Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Toru Nishikawa (right), co -founder and chief executive director of preferred networks, and Daisuke Okanohara, co -founder and chief executive director, with Fanuc Corp. Robotic ARM at their company’s headquarters in Tokyo, Japan, March 16, 2018.
Bloomberg | Bloomberg | Gets the image
The preferred networks are a little-known name outside Japan, but a unicorn supported toyota, a unicorn has big plans to use deep learning to correct the “real world”-to go to global.
“In the last 10 years we have been working on using AI to solve problems in the real world”, Chief Executive Director Daisuke Okanohara told CNBC “Managing Asia“ The company dared into industries, as well as a variety of freight, health and robots.
The preferred networks have attracted the attention of great names such as a car giant Toyotawhich in 2017 invested 10.5 billion yen ($ 95.4 million) in the company except the initial 1 billion yen in 2015.
Preferred networks are one of the few unicorns in Japan. Although Japanese Japanese ecosystem has grown, Japanese companies worth more than $ 1 billion make up only 0.5% of the total world, According to the International Workers’ Document of the Monetary Fund Posted in December.
Launching deep learning is also in Joint venture with the Mitsui & Co. in the field of autonomous ride. Both have been invested in the company, T2 to solve the problems related to the auto -traditional, including the “extremely long” working time, and the driver’s pool is reduced, said Tor Nishikov CEO.
But the competition in the automotive industry was also a factor. Recognizing the current restrictions on the company, Okanohara added: “If we solved the problem of ordinary cars, it would be extremely difficult. There are many companies working on autonomous management and this is a very competitive sphere.”
Since then, the AI ambitions of the prevailing networks have gone much further. But compared to other AI decisions in digital space, in which “relatively easily achieved” with users “, fighting problems in the real world requires longer, Okanohar said.
Asked how much time it takes to make money technology, Nishikawa said it could take as many as five years in an area such as detecting AI -based materials.
“First, we start joint research with companies, and then evaluate whether it can be commercialized. Once we determine that it can be commercialized, we are working on the creation of services and products that can be provided to customers,” he said, adding that this process is about three years. “
As for the latest developments AI, Okanoharo said the predominant networks have recognized “very good technologies” Deepseek and “paid close attention” before it was catapult to glory, adding that “different players” would soon enter the branch.
Preferred networks hope to be among them. By recognizing this Nvidia’s The AI generative strategy is still underway, but has a graphic processing unit that is not yet “optimal”, Nishikov said that the preferred networks are in the midst of “creating more advanced processors” to get a competitive advantage.
IPO plans
Whether the company will be able to achieve the scale of its ambitions will still have to be learned. According to Nishikov, he looks at foreign investors and receives “many requests in the field of entertainment, semiconductors and computers”, which noted that “there are restrictions that we can only do in Japan, and the world market is much greater than the Japanese market.”
Getting an initial public offer is one of the goals, he added. When asked about the terms, Okanohara said: “I think the need for huge means will come at a time when we present our hardware products such as semiconductors, the world. So I think we will strive for the public in these terms,”
