capable while the company focuses its most powerful AI5 silicon on the more demanding tasks required for robotics and large-scale AI training.