AI companies are raising AI chips | it says | 'No need to do'

After the popularization of the AI ​​concept, all parties are looking for business models, expecting AI technology to land as soon as possible, but most of the current AI startups are in the stage of relying on financing and burning money.

From last year to the first half of this year, many AI companies have announced that they have entered the chip field and launched their own AI chips, including Yunzhisheng. They went out to ask questions. Rokid, Spirent also announced that it will launch AI chips last month.

Is this story good to tell?

Not involved in the AI ​​chip

Founded in Shanghai in 2001, Xiaoyi Robot (Intelligent Intelligence) is a provider of artificial intelligence technology and industrialization platform, providing artificial intelligence core technologies including natural language processing, deep semantic interaction, and communication, finance, Government, e-commerce, medical, manufacturing and other industries with a deep integration of solutions and service systems, the most important product is the intelligent customer service robot (VCA). However, Xiaoyi robot CEO Zhu frequently told reporters that the company will not be involved in the chip field for the time being.

First of all, in terms of company positioning, Xiaoyi Robot is a technology company. Zhu frequently pointed out that AI is generally divided into three layers, the upper layer is the application layer, the middle is the technical layer, and the bottom layer is the base layer. 'We are a technology company, these Technology needs to have application scenarios to bring out the value of technology. We believe that the company is still moving toward technology. 'Based on this idea, on July 5th, Little i Robot launched the latest version of the intelligent interactive system iBot Pro (iBot 10) Professional Edition), and released intelligent customer service, smart finance, smart city, smart office, smart medical, intelligent robot, intelligent hardware, intelligent manufacturing 'AI+' 8 field solutions.

For the lower level of technology and computing power, he believes that it depends on the entire ecological chain, 'because we are semantic, we are mainly facing text, and the requirements of text for computing power are not as big as images and speech. So our urgency for the current computing power of the chip is actually not that big.

Secondly, the chip is very difficult and has its own ecology. In Zhu frequency, there is no need to make chips by itself. Now, with the development of the ecological chain, the chips that are optimized for deep learning are also coming out. Based on these chips to develop our applications, there is no need to do their own independent things.'

In addition, from the business model, he believes that the purpose of making chips is to reduce the size, power consumption and cost of the chip. To achieve low cost, a certain amount of sales must be achieved. 'This means the application scenario you are going to do. It must be a standard scene that can be copied, not too specific. Once you find a very standard application scenario, everyone comes in, the giants are coming in, whether it is NVIDIA or a variety of chips, even Baidu. The AI ​​chip was released. In this case, it is not what our company can do.'

He pointed out that the strategy of the enterprise is different from the development of the industry. The development of the industry can see how many executives are developing in the industry. The corporate strategy needs to combine the capabilities and stages of the enterprise itself. 'At present I will not do chips.'

From loss to profit

Such a strategy is also inseparable from the actual development of the company. In the first half of this year, the company delisted from the New Third Board.

From a loss of more than 70 million in 2016 to a profit of more than 40 million in 2017, Zhu frequently analyzed his exploration path to reporters. Although the field of intelligent customer service has brought small and medium-sized companies the cash flow of continuous research and development of artificial intelligence, it is difficult to single-field. Do it on a large scale quickly.

2017 is the first year of Zhisheng's intelligent profit. The company's annual income is 2-3 billion yuan, and its profit exceeds 40 million yuan. In 2018, the target is 500 million yuan and the profit is 100 million yuan. According to the previously disclosed financial data, In 2016, the company's net loss was 73.351 million yuan. In the first half of 2017, Zhisheng's intelligent revenue totaled 65.116 million yuan, with a net loss of 14.65 million yuan. The second half of last year was an important time point for the company to turn a profit.

Why is there a higher profit in the second half of the year? Zhu frequently pointed out that for the 2B industry, the revenue in the first half of the year is usually 1/3 of that in the second half of the year, 'this is related to the decision-making process of large enterprises facing the big ones. Planning, implementation and launch in the second half of the year, so this will cause most of the contracts and revenues of these enterprises, especially telecommunications and financial companies, to be confirmed in the second half of the year.

Zhu frequently said that in the revenue of more than 200 million in 2017, revenue from the government and financial industry each accounted for about 20%, and other industries generally average 7%, 8%. In the 20 billion revenue target of 2018 Smart customer service, smart city and smart finance will still be the main source of income. And smart office, intelligent robot will have a process of growth, 'may produce explosive growth'.

AI+ commercialization exploration

After the popularization of the AI ​​concept, all parties are looking for business models, expecting AI technology to land as soon as possible, but most of the current AI startups are in the stage of relying on financing and burning money. From 2016, the loss of more than 70 million to 2017 profit of more than 40 million, Zhu frequently analyzed his exploration path to reporters.

Zhu frequently pointed out that in 2016, the company originally hoped to be a strategic emerging board in Shanghai, and wanted to be the leader of the sub-sectors in the strategic emerging industry. 'As long as you are the leader, whether it is a loss or not, as long as you have the status of the industry, you will You can go to the capital market. If you can, in this way, there is of course a better driving force for sustainable development. ' He admits that the company's strategy in 2016 is to fully develop and invest heavily. Whether it is sales investment, industry development The investment and technical investment are very huge, especially in terms of hardware. 'We are actually a software company, we are not very good at hardware, but at that time it may be hot.'

In addition to the hardware, there are other commercial investments, which caused the company's huge losses in 2016. Zhu frequently said that after the cancellation of the strategic emerging board, the company lost a stage of development goals. It is at this time, realized Businesses need to be able to continue to grow and have the ability to continue to operate.

In 2017, the State Council issued the “New Generation Artificial Intelligence Development Plan.” Zhu frequently pointed out that 'although it is very important for national artificial intelligence, many places think that artificial intelligence can't be shown to show off every day. To generate value, it needs to be integrated into the industry. Since the second half of last year, the topic of artificial intelligence industrialization has also become a hot topic. He believes that the strategy of the small i robot is to shrink and focus on its core competence. He said that he subsequently cut down the hardware. A lot of non-core businesses, and to a greater extent emphasize the deep cooperation with industry partners.

In addition, he also talked about the high valuation of AI startups. 'In fact, many AI companies are very expensive now. Why would everyone want to invest in such companies? They think that the threshold of artificial intelligence technology is relatively high, and they can continue to invest. Formed a technical threshold, especially to a certain turning point, capable of forming a scale capability.

According to the Gartner report, global business intelligence related to artificial intelligence in total will reach 1.2 trillion US dollars in 2018, an increase of 70% compared with 2017. By 2022, artificial intelligence related business value is estimated to reach 3.9 trillion US dollars. 2018 is artificial intelligence In the year of technology outbreak, the growth curve is very steep.

2016 GoodChinaBrand | ICP: 12011751 | China Exports