Horizon Yu Kai: Let the outcome be decided before the battle.

Written by | Chen Hao, edited by Yu Jie | Produced by |Clink | Produced by Automobile 2065438+In the first half of 2009, Che Hejia just changed its name to Li, and the money in the account was only enough for three months. It is said that Li Xiang was in poor health because of excessive pressure.

A few months later, due to the stagnation of financing and problems in the capital chain, Xie Dongying, CFO of Weilai, announced his resignation. The next day, an article entitled "The Worst Man in Weilai Li Bin 20 19" was screened online.

20 19 is the "dark moment" for the upstarts who are betting on smart electric vehicles.

Also hovering in the jaws of death are the supporters and enablers of smart electric tracks like Horizon.

Like Weilai and Ideality, at the end of 20 19, Horizon couldn't get the money, and hundreds of people were laid off (nearly one third). Yu Kai recalled, "At that time, it was overwhelming on the Internet that the horizon was dying."

Although Horizon had released the first generation of automotive AI chips at that time, it was also the first automotive AI chip in China-Journey 2.

In fact, when Yu Kai returned to China to join Baidu in 20 12, he was already an influential figure in the global machine learning field.

A year later, the Baidu Deep Learning Research Institute (IDL) led by him not only became the only deep learning research institution in the world (the other is google), but the establishment of IDL marked that the entire Internet industry in China began to pay attention to cutting-edge technology research and development.

Baidu's famous autopilot project was also born in IDL.

Later, with the gradual warming of autonomous driving, many people in Baidu successively entered car companies or created autonomous driving companies. Baidu is dubbed the "Whampoa Military Academy" in the domestic autonomous driving industry.

However, Yu Kai, who built Baidu's autopilot project, chose to be the underlying smart chip hardware.

When Yu Kai founded IDL on 20 12, people didn't know what deep learning was. On 20 15, Yu Kai became the first person to leave Baidu's autopilot team. When he switched to a chip, others still "don't understand."

"In the past 20 years, no one in China has made money from chips." Because it requires huge investment in the early stage, there is no return.

From 20 15 to 20 19, it is a lonely five years with no return on the horizon. Yu Kai said that time was from darkness to darkness and then to darkness. ...

Later, with the mass production of Journey 2 and Journey 3, the customer scale of Horizon increased from 0 to dozens, and the company scale increased from dozens to more than 1200 now.

The reaction of the capital market is more direct and rude:

Horizon has rapidly developed into a "unicorn" in the industry because it is involved in chips and autonomous driving at the same time. It is reported that Horizon's post-investment valuation is as high as $5 billion.

According to industry investors, many investment institutions want to participate in Horizon's C round of financing, but they can't get in at all. Limited by the barriers of R&D and the mass production of automotive chips, such investment targets as China are very scarce at present.

This "barrier" and scarcity were built by Yu Kai and Horizon six years ago.

On the afternoon of August 10, located on the third floor of Horizon Building in Yongfeng Industrial Park, Zhongguancun, Fenghao East Road, Yu Kai, founder and CEO of Horizon, told the automobile production network, "Let the victory or defeat be decided before the battle."

We are number one in the world.

There is a culture inside Horizon Company, which is called pirate culture.

The core of pirate culture is to compete where there is no competition. "We must decide the outcome before the game."

The start and development of Horizon was carried out under the guidance of this "anti-knowledge" strategic thought.

"Don't chase hot. Seeing hot things now, there must be some companies that made some key decisions and long-term organizational preparations four or five years ago. "

If Yu Kai returned to China to create IDL in 20 12, he was the only one in China, and when Horizon was founded as an artificial intelligence chip in 20 15, he became the only one in the world.

"When we make chips, no one wants to make chips. We are the first company specializing in deep neural network chips in China, and now it is probably the first in the world. " Yu Kai said that after searching, there is really nothing earlier than the horizon.

And why choose smart chip, because of a prophecy:

"In the future, the automobile industry will certainly copy the laws of the mobile phone industry and the PC industry: one type of enterprise (like Apple) chooses vertical integration and does everything by itself; Most enterprises will pay more attention to products, brands, channels and users, and it is impossible to have all the underlying technical capabilities. "

Horizon is one of the core technology suppliers at the bottom.

Yu Kai's anti-knowledge thinking and accurate forward-looking judgment on the industry have won the time advantage for Horizon.

When autonomous driving becomes more and more popular and chip shortage becomes a global problem, more and more people begin to pay attention to and invest in the field of automotive smart chips.

However, at this time, the horizon has gone through a "lonely period" with no return for several years and entered the harvest stage. Now Horizon is the only AI chip enterprise in China that achieves mass production and forms a closed-loop business.

In addition to "early", the horizon is still "fast".

Horizon released 20 19 to 202 1 three automotive AI chips.

Among them, Journey 2 took 10 months from release to mass production, and Journey 3 took 7 months. Journey 5 is expected to take 12 months from release to mass production (mass production is planned for the third quarter of next year).

The general situation in the industry is that it takes an average of two years for a chip to be released to mass production. Why so soon?

Yu Kai said, "Let's fight."

On the platform of the new generation of Li ONE, Li would like to mention: Horizon students often work overtime with our Li team until evening 12.

Under the stage, a person in charge of autonomous driving whispered beside Li: No, it's always all night.

"I even offer recipes.

But I can't cook. "

On July 29th, Horizon Journey 5 conference invited 8 car companies and dozens of tier 1 suppliers to stand for themselves. In addition, Yu Kai spent nearly one third of his time introducing these customers on the stage to pay tribute to the horizon.

This is almost unprecedented in the history of previous press conferences.

"So many car companies are willing to give us a platform that other suppliers can't. What is the reason? "

Yu Kai said, "Any enterprise should think clearly about its own strategy. Never say that what you think is different from what you do. I think every customer can see it. "

On June 30th, Chen Hong, Chairman of SAIC, released the "Soul Theory" of automobile enterprises at the shareholders' meeting, which put the game relationship between traditional automobile enterprises and core technology suppliers in the process of electrification and intelligent transformation of automobile industry.

There is no doubt that Horizon will continue to face the same problem in the era when smart chips in automobiles have become or will become the "brains" of automobiles.

In the interview, Yu Kai made a metaphor for the positioning of the horizon:

We are the underlying technology provider of autonomous driving. If the self-driving company (tier 1) is cooking, we (tier2) will provide all the cooking materials, seasonings, chopping boards and kitchen knives ... I even provided a "recipe", but I just don't fry this dish.

"This is our border." A more professional expression is, "We should have the ability of full stack, but we can't make full stack money."

Yu Kai believes that the game of discourse right between OEM and supplier has always existed. For example, there have been many automobile manufacturers complaining about the bundled packaging scheme provided by a well-known supplier in the industry.

But obviously, Yu Kai didn't want Horizon to face such a situation, but chose to try his best to build an open platform for customers to take their own needs. Horizon "only took a ladle of drinks".

The choice between Yu Kai and Horizon is easy to ask: If you can do it, why not do it? In other words, if you can make money, why not?

"Do you think Bosch has the ability to build a car?" Yu Kai asked.

"Too many people like being rabbits,

I still think it's better to be a turtle. "

When Baidu Deep Learning Research Institute was first established, it wanted to recruit some talents in artificial intelligence, but found that it could not be recruited.

"I don't think so. Before I went abroad, all major universities in China set up artificial intelligence majors. Where have all these people gone? "

Later, when Yu Kai inquired, he found that some went to banks and telecommunications, and some became civil servants ...

This reality may be a little incomprehensible to Yu Kai. He was obsessed with artificial intelligence and neural networks during his undergraduate years, and never changed his direction in the next 20 years, whether he was studying PhD at Munich University, working in the United States, returning to China to enter Baidu, and then quitting his job to start a business.

"My undergraduate thesis, master's thesis, doctoral thesis, my 1 job, my second job, my third job and my fourth job have not changed their direction until today."

Yu Kai also used the metaphor of a race between a tortoise and a rabbit.

"Too many people like to be rabbits, so this popularity jumps here and that popularity jumps there ... I think it's better to be a turtle, that is, to find a long-term direction, not to think too much and not to be too utilitarian."

Yu Kai's persistence in the long-term concept also played a decisive role in the development direction of vision and the shaping of values.

"Endure loneliness and make customers." It is the eight-character cultural values of the horizon.

In the interview, Yu Kai mentioned these words many times. He believes that in the long run, the "most virtual" things of an enterprise-mission, vision and values-are the "most real" core competitiveness.

"If we have a little humility, our team and our classmates will also show arrogance in their daily behavior, and our customers can see it, so we can't go far."

Yu Kai's long-term theory reminds people of what Zhu Yanfeng, then general manager of FAW Group, said in 2002: Building a car should "withstand 20 years of loneliness"

Looking back 20 years later, Zhu Yanfeng's words should have won applause.

The next challenge

However, some experts pointed out that although Horizon has become a "unicorn" enterprise in the industry, it has not really experienced the test.

Compared with the tens of millions of shipments of NVIDIA and Mobileye, Horizon has just started (at present, the cumulative shipments of Horizon have reached 400,000+,and it is expected to reach 1 10,000 this year).

This, Yu Kai also admitted.

"If you deliver 65,438+10,000 vehicles, the probability of making mistakes is 1. If you don't continue to improve the quality and deliver 100,000 vehicles, you will have the possibility of making mistakes100 times."

Now, two of Horizon's three-generation products (Journey 2 and Journey 3) have been proved to be successful, and Horizon has a lot of customers, which proves that the business model predicted and selected by Yu Kai at the beginning exists and is also tenable.

After the "Four Crossing Chishui" stage, Yu Kai said that the next scale expansion should be accompanied by the simultaneous improvement of operating efficiency.

"If there is no efficiency growth, it is actually a loss."

In order to increase the scale and improve the efficiency at the same time, Yu Kai thinks that Horizon's organizational ability, professional ability and management mechanism need to be greatly upgraded.

postscript

Unlike expected before the interview, Yu Kai doesn't look like a science and engineering man in the traditional sense.

He has the specialty and concentration of science and engineering people, but he will metaphysically think and answer more and higher-level questions (philosophy and values).

These hybrids have become a rare trait-focusing on the essence of things (or getting closer to the essence), and then making more accurate, forward-looking and long-term judgments than others.

For example, he thinks that the entrepreneurial environment in China is no longer the state of barbaric growth and entrepreneurship 20 years ago. "Today, the development of any enterprise must truly create value, and the norms and commitments of enterprises must be built according to world-class standards. Otherwise, there will be problems. "

For another example, he said that all living beings are afraid of fruit, while bodhisattvas are afraid of cause, and the real "cause" is more important.

When asked what is the most unbearable thing?

Yu Kai replied without thinking, "short-termism, choose short-term immediate satisfaction." ?

Left: Yu Kai, founder and CEO of Horizon; right: Chen Hao, assistant to President of Easy Car and editor-in-chief of Auto Production Network.