Tesla targets 1,000 Optimus robots weekly amid production and capability limits

By Splitfeed EditorialLast updated Sep 26, 2026, 19:06 UTC

Tesla has increased Optimus humanoid robot production capacity tenfold and is now targeting output of more than 1,000 units per week by the end of 2025, according to reports published on September 25 and 26, 2026. Electrek reported on September 25 that the company has already reached hundreds of units per week, representing a meaningful step up from earlier volumes. At the same time, Elon Musk publicly acknowledged that Optimus faces a significant problem, and separate reporting confirms that the robots struggle to generalize learned tasks to new environments, meaning skills acquired in one setting do not reliably transfer to others.

The twin obstacles here are distinct and both consequential. Scaling a manufacturing line to thousands of units weekly is an industrial challenge involving precision tolerances, supplier reliability and assembly complexity. Generalizing robot behavior across varied real-world situations is an AI and robotics research challenge that has resisted easy solutions across the entire field. Progress on one does not guarantee progress on the other, and the tension between ramping hardware production before software capability is mature sits at the center of the debate around Optimus's near-term viability. Polymarket, cited in the pro-AI camp reactions, puts the probability of Tesla shipping consumer Optimus units by year-end at just 11 percent.

The pro-AI camp points to the production ramp itself as validation, with Polymarket noting that Tesla is targeting over 1,000 units per week by year-end as evidence of real manufacturing momentum. Skeptics push back hard on what that production actually represents. Niedermeyer noted that Tesla is effectively the only customer for Optimus, that the robots operate in tightly controlled and supervised areas, and that they are programmed for specific tasks rather than functioning as general-purpose machines. Captain Carrot questioned earlier product demonstrations, and Zubiqo argued that mass-producing fragile hardware before the software works is an expensive manufacturing trap. Tim Fernholz's observation that dexterous manipulation is simply hard points to a structural difficulty, not a temporary lag. The middle ground, represented by Kautious, Sean O'Kane and NOTRELOAD AI, frames the story around execution history. O'Kane noted that Musk has previously pushed automation before the technology was ready, citing the near-disastrous over-automation of Model 3 production, while NOTRELOAD AI described a messy scale-up marked by failing actuators, factory glitches and supplier delays.

The clearest near-term test will be whether Tesla reports verified third-party deliveries or confirmed commercial customer orders before the end of 2025, and whether any production figures released in quarterly updates reflect units actually shipped rather than manufacturing capacity alone.