Tesla put its Optimus robot behind glass in Shanghai

Attendees hoping to see Tesla's humanoid robot walking around in Shanghai got a bit of a museum tour instead.

At the World Artificial Intelligence Conference (WAIC), Tesla showcased its second-generation Optimus behind a glass display case. The robot didn't move, walk, or interact with visitors. Yet, as one of the few high-profile American AI products on the show floor, it still drew massive crowds.

Glass cases and 18 local challengers

While Tesla kept Optimus frozen in place, Chinese robotics companies used the home-court advantage to put their hardware front and center.

A total of 18 humanoid robots were on display across the floor, including the locally developed Healthy Loong. A long roster of domestic contenders took the stage, including Fourier, Tlibot, Dataa Robotics, Robotera, Kepler Robot, Ti5 Robot, and Leju Robot.

Leju Robot demonstrated its Kuavo bot, which runs on Kaihong OS—an operating system derived from Huawei's open-source OpenHarmony. Company engineer Wu Changxuan pointed out a hard truth: building the software system to make these machines useful is proving far more difficult than assembling the physical hardware.

The massive cost gap

Despite the glossy demos, manufacturers admitted that humanoid robots are largely stuck in the lab. While they are being tested across education, healthcare, elder care, and factory floors, companies confirmed they are not yet generating meaningful commercial revenue.

The real issue comes down to pure economics. An engineer from Sichuan-based Tlibot explained that building a single humanoid robot currently costs between 500,000 and 1 million yuan ($70,000 to $140,000), depending on how capable it is.

That makes Elon Musk's aggressive projections look even more ambitious. Musk recently told shareholders that Optimus could eventually carry a price tag under $30,000—and single-handedly propel Tesla to a $25 trillion valuation.

Why it matters

Putting Optimus in a glass box was probably a smart PR move for Tesla. It drummed up buzz across Chinese social media without risking a disastrous on-stage stumble, but it also signals that real-world deployment is still far off.

For the broader industry, Tesla's sub-$30,000 price target sets an aggressive benchmark. Chinese robotics startups are showing off dozens of functional prototypes, but at current production costs topping $70,000, wide-scale adoption remains impossible. We think the real bottleneck isn't mechanical dexterity anymore—it's driving down actuator and sensor costs while developing a general-purpose AI brain capable of handling complex physical work.