Tesla Moves Optimus Robot Delivery Target to Late 2027

Tesla has pushed back the scheduled start of external deliveries for its Optimus humanoid robots to the end of 2027. This update, provided by CEO Elon Musk during the company's Q3 2026 earnings call, highlights a significant shift in the timeline for a product that has long been a centerpiece of the company's future growth strategy. The delay is attributed to ongoing technical hurdles, specifically regarding the dexterity of the robot's hands, its overall durability, and the maturity of its artificial intelligence systems. These are the same core challenges that define the current state of the humanoid robotics industry.

Production is currently underway at the Fremont factory, where the Model S and Model X lines were previously located. Musk indicated that most of the equipment for a high-volume line capable of producing one million units per year is already installed. However, he noted that reaching the target rate of approximately 20,000 units per week will take at least another year. This is because portions of the design are still in flux, preventing the line from running at full efficiency. The company expects to begin shipping to outside customers in industrial settings, focusing on jobs that are dangerous, repetitive, or require around-the-clock operation.

The primary technical bottleneck remains the robot's hands. During the call, a Tesla engineer identified the hand as one of the most difficult hardware components to finalize. The challenge involves creating fingers that are soft enough to manipulate delicate objects yet durable enough to withstand constant wear without the ability to heal like human skin. Custom gears also remain a concern for long-term reliability. Musk agreed that the hand is the main issue holding back broader deployment, alongside broader electromechanical design questions.

On the software front, Tesla is still refining its AI architecture. Unlike its vehicle fleet, which generates massive amounts of data for its Full Self-Driving program, Tesla does not yet have a large fleet of robots collecting real-world data. To address this, the company plans to build between 10,000 and 20,000 robots to practice tasks in a dedicated facility called the Optimus Academy. Additionally, Tesla intends to integrate its AI5 chip into the Optimus units around April 2027 to improve processing capabilities.

The delay follows a history of shifting timelines for the robot. In 2024, Musk suggested internal use in 2025 and high-volume external sales in 2026. By January 2025, he projected that about 10,000 robots would be built that year, with several thousand performing useful work. He later acknowledged that none were actually doing so. The company's Q3 shareholder letter now describes the current output from Fremont as the first production-quality bots, implying that earlier units did not meet that standard.

Competitors in the space are also navigating similar timelines and technical realities. XPeng turned on an automated production line for its IRON robot in Guangzhou in September, with over 80% of core processes automated. XPeng aims for mass production by the end of 2026 and deliveries in the first half of 2027, starting with retail environments. Meanwhile, Boston Dynamics began producing its electric Atlas robot in January and has committed all of its 2026 deployments, including to Google DeepMind. Boston Dynamics is also leveraging a more pragmatic approach to hands, favoring sturdy industrial grippers over human-like dexterity to solve reliability issues faster.

While Musk stated that no other company is even close to Tesla in general-purpose humanoid development, the progress of rivals suggests a more competitive landscape. Tesla faces the same fundamental problems as its peers: creating hands that survive real work, ensuring robots are reliable enough to justify their cost, and developing AI that can generalize to new tasks. The difference lies in the public narrative, with competitors not claiming a lack of competition. For investors, the key question is whether the late-2027 target will hold or if it will continue to shift as the technology matures.