Founder Wang Xingxing made his first public appearance after the IPO at the 2026 World Robot Conference in Beijing . In his speech, he addressed the industry's key challenges, stating that the biggest bottleneck preventing humanoid robots from entering daily life is the insufficient generalization ability of embodied AI . He emphasized that the company is heavily investing in AI models and revealed that Unitree had begun exploring world models as early as 2020 .
#GateStockInsightsChallenge

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

5 Likes44 views

Reward
5
11
1
Share

Comment

Add a comment

ContractEscapePod

· 1h ago

It’s 2026, and humanoid robots are still talking about generalization capabilities, which shows this industry really is a tough nut to crack. But Wang Xingxing dares to say it outright, putting him a cut above companies that only know how to release CG.

Reply0

TimeChainTraveler

· 2h ago

So after all, the reason Unitree’s stock price has been hyped so high is that it still has to compete on AI software. Its hardware moat isn’t deep enough, so now it all comes down to whether its world model can come up with something others don’t have.

Reply0

AuditNightOwl

· 2h ago

I believe in making major investments in AI models—after all, Unitree’s previous demos clearly showed that its software is evolving rapidly. However, how much of the world model can actually be put into practice remains to be seen in its subsequent products.

Reply0

TradFiCrossover

· 2h ago

He said the company is investing heavily in AI, but R&D for something like this is a bottomless money pit—can its financials withstand it? Hopefully it’s not purely to spin a story and pump the price.

Reply0

SignBlind

· 2h ago

From quadrupeds to humanoids, and then from a hardware arms race to world models, Unitree’s path has been pretty clear. Now we’ll see whether the explorations from 2020 can truly turn into robots that are useful in everyday life, rather than ending up as just another industry PowerPoint.

Reply0

SybilHunter

· 2h ago

As someone who has worked on RL, I really relate to him saying that embodied intelligence is the biggest bottleneck. So many teams get completely stuck on the sim-to-real gap.

Reply0

ShitcoinRadar

· 2h ago

Insufficient generalization ability—translated, that means: the robot can do everything in the lab, but turns into an idiot the moment it enters your home. This pain point is all too real.

Reply0

Trending Topics
View More
#
BTCBreaks71000Up10.5%
609.82M Popularity
#
ETHSurges20%BreaksThrough2300
33.36M Popularity
#
BTCETHReboundTradeIdeas
78.21K Popularity
#
USTreasuryBuybacksAndRegulatorySignalsDriveCryptoSurge
2.79M Popularity
#
StrategySurgesNearly12%
7.69M Popularity

Pinned

Sitemap

Founder Wang Xingxing Delivers First Post-IPO Speech at World Robot Conference

Trending Topics

BTCBreaks71000Up10.5%

ETHSurges20%BreaksThrough2300

BTCETHReboundTradeIdeas

USTreasuryBuybacksAndRegulatorySignalsDriveCryptoSurge

StrategySurgesNearly12%

Pinned