Elon Musk has a habit of setting bold targets that stretch the boundaries of what seems possible. This time, the stakes feel different. During Tesla’s second-quarter earnings call, the chief executive described the company’s humanoid robot project as something altogether more demanding than anything that came before it.
“This is going to be the hardest product to scale manufacturing that we’ve ever made at Tesla because everything on the robot is new,” Musk told investors. The words landed with weight. Not hype. A blunt admission from a leader who has steered the company through electric vehicle production hell and autonomous driving delays.
Analysts and engineers who follow the sector closely understand the difference. Cars borrow from decades of supplier networks. Optimus does not. No ready supply of actuators, sensors or dexterous hands exists at the volumes Tesla envisions. The robot demands an entirely fresh chain of components, each one engineered from first principles. But first the company must solve basic constraints around chips, memory and packaging. Shortages already bite.
The Next Web captured the moment in detail. Musk emphasized the human hand as the core obstacle. “The hand is an incredible thing,” he said. No robotics outfit has matched its capability at scale. Tesla’s version needs at least human-level dexterity to handle real factory work. Anything less falls short.
Production realities have shifted since earlier promises. Tesla once spoke of several thousand Optimus units in 2025. Internal updates show only hundreds built so far. A redesign of the robot’s hands paused assembly for months. Overheating motors, short gearbox life and weak payload capacity forced changes. ROIC AI reported the gap between ambition and output in July 2025. The 5,000-unit goal looked out of reach. Insiders called it unachievable.
Yet progress continues in pockets. Tesla converted part of its Fremont factory from Model S and X production to an Optimus line. Musk shared photos of the new setup on X in early July 2026. “Walking the Optimus production line in Fremont,” he posted. The line runs in pilot mode now. A much larger Gen 3 line is slated for 2026. Output will start slow. Painfully slow.
“No, Optimus production will be extremely slow at first, as everything is new. This is not like making a car,” Musk wrote on X days later. The S-curve pattern he often cites applies here with force. Early volumes crawl. Later acceleration can surprise. Still, the initial ramp looks flat and extended. Reliability must improve before any useful factory deployment.
Cost targets add pressure. Tesla aims for $20,000 cost of goods sold per robot at volume. That figure would open doors to homes and factories worldwide. Current prototypes sit far above it. Vertical integration helps. Tesla designs its own actuators, hands and inference computers. The approach mirrors how the company tackled batteries and power electronics. Success there does not guarantee smooth sailing with thousands of unique robot parts.
Supply chain gaps remain the dominant hurdle. Unlike electric vehicles that tap established vendors for everything from glass to tires, Optimus starts from scratch. Musk admitted the team tried desperately to adapt existing motors, actuators and sensors. Nothing worked at the right price or performance. “We had to design everything from physics-first principles to work for humanoid robot and with the most sophisticated hand that has ever been made before by far,” he said in a January 2025 interview covered by Fortune.
Chip supply compounds the issue. Tesla needs far more AI processors for training and onboard inference than current markets provide. Memory, logic and advanced packaging all face constraints. The company builds some components internally. It partners with Samsung and TSMC for others. Scaling those relationships to millions of units will test execution.
Competitors watch closely. Figure AI deployed robots at BMW plants and logged thousands of productive hours loading parts. The data gives that startup an edge in proving real-world value. Tesla’s own Optimus units have not yet performed material useful work inside its factories. Musk acknowledged as much in early 2026. “It’s not in usage in our factories in a material way. It’s more so that the robot can learn,” he told Electrek.
The pattern echoes past Tesla forecasts. Optimus timelines slipped repeatedly. Early 2025 targets for factory deployment went unmet. Leadership changes hit the robotics team. Production pauses followed. And yet the vision persists. Musk believes Optimus could dwarf every prior Tesla product. He predicts it will lift global GDP by an order of magnitude once scaled. Valuation models already bake in some of that optimism. Autonomy and robotics together could eclipse the rest of the auto industry, he argues.
Recent factory moves signal commitment. Shutting down Model S and X lines to free space for robots carries risk. Fremont crews completed the conversion in roughly four months. Musk called the speed “insanely fast.” New modular equipment arrived from Germany. Sub-lines for actuators and batteries now sit ready. Initial output targets remain modest. The company expects the first high-volume Gen 3 line to reach one million units per year eventually.
Training the robots presents another layer. Real-world AI data collection differs sharply from car-based systems. Musk noted that Optimus training needs could run ten times higher than for vehicles to cover the full range of useful tasks. The company invests heavily in Texas-based infrastructure to support it. Videos show prototypes folding shirts, serving drinks and practicing basic movements. Demonstrations impress. Sustained factory productivity is the true test.
Hand redesigns consumed much of 2025. Dexterity shortfalls caused the temporary halt in assembly. The Information first reported the issue, later picked up by TrendForce and TechSpot. Tesla engineers iterated on finger mechanisms, joint durability and grip strength. Gen 3 prototypes are expected by late 2025 or early 2026 with manufacturing-optimized designs. Unveiling could come in the first quarter of 2026, Musk hinted last fall.
Broader economic implications draw attention. If Tesla succeeds, millions of humanoid robots could handle repetitive, dangerous or boring jobs. Manufacturing lines, warehouses, even elder care might change. Skeptics point to the long list of broken promises and the capital intensity required. Tesla burned through cash in recent quarters while investing in AI and robotics. Free cash flow turned negative even as automotive revenue held steady.
So what separates Optimus from previous efforts? The absence of any meaningful precedent. No company has mass-produced dexterous humanoids at low cost. Boston Dynamics focused on research machines that sell for hundreds of thousands. Chinese firms like Unitree offer lower-priced models but lack the AI brain Tesla pursues. The combination of low-cost hardware, end-to-end neural networks and vertical manufacturing forms Tesla’s bet.
Musk tempered language in recent shareholder materials. References to “mass production” disappeared from the Q2 deck. Expectations need calibration, he said. The initial ramp will test patience. Observers who lived through Model 3 production hell recognize the signs. Painful learning curves. Supplier qualification marathons. Yield problems that appear overnight.
Still, Tesla holds advantages. Its AI team trains models on vast video datasets from its vehicle fleet. That data transfers to robot perception. Dojo supercomputers accelerate iteration. Factory know-how from millions of cars built annually provides a base. The question is whether those strengths overcome the novel complexities of bipedal movement, balance and fine manipulation at scale.
Recent X posts from Musk and Tesla reinforce the dual message. Optimus pilot lines run today. Significant volume waits until later this year or 2027. The company tests units in offices and factories for real tasks. Goals remain aggressive. Cost targets sit at $20,000. Production capacity ambitions reach into the millions annually over time.
Investors weigh the risks. Tesla stock reacts to each new robot video and each missed milestone. Optimus represents the biggest swing yet. Success could justify valuations far beyond today’s auto and energy businesses. Failure would burn capital and erode confidence. The manufacturing challenge sits at the center.
Everything on the robot is new. That single fact explains the difficulty. New actuators. New hands. New inference hardware. New software stacks for whole-body control. Each element must reach high reliability before volume matters. Early units serve mostly as learning platforms. Later generations will carry the economic load.
Tesla’s history suggests it can iterate fast once problems are isolated. The company turned the Model 3 ramp from near-disaster to profit engine. Cybertruck production stabilized after early stumbles. Optimus enters a different arena. Physical robots in unpredictable environments demand robustness that software updates alone cannot deliver.
Watch the production line updates closely. Fremont output numbers. Gen 3 design freezes. Supplier announcements. Each data point will reveal whether the hardest manufacturing task Tesla has faced will yield the biggest product it has ever built. The coming quarters will test every part of the plan. Slow at first. Then, perhaps, faster than anyone expects.
WebProNews is an iEntry Publication