Below is the transcript of a CNBC interview with Qiu Heng, President of Global Marketing and CMO for the Enterprise Business and CNBC's Arjun Kharpal. The interview took place at CNBC's inaugural tech conference, East Tech West, in Nansha, Guangzhou.
AK: Thank you very much, Mandy, and I'm hoping Qiu's going to be able to tell us about some secret cars with lasers on them, and-,
AK: Controlled by robots, but let's find out what's going on over at Huawei. Qui, you talk a lot about Huawei's vision, when it comes to the way that AI's going to change enterprises going forward, so in your opinion, what does that mean? How will it change enterprises?
QH: Okay, because, first-, because most people think of Huawei as a smartphone company, I need to clarify that I'm from Business-, Enterprise Business Group, and Huawei has a total around $100 billion US dollars of revenue, and for Enterprise Business Group, we double our revenue every two years, and this year, we'll reach $10 billion US dollars. And I'll come back to the AI topic, so this October, we just released our full-stack all-scenario AI portfolio. The target is to realize inclusive AI. What does inclusive mean? It means affordable, effective and reliable AI. It means that AI is not only for giant companies, it should be for everyone, for every industry. And full-stack means that our AI portfolio will include a chipset, chipset enablement, unify the training, and (inaudible) architecture, and the application enablement. So, everything needed for depth, then we call can use this to develop the applications very easily. And all-scenario means the portfolio will cover cloud, edge computing, industrial terminal, consumer terminal. Every scenario. So, it will be a full-stack.
AK: So, Qui, what you mean by full-stack, of course, is owning all parts of that chain required for a business to implement AI in to their company, essentially, and transform the business. So, let's dig in to parts of that stack. You recently announced some new chips-,
AK: The Ascend 310, Ascend 910. Can you just give us a brief outline of what these chips are for, what are they designed to solve?
QH: Okay. So, before talking about the chipsets, I would like to say why chipset is so important. Because we say for AI, three factors are very important: computing power; data; algorithm. But just to check which factor contributes the most. From the IDC data, from 2014 to 2020, the algorithm, for the recognition success rate, only 4% increase, for the data volume, 5.5 times increase, very encouraging, while for computing power, 30 times increase. So, the (inaudible) factor, computing power, contributed most to the recent upsurge of AI trend, so it's very important. Then, what contributes the most to computing power? Chipset. Okay? So, in October, we released two chipsets and a series of IP for chipsets. One is Ascend 310. Another is 910. What's the difference? The Ascend 310 is designed for edge computing, it means it's very low power consumption, because at the edge, you can't get a lot of power supply, while for 910, it's with the highest computing density in the world, you (inaudible) compare, it is with the highest computing density. So, both these chipsets will contribute a lot to increase the computing power, and to realize affordable, effective and reliable AI.
AK: And how do you plan to commercialize these chipsets, then? Because-, are you planning to sell these to, perhaps, some of your competitors? Because these are-, well, the 910 is a chip designed for servers and datacenters-,
AK: And so, you know, will you be making the servers to go in to datacenters, and selling that as a package? Or are you selling these chips separately, to become a chipmaker specifically for big data, and cloud and AI?
QH: Okay. So, first, we will not sell the chipset so there will be direct competition with chipset companies. We will provide the capability of the AI computing power in the format of-, first as an acceleration module. Our customer can buy the module, they can buy the acceleration card, they can buy an AI server, they can also buy the mobile datacenter we announced, together with Audi, for the autonomous driving, installed in the car, and also, they can buy cloud, we provide AI capability with Huawei Cloud, called EI, Enterprise Intelligence.
AK: How do you plan to-, because you're not the only player in this space, we've heard, earlier this year, Alibaba planning to make its own chips for datacenters, of course, you've got the likes of Nvidia, over in the US. How do you plan to compete with some of these other players in the space? What sets you apart?
QH: So, in fact, if you check the capability of our chipset and our offerings, you will see it's very competitive. I will give you several examples. For example, when we install the chipset 310 in to the camera, before, also an AI camera, it can only recognize 30 objects at the same time, while after we installed the chipset, it can be increased to 200 objects, including people, cars, obstacles, traffic, (inaudible), so autonomous driving and traffic control can be realized easily, and we will further increase to 600 objects at the same time, just one example. The other example is for the common training case for AI. Usually, it's very, very time consuming. In the past, Fast.ai set a record, to train a common picture recognition case with 18 minutes, that's the record, for the past. But now, with our 910, with EI, we can realize only 10 minutes, so we set a new record. And there are a lot of these kind of examples. And not only is it very powerful, it saves energy. I'll give you an example for MDC, mobile computing, mobile datacenter for the autonomous driving. All the chipsets inside an autonomous driving car (inaudible) from Huawei for Ascend 310, it only consumes 200 watts. 200 watts is nothing, it's almost a similar level as a stereo speaker, so you understand, it's very, very-, it saves the power consumption.
AK: And, Qui, apart from AI, it does link, in some senses, to what you're trying to do in 5G networks, as well, with optimization, 5G, of course, is a key technology that Huawei is looking in to. People hear a lot about 5G, that it's going to allow us to download movies in a matter of seconds, but is this actually more of a business-focused product, 5G, allowing some of the things you've spoken about, like the Internet of Things, and allowing businesses to use these increased cloud computing capabilities? What kind of impact will it have on business?
QH: Yeah, first, we say, in the (inaudible), it's already defined what 5G can be used for. First of all, extremely high speed, for people communication, and then for extremely low latency, Internet of Things, and then for massive connection, because the number of things is much more than the number of people. So, for these three kinds of scenarios. For Huawei, now we are doing trials, and pre-commercial verification, in 66 countries already, and we already shipped out more than 10,000 base stations supporting 5G.
AK: I want to talk about some of the challenges, however, around some of this technology, in particular AI, it's a conversation we've been having here at East Tech West, there aren't, of course, many opportunities, but there are potential fallouts from the widespread use of AI, and we've talked about the impact, perhaps, on jobs, and broader society, as well. So, how do you develop this technology responsibly, to make sure that the disruption that it causes is kept to a minimum, and those people that perhaps are impacted are able to reenter the workforce, perhaps in a different way?
QH: So, first, we are optimistic about the impact of AI to society. We think AI will help to improve the production value, improve the internal management efficiency, and also it may help the society become even more equal. Why? I can give you some examples. So, take our own business, first, as an example. We are doing the enterprise business, we double the revenue every two years, but we can't double our employee number every two years, right? So, the work is increasing very fast. How to handle all this work? I'll give you some examples. Our homepage supports more than 80 different languages. We have more than 12,000 documents. If we tried to translate all these documents in to 80 languages, you can imagine, by a human being, how much work we would be needing. So, in the past, we had to select only the most important documents to be translated, but with the help of AI, we can realize all language translation at the same time. So it means that for even the language without a big population, they can enjoy the full translation of all documents, the same level of service. So that's the first thing. Again, I will give you several examples of our customers. For example, in Shenzhen, just a nearby city, the traffic congestion becomes more serious, because for every kilometer of road, there are 510 cars. You can imagine, it's a very difficult situation. Now we are using AI for the automatic traffic control. In the past, the red light turned to the green, just according to some timer, but now it's different. The camera is watching, this line, how many cars are coming? That line, how many cars are coming? And they decide when to turn the light, from green to red, and from red to green. And it improved the speed of cars around 15%. That means, in the morning, we can sleep 10 minutes more-,
QH: It's very, very-,
AK: Very useful.
QH: Yeah, very, very valuable.
AK: Qui, I've got to ask you a question specific to Huawei, as well. You've got all these new products, you're rolling out 5G, you've got the new AI chips, products that are built with those chips. How worried are you about the political headwinds towards Huawei, particularly from the US, and your ability to sell these products to other customers, as well?
QH: So, first, now Huawei are serving global customers in over 170 countries, and our customers include 46 of the top 50 telecom operators, 211 global Fortune 500 enterprises, and hundreds of millions of consumers. So, why do they choose Huawei? Because they fully trust Huawei, and also, they recognize the value Huawei provides. So, we believe our partners and customers, all over the world, will make their own right choice, with their own decision, and their own judgement, and, based on their experience of working together, we will continue to serve our global customers with our innovative solutions.
AK: We've got about a minute left, Qui, so as we close out this session, I want you to give our audience an insight in to what you think is the one biggest change, or impact, that artificial intelligence is going to have on enterprises in the future.
QH: First, it will change all industries. AI will not only be used for some rich industry, or only rich customers. Our inclusive AI will make AI affordable, effective and reliable, in all industries we provide with the format of cloud, with the format of equipment, and it will dramatically improve the production efficiency, and reduce the cost. I think it will benefit to the whole society.
AK: Qui Heng, President of Global Marketing and CMO for the Enterprise Business Group at Huawei, thank you very much for your time.
QH: Thank you very much.
CNBC is the leading global broadcaster of live business and financial news and information, reporting directly from the major financial markets around the globe with regional headquarters Singapore, Abu Dhabi, London and New York. The TV channel is available in more than 410 million homes worldwide.
CNBC.com is the preeminent financial news source on the web, featuring an unprecedented amount of video, real-time market analysis, web-exclusive live video and analytical financial tools.
CNBC is a division of NBCUniversal. For more information, visit www.cnbc.com.