The Day Robots Surpassed Human Speed: China's "9.39 Seconds" and the Reality of the Humanoid Competition for Japan

100 meters, 9.39 seconds.

Anyone familiar with athletics will immediately recognize the abnormality of this number.

The world record for the men's 100 meters is 9.58 seconds, set by Jamaica's Usain Bolt at the 2009 World Championships in Berlin. For 17 years, humanity has not been able to break this record.

However, in August 2026, at a stadium in Beijing, China, an "athlete" emerged who ran 100 meters 0.19 seconds faster than that number.

However, this athlete was not human.

It was a humanoid robot with a white body called "Tiangong Ultra."

At the 2nd World Humanoid Robot Competition, which opened on August 22 in Beijing, Tiangong Ultra ran 100 meters in 9.39 seconds, breaking the "9.58-second barrier" that humans had long been unable to reach with its mechanical body.

What's important is not just that it surpassed Bolt's record.

This robot took 21.50 seconds to complete 100 meters in the previous year's competition. It shortened that time to 9.39 seconds in about a year.

This "evolution speed" might be the number Japan should pay the most attention to.

Over 2,000 Robots Gathered in Beijing

The competition is being held at the National Speed Skating Oval, known as the "Ice Ribbon," which was also used for the 2022 Beijing Winter Olympics.

This year's competition registered 666 teams and 2,056 robots from 16 countries across six continents. The number of events expanded from 26 in the previous year to 51, with 1,301 competitions held over five days.

It's not just track and field events like the 100 meters or 400 meters.

Events include soccer, martial arts, wushu, table tennis, weightlifting, tug-of-war, and dance, resembling a comprehensive human sports tournament.

What's more noteworthy is the significant increase in "scenario competitions" that simulate real jobs such as factory work, hotel service, household chores, logistics, and firefighting/rescue.

In other words, what China is hosting is not just a "robot sports festival."

They are testing at a tremendous speed the abilities needed for future industrial and service humanoids, such as high-speed movement, posture control, recovery from falls, object recognition, dexterous manipulation, coordination among multiple robots, and autonomous decision-making, in the form of competitions.

"Surpassing Humanity" Beyond Just 9.39 Seconds

In the 100 meters, besides Tiangong Ultra's 9.39 seconds, "Lightning," related to the smartphone maker Honor, also recorded 9.47 seconds in the race, surpassing Bolt's 9.58 seconds.

It has been reported that Lightning recorded 9.32 seconds in pre-competition tests.

This 9.32 seconds and the 9.39 seconds recorded in the competition are sometimes confused on social media and in some articles, spreading as "Chinese robots broke Bolt's record with 9.32 seconds." At least, the widely reported record of Tiangong Ultra in the competition is 9.39 seconds, and the numbers need to be distinguished.

Furthermore, in the 400 meters, times exceeding the human world record were reported.

From a sports perspective alone, one might think, "It's not surprising that machines powered by motors are faster than humans."

Cars are, of course, faster than humans. There's no necessity for bipedal locomotion.

However, the point this time is that "bipedal machines with structures almost identical to humans" are approaching the stage where they can move quickly in spaces designed for humans.

Stairs, doors, workbenches, shelves, tools, factory equipment.

Most of the current social infrastructure is designed based on the human body.

If humanoids can move, carry things, and work in that environment without significant changes, a huge market could emerge for them.

The 100 meters in 9.39 seconds serves as a "billboard" to clearly show that future to the world.

But Being Able to Run Fast Doesn't Mean "Being Able to Stop"

On the other hand, the reason the competition footage became a global topic wasn't just the spectacular records.

Some robots fell right after the start.

Other machines could only run at a walking pace.

In martial arts, some robots lost balance and became unable to move, leaning against the ropes of the ring, and had to be carried out by human staff.

Among the high-speed running machines, some couldn't decelerate in time after the finish line and crashed into the shock-absorbing equipment prepared for them.

This reveals an interesting contradiction in current humanoid technology.

They are beginning to surpass humans in the single goal of "running fast."

However, actions that humans perform casually, such as "understanding the surroundings and stopping at the appropriate position," "responding to unexpected events," "correcting posture when about to fall," and "grasping and inserting objects accurately," remain difficult.

For robot development, the real high wall is not the top speed.

It's the countless "exception handling" that exists in everyday life.

"Can It Plug a Cable?" Is More Important

That's why seemingly mundane competitions are also prepared for the event.

Connecting cables to the correct places.

Identifying and assembling parts.

Transporting materials to designated locations in factories.

Providing services in hotels and restaurants.

Charging EVs.

Handling clothes and items within a household.

In terms of social media appeal, these tasks lose out to the 100-meter dash.

However, for the industry, these are far more important.

For example, there's almost no need for robots working in factories to be faster than humans.

What is required is to work stably for 8 to 10 hours, adapt even if the position of the target object changes slightly, recover on their own when they fail, and not harm the surrounding humans.

Ultimately, even considering the purchase and maintenance costs of robots, they need to generate sufficient economic value compared to when humans are in charge.

The glamorous "surpassing Bolt" impresses the world with technological prowess.

However, the winner of the humanoid industry will likely be determined not by the 100-meter time but by "how many tasks can be processed per hour, without failure, and at what cost."

On Social Media, "Amazing" and "Scary" Spread Simultaneously

This news quickly spread on Japanese social media as well.

When broadly categorizing the reactions, the first noticeable one is pure amazement.

"When did they surpass the fastest human?"

"The robot's form is beautiful despite being a machine."

"The evolution is too fast compared to last year's footage."

Posts with such sentiments were observed.

Many people particularly focused on the progress from 21.50 seconds the previous year to the 9-second range.

Simultaneously, a reaction unique to Japan was the voice saying, "It reminds me of ASIMO."

For Japanese people, humanoid robots are not a new presence.

Japan has been watching Honda's ASIMO climb stairs, run, and kick soccer balls for over 20 years.

Because of this, there is a complex feeling towards the current situation where Chinese humanoids are monopolizing global attention, with sentiments like "There was a time when humanoid robots meant Japan."

On the other hand, there are also many posts expressing concern about the dangers.

Some people, watching footage of bipedal machines running at high speeds, associate it with military use and express anxiety about "what if they carry weapons."

Regarding the footage of robots failing to stop properly after the finish line, reactions like "They should research stopping methods more than speed" and "It's scary that they can run but can't stop" were also seen.

Similar trends were observed on overseas forums like Reddit, where technical questions like "Why is it so hard to stop after the finish line?" mingled with jokes reminiscent of the movie 'Terminator,' showing a mix of surprise and caution.

In a tally where Yahoo! Real-Time Search used AI to classify specific related posts, there were cases where posts judged to be on the negative side reached 75-88%.

However, this was an automatic classification of a limited number of about 64-65 posts, not a public opinion poll. It should be considered as reference material to see that "not only admiration for technology but also concerns about safety and military use are spreading simultaneously," rather than focusing on the numbers themselves.

Is It Enough to Conclude "China Is Amazing, Japan Is Finished"?

When such news comes out, two extreme debates tend to arise quickly in Japan.

One is the pessimistic view that "China's technology is overwhelming, and Japan has completely lost."

The other is the optimistic view that "Japan has industrial robot technology. Running robots are just a spectacle."

In reality, it's necessary to look at the middle ground.

Japan still has strengths in important fields supporting robotics, such as motors, reducers, sensors, precision machinery, industrial robots, and production technology.

The experience accumulated by Japanese companies in ensuring long-term stable operation, safety, and quality in manufacturing sites, and tailoring devices to fit the site, is significant.

On the other hand, China's strengths lie elsewhere.

A very large number of companies, universities, and research institutions are developing simultaneously.

They produce a large number of actual machines.

They are deployed in competitions and demonstration experiments.

They break.

They improve.

They are deployed again.

They repeat this cycle at an extremely short interval.

Although 666 teams are participating in this competition, 641 of them are from within China. Of the 2,056 robots, 1,975 are from China.

Although it is called a "world competition," in reality, it has a very strong aspect as a competition and demonstration venue for the huge Chinese domestic robot industry.

However, this might be exactly what China is aiming for.

The "Competition" Has Become a Huge R&D Apparatus

If 641 teams tackle the same challenge at the same time, not only success stories but also a large amount of failure data will be generated.

They fall.

They drop things.

They make recognition errors.

Motors overheat.

Joints break.

Communication is interrupted.

Situations that would end as failures for humans can become learning data in AI robotics.

Furthermore, if the competition rules are common, machines from different companies can be compared.

This is a benchmark for robots.

While it has the outward appearance of a sports competition, behind the scenes, a huge demonstration test of hardware, AI, batteries, motors, reducers, sensors, control software, and data collection is taking place simultaneously.

Thinking this way, the news of 9.39 seconds in the 100 meters takes on a different perspective.

China has not created just one "robot faster than humans."

They are trying to create a "system" where a large number of robot companies repeat failures and improvements.

Japan Once Showed the World the "Future of Humanoid Robots"

From Japan's perspective, there is a sense of déjà vu in this scene.

Honda's ASIMO, Sony's AIBO, Toyota's partner robots—Japan was a country that had long presented the future image of a robot society to the world.

In the era when humanoid robots climbing stairs became global news, Japanese companies were certainly at the forefront.

Japan's research and development capabilities have not disappeared even now.

At the Humanoids Summit held in Tokyo in 2026, Japanese companies and researchers also showcased high-durability and high-precision technology.

However, at the same time, relatively inexpensive humanoids from Chinese companies attracted attention at the Japanese venue, and there are cases where Japanese companies themselves use Chinese-made robots for demonstrations.

In relation to Haneda Airport, demonstrations using humanoids from China's Unitree are being advanced as a countermeasure to labor shortages.

This can be said to be a symbolic change.