Tesla is taking the next steps in developing its humanoid Optimus robots—or at least it’s hiring workers to take those literal steps. For up to $48 an hour, you could help collect data to train Tesla’s AI-powered robots, designed to automate work in company factories.

According to job listings on LinkedIn, Indeed, and Tesla’s website, Elon Musk’s electric vehicle company is hiring “data collection operators” to gather movement information and provide equipment feedback on the Optimus robots. Workers are required to wear motion-capture suits and virtual reality headsets to simulate the movements and actions of the bots. Per the job listing, applicants must be able to walk for over seven hours a day and should be between 5’7″ and 5’11″ in order to operate the motion-capture suits. Payment ranges between $25.25 and $48 per hour.

In a video shared across Tesla’s social media in May, dozens of teleoperations workers wear motion-capture and VR equipment and stand alongside the Optimus bots. Standing on a black mat in a large white room, workers mimic lifting and placing items slowly and mechanically, while Optimus bots complete that same motion, but with real items in front of them. In the same space, the robots practice walking and completing other tasks like folding laundry, unaccompanied by human workers.

Tesla announced its Optimus project in 2021, with the goal of using it to complete factory tasks that were “unsafe, repetitive or boring.” The initiative aligns with greater industry-wide investment in automation, spurred by pandemic-era labor shortages. As of 2016, 10% of warehouses reported using meaningful automation technology, according to Westernacher Consulting, with robot shipments expected to increase by up to 50% each year until 2030, per McKinsey data.

But Tesla’s automation efforts have not yet come to fruition. Like Musk’s other ambitions, Optimus’s timeline has overpromised and under-delivered. In 2022, Musk suggested Optimus production could begin as early as 2023. Optimus was initially met with disappointment from roboticists, who, after seeing the 2022 Bumble C prototype of the machine, found it largely underwhelming.

“While there’s absolutely nothing wrong with the humanoid robot that Musk very briefly demonstrated on stage, there’s nothing uniquely right, either,” Evan Ackerman, robotics editor for technology magazine IEEE Spectrum, wrote for the publication. “We were hoping for (if not necessarily expecting) more from Tesla.”

Since then, the Optimus project has made strides.The most recent iteration of the robot saw it perform its first autonomous task of handling batteries at one of Tesla’s facilities, according to its 2024 second-quarter earnings. Musk said in July the robots will be in production for internal use as early as next year, with the goal of being sold to other companies in 2026. Tesla did not respond to Fortune’s request for comment.

Optimus optimism

The Tesla CEO has been bullish on the impact of Optimus on Tesla’s operations, confident the robots could drive the EV company to a $25 trillion valuation. That’s over 36 times its current valuation of $692.94 billion, and over seven times that of Apple’s $3.41 trillion market cap. Musk estimated the eventual production of over 20 billion units, arguing the planet’s 8 billion people will want the product, alongside the growing industrial demand.

But competition for task-automating robots is heating up. In February, Figure, another AI-powered robotics firm, announced $675 million in its latest funding round, giving it a valuation of $2.6 billion. At the beginning of the year, the company—backed by OpenAI’s start-up fund, Microsoft, Nvidia, and Jeff Bezos’s Bezos Expeditions—minted a commercial agreement with BMW to assist the carmaker in vehicle manufacturing. After identifying the best uses for the technology, Figure and BMW will implement the bots at the auto company’s Spartanburg, South Carolina, factory.

If Tesla wants to put up a fight against competing AI robotics companies, it has its work cut out for it. Optimus robots will require constant updates and new prototypes, argued Animesh Garg, robotics professor at Georgia Institute of Technology and senior researcher at Nvidia Research. Each component and upgrade is custom to the bots, and early research in the worker-powered efforts of gathering data on its functioning is a steep investment—and risk.

“It [is] extremely difficult to produce robots at scale,” Garg told Business Insider. “The amount of data collection you’d need would easily be half a billion dollars, and the real question is ‘Even if you do that, do you succeed?’ Because there is no guarantee of success.”