Tesla Advances Automated Manufacturing for Semi

- Tesla is targeting volume production of the Semi to begin by March 2026 at a new facility adjacent to the existing Giga Nevada, with a goal of producing 50,000 trucks annually. Initial pilot production began in October 2022, with the first deliveries made to PepsiCo. - The Semi's powertrain features three independent electric motors on the rear axles, a 1,000-volt system, and a battery capacity estimated around 900 kWh. This delivers a range of up to 500 miles on a single charge while carrying a gross vehicle weight of 82,000 pounds. - To support the Semi, Tesla is building a network of "Megachargers," which utilize V4 power electronics and can deliver up to 1.2 MW of power. This allows the Semi to regain up to 70% of its range in about 30 minutes. The company plans to have 46 Megacharger stations operational by early 2027. - The manufacturing process for the Semi will leverage extensive automation, including AI-guided robots for tasks like welding and parts assembly. Tesla is also expected to use a massive stamping press, sometimes referred to as a "Giga Press," to produce large single-piece body panels, which increases strength and production efficiency. - Tesla’s "unboxed" manufacturing strategy, which involves assembling large submodules of the vehicle independently before final assembly, is expected to be used for the Semi. This modular approach can reduce factory footprint by approximately 40% and cut manufacturing costs by up to 50%. - The Long Range model of the Semi has a curb weight of 23,000 pounds, which, when accounting for a 2,000-pound allowance for zero-emission trucks, makes its payload capacity competitive with diesel counterparts. The truck achieves an energy consumption of about 1.7 kWh per mile. - Beyond vehicle automation, Tesla is positioning itself as an AI and robotics company, developing its own AI inference chips for its Full Self-Driving software and leveraging AI for vision and planning. The company's AI training for neural networks involves 48 networks and requires 70,000 GPU hours for a full build.