Sun Zhengyi: After 50 years, human driving cars will disappear
“In less than 50 years, people will no longer be allowed to drive on their own in the city or on the highway unless they have a special permit. Because the driver may cause traffic jams and accidents.”
"In the future, people will drive like an old-fashioned sport."
On the 19th of last month, as the head of the Softbank Group, Sun Zhengyi, the CEO of the 100 billion dollar vision fund, made a prediction without hesitation in the annual meeting of Softbank. Standing next to him, Daniel Ammann, president of General Motors, and Liu Qing, CEO of Didi.
(Image from nikkei, copyright belongs to the original author)
When is human driving replaced by autonomous vehicles? Behind the conclusion of Sun Zhengyi is the autopilot industry where Softbank has already begun to lay out.
Today, the secret agent tried to sort out the investment situation of Softbank and its vision fund in the autonomous driving track, and analyzed why Softbank chose these companies, and today's startups have any chance.
Maybe, the next person Sun Zheng is eyeing is you.
What to vote for: automatic driving software and hardware are cast
Speaking of autonomous driving, what do you think of? Is Google's Waymo, Uber, or a traditional car company like GM, Ford? Whether it is a technology giant, a shared travel company, or a traditional car manufacturer, many companies are greedily competing for the auto-driving car market. But behind this battlefield, the most notable is the large-scale flow of funds from companies and venture capital institutions.
According to Pitchbook data, there were 69 investments in autonomous driving in 2017, totaling more than $4 billion. In 2016, 34 investment transactions were only $626 million. What position does Softbank occupy in it?
Before combing Softbank's investment in autonomous driving, the agent may wish to briefly divide the important chains in the autonomous driving industry, as shown in the following figure:
Compared with traditional car manufacturing, autonomous vehicles have specific hardware and software technology requirements. The hardware involves AI chips and sensors, and the sensors include cameras, millimeter wave radars, and laser radars. Related software technologies involve high-precision maps, algorithms, and vehicle-to-everything (V2X).
The agent found that in addition to the relevant investment information in the V2X technology field, Softbank almost touched all the remaining fields.
(Mission is organized according to Crunchbase data)
In the AI chip space, Softbank successfully acquired Nvidia's approximately $4 billion in shares last year and transferred it to its Vision Fund. Nvidia's support for autonomous driving has ranged from the in-vehicle chip, known as the autonomous driving "brain", to working with Bosch to create multiple actions such as an autopilot that is capable of meeting Level 5 levels.
After talking about the "brains," let's take a look at "eyes."
For autonomous vehicles, the sensor is equivalent to an "eye." Because only through sensors, cars can identify roads, traffic signs, pedestrians, vehicles and various basic transportation equipment.
At present, the main sensors include cameras, millimeter wave radar and laser radar. What is the difference? The simple understanding is that, like the car company represented by Tesla, the sensor is dominated by visual technology, and the related functions are mainly realized by the combination of camera + millimeter wave radar, which is relatively more economical and cost-effective. Like Google Waymo and Baidu, Lidar is the dominant factor, and the relative cost is higher, but the accuracy is better.
When Sun Zheng met with the founder of Light, he did not hide the reference to Light's precision camera. The most valuable thing in the future is autopilot. In Sun Justice's view, Light can be seen as a potential replacement for lidar.
After receiving $121 million in investment, Light's founder Grannan admitted in an interview that he hopes that the technology used in the consumer sector can be pushed to vertical areas such as automobiles, aerospace, and robotics. This is believed to be the development of a camera for autonomous vehicles.
However, this does not mean that Softbank has not invested in laser radar. The most typical one is the Israeli startup Innoviz. Softbank has led its B round of financing, and Innoviz's first product is HD solid state laser radar (HD- SSL).
In terms of software, Softbank's Vision Fund led the $164 million C-round financing of Maps Mapbox. Today, Mapbox collects more than 200 million miles of anonymous sensor data every day and processes it in real time to update maps in real time. Updating this item in real time is valuable for the high-precision maps that are needed for autonomous vehicles.
For autonomous driving, various intelligent algorithms are important. In this field, Softbank has taken a look at Nauto, a startup that focuses on autopilot data collection and related software development. At the same time, Brain Crop, a company that develops robotic navigation technology, has also received $141 million from Softbank and Qualcomm. investment.
Brain Crop's main business is to control robots to navigate using visual cues and landmarks, allowing robots to see and avoid people and obstacles. The use of software to handle chaotic and unstructured environments has something in common with the "last mile" of autonomous vehicles.
(The secret agent sorts, the copyright belongs to the Silicon Valley secret agent)
In this way, only V2X technology, Softbank has no relevant investment behavior. Simply put, V2X refers to the technology that the vehicle interacts with the surrounding mobile traffic control system. X can be a vehicle or a traffic light such as a traffic light. The purpose is to help the self-driving vehicle to grasp real-time driving information and road condition information. This is also the key link for the autonomous vehicle to move toward the driverless stage.
However, the agent found that Softbank is not doing nothing in this piece. For example, the R&D department of Honda has been working with it since November last year. It is aimed at car networking technology and hopes to connect cars, moving items and various other things. .
Why vote: the king strategy
What does it feel like to have 16 lenses on one camera?
This is Light, the startup company of Palo Alto in the Silicon Valley Universe Center. They installed 16 cameras on a smartphone-sized camera. In the later stage, high-quality images can be automatically synthesized by algorithms, with more than 50 million pixels.
(Image from Light, copyright belongs to the original author)
On the 18th of last month, Softbank's Vision Fund announced that it had placed $121 million in Light D. Prior to this, Light's CEO Dave Garnan only met Sun Zhengyi three times.
“At our second meeting in Tokyo, he (Sun Justice) said, 'I believe the most valuable part of this product is to use autonomous driving.'” Dave Garnan recalled the scene of meeting with Sun Zheng.
Although this vision fund is invested in a company that manufactures precision lenses, it is precisely the ambition of Softbank for the autopilot layout.
If it is said that investing Light and giving the startup a new direction is Sun Zheng’s need for autonomous driving, then Softbank’s direct investment of $2.25 billion in news for GM’s autopilot company is even more straightforward.
At the end of May this year, the same vision fund of Softbank announced that it would make a strategic investment in the auto-driving company Cruise, which was previously acquired by GM. The investment will be made in two parts: the first $900 million at the end of the transaction; once Cruise Auto-Drive is ready for commercial deployment, Softbank will complete a second $1.35 billion investment.
Softbank Management Partner Michael Ronen has revealed that Softbank has been focusing on the autonomous driving industry for several years. One of the reasons Softbank decided to invest in General Motors and Cruise was the shock of the ability of the Cruise team to quickly iterate.
This is true.
If the combination of auto-driving is mostly the veteran giant + startup company, although GM is much smaller than Waymo and Tesla, it is undoubtedly going fast and good. In the past year and a half, the GM autopilot has iterated four times. Judging from the product image of the fourth exposure, there is no steering wheel, and there is no brake in the car. It can be regarded as "unmanned" in the true sense.
From the perspective of smart driving technology, GM is equipped with the L2 level Super Cruise technology on existing production models such as the Cadillac CT6. GM claims that this is the industry's first mass-produced L2-class intelligent driving system that allows drivers to release their hands on the highway.
(Image from General Motors, copyright belongs to the author)
At the beginning of this year, research institute Navigant Research released a list of autonomous driving competitiveness, from the company's vision, market development strategy, partners, production strategy, technology, capacity, product quality and reliability, product mix and long-term investment. A comprehensive evaluation of ten dimensions. GM is the number one player in the global autopilot field, followed by Google Waymo, Daimler-Benz-Bosch, Ford and Volkswagen.
(Image from the web, copyright belongs to Navigant Research)
As the Secret Agent reported earlier, the vision fund's investment is a Kingmaker strategy. In addition to making the king, Softbank does not hesitate to start with the existing kings, such as Nvidia and ARM.
Back in the autopilot industry, Softbank has concentrated on the field of travel. From the Southeast Asian version of "Uber" - Singapore company Grab to India's shared car company Ola, then to Brazil's 99Taxis, plus Didi and Uber, these representatives are leaders in different regions of the world.
“The leaders we cast on those tracks are not from the first to the fourth.” In a conversation at Fortune, Lydia, an investor from SoftBank, said.
In addition to investment, Softbank Group also personally set up a research and development team to develop autopilot technology, but only for electric buses.
In April 2016, Softbank Group and Advanced Smart Mobility jointly established SB Drive to focus on the development of self-driving buses. On the 4th of this month, SB Drive and Baidu Japan reached a cooperation, and 10 Apolong mini electric buses jointly developed by Baidu and Xiamen Jinlong United Automobile will be shipped to Japan early next year.
(Apollon mini electric bus, picture from SB Drive)
Where is the next opportunity?
The Brookings Institution has reviewed the investment transactions of different participants in autonomous vehicles or core technologies over the past three years (from August 2014 to June 2017). The study found that there were 160 investment events in the autonomous driving field with an investment of US$80 billion.
According to previous statistics, Softbank's investment in auto-driving related fields in the past three years is far more than US$6 billion, which does not include the travel market and the surrounding markets of auto services.
Compared with traditional cars, autonomous driving has higher requirements for the electronicization of automobiles. Therefore, fundamentally, the final performance of autonomous driving depends on AI chips, sensors, intelligent algorithms and so on.
However, the secret agent combed Softbank's investment from 2014 to the present, in addition to the investment in the upstream software and hardware of various types of vehicles, the major travel companies and manufacturers in the midstream, the service chain of Softbank in the automotive industry did not fall:
Softbank has invested in the A and B rounds of the platform company YourMechanic, which is communicated by a car mechanic to the car owner; invested in the B and C rounds of a car auction platform ACV Auctions, and the Neighborhood Fuel and airport involved in mobile refueling services. Many companies such as FlightCar in the field of car rental. From the perspective of the entire automobile industry chain, these enterprises are more inclined to value-added service providers downstream of the industrial chain.
“In less than 50 years, people will no longer be allowed to drive on their own in the city or on the highway unless they have a special permit. Because the driver may cause traffic jams and accidents.”
"In the future, people will drive like an old-fashioned sport."
On the 19th of last month, as the head of the Softbank Group, Sun Zhengyi, the CEO of the 100 billion dollar vision fund, made a prediction without hesitation in the annual meeting of Softbank. Standing next to him, Daniel Ammann, president of General Motors, and Liu Qing, CEO of Didi.
(Image from nikkei, copyright belongs to the original author)
When is human driving replaced by autonomous vehicles? Behind the conclusion of Sun Zhengyi is the autopilot industry where Softbank has already begun to lay out.
Today, the secret agent tried to sort out the investment situation of Softbank and its vision fund in the autonomous driving track, and analyzed why Softbank chose these companies, and today's startups have any chance.
Maybe, the next person Sun Zheng is eyeing is you.
What to vote for: automatic driving software and hardware are cast
Speaking of autonomous driving, what do you think of? Is Google's Waymo, Uber, or a traditional car company like GM, Ford? Whether it is a technology giant, a shared travel company, or a traditional car manufacturer, many companies are greedily competing for the auto-driving car market. But behind this battlefield, the most notable is the large-scale flow of funds from companies and venture capital institutions.
According to Pitchbook data, there were 69 investments in autonomous driving in 2017, totaling more than $4 billion. In 2016, 34 investment transactions were only $626 million. What position does Softbank occupy in it?
Before combing Softbank’s investment in autopilot, the agent may wish to