Hello ladies and gentlemen and thank you for standing by for Baidu’s fourth quarter and full year 2020 earnings conference call. At this time, all participants are in listen-only mode. After management’s prepared remarks, there will be a question-and-answer session. Today’s conference is being recorded. If you have any objections, you may disconnect at this time. I would now like to turn the meeting over to your host for today’s conference, Juan Lin, Baidu’s director of investor relations.

Hello everyone and welcome to Baidu’s fourth quarter and full year 2020 earnings conference call. Baidu’s earnings release was distributed earlier today and you can find a copy on our website as well as on newswire services. On the call today, we have Robin Li, our co-founder and chief executive officer, Herman Yu, our chief financial officer, and Dou Shen, our executive vice president in charge of search and feed. After our prepared remarks, we will hold a Q&A session. Please note that the discussion today will contain forward-looking statements made under the “safe harbor” provisions of the U.S. Private Securities Litigation Reform Act of 1995. Forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from our current expectations. Potential risks and uncertainties include but are not limited to those outlined in our public filings with the SEC, including our annual report on Form 20-F. Baidu does not undertake any obligation to update any forward-looking statement, except as required under applicable law. Our earnings press release and this call include discussions of certain unaudited non-GAAP financial measures. We have made minor adjustments to our non-GAAP measures and retroactively applied these changes for comparison purposes. Our press release contains a reconciliation of the unaudited non-GAAP measures to the unaudited most directly comparable GAAP measures and is available on our IR website at ir.baidu.com. As a reminder, this conference is being recorded. In addition, a webcast of this conference call will also be available on Baidu’s IR website. I will now turn the call over to our CEO, Robin.

Good morning, everyone. Baidu delivered another solid quarter in Q4, with Baidu Core revenue reaching RMB 23.1 billion, or US$3.5 billion, up 6% year over year and up 8% quarter on quarter. The latter was much higher than our historical flattish or low single-digit growth from Q3 to Q4. In addition to the improvement of macroeconomic conditions in China, we are benefiting from the digitalization of industrial Internet. Baidu Core non-advertising revenue in the fourth quarter grew 52% year over year to an annualized run rate of USD $2.6 billion, driven by the convergence of AI solutions, cloud services and consumer Internet. Such convergence is expected to accelerate into the next decade and Baidu stands to benefit from it. Let me elaborate on the trends that we are seeing, starting from smart transportation. China aims to raise its global competitiveness in the transport sector through aggressive infrastructure building. The Chinese government is promoting a multi-year plan to install smart transportation networks in cities with sensors and cameras on public roads, connecting vehicles with road side units to foster safer driving, improve throughputs, and reduce carbon emissions. Such government initiative encourages municipal governments to shift their budgets toward a more software defined transportation network. Apollo’s strong brand in intelligent driving and our holistic approach to smart transportation, make us well suited to capture the opportunities in smart transportation infrastructure development. For example, Baidu signed a strategic partnership with Guangzhou in southern China, to deploy the world’s first multi-modal autonomous driving Mobility as a Service, or MaaS, platform. Utilizing Apollo robotaxi and robobus fleets, Guangzhou begins this month to allow commuters in the Huangpu district to call up public transport services on demand, in lieu of taxis, ride hailing and other traditional gasoline-powered public transport services. The Guangzhou smart transportation project takes…

Thank you, Robin. Hello everyone. Welcome to Baidu’s fourth quarter 2020 call. All monetary amounts used in my discussion are in RMB, unless stated otherwise. Despite an unprecedented year, our business has returned to growth and our profits were solid in the fourth quarter. Baidu’s online marketing is widely diversified, covering many industries in the Chinese economy, and we are benefiting from a rebound of the Chinese economy. Concurrently, our non-advertising revenue is growing rapidly, especially from the new AI business. Non-advertising revenue was up 52% in the fourth quarter, reaching 18% of Baidu Core revenue. We recently conducted a TAM exercise, as part of our strategic planning process, and found that the TAM for our non-advertising business, including Internet value-added services, cloud services and intelligent driving, before adding autonomous driving, was 10 times the size of our online marketing TAM and the expected CAGR to 2025 is 3 times that of our online marketing TAM. The potential for our non-advertising business is quite significant. For example, we recently signed a smart transportation project, whose contract amount was more than half a quarter’s worth of our auto online marketing revenue. There are over 100 cities in China with a population of over 1 million. When you consider a robobus project recently experimenting collecting 25 RMB for each ride, if fare collection becomes feasible, we could see an increase in robotaxi demand, which would drive our MaaS revenue under transportation cloud, the service component of fleet management. For fiscal 2020, Baidu Core revenue reached US$12.1 billion, and we spent 21% on research & development, which is a testament to our commitment to technology, especially in the area of research. Our consistent heavy investment in technology has made Baidu PaddlePaddle the No. 2 deep learning framework in the world, Baidu open AI…

[Operator Instructions] Your first question comes from the line of Alicia Yap from Citigroup.

My question is related to your partnership with Geely. Can you share with us some colors on the expected timeline for this new EV to be released and what Baidu want to achieve out of it? Will Baidu also continue to seek for partnership with other OEMs to co-produce other EV? And just related to that is it seems like Geely is also working with multiple AD companies and also working on developing their own ADAS. So how would you describe the partnership and your relationship with Geely? Thank you.

Hi Alicia, this is Robin. Let me answer your question. Our partnership with Geely is based on the belief that end-to-end integration, hardware and software, will provide the best experience for autonomous driving or smart EV. And in the past, we have developed a lot of very advanced autonomous driving technologies. We try to push this to integrate into our OEM customers. But most of the traditional OEMs tend to wait until someone else adopts the newest and latest technology. That's why we decide it's time for us to just build a benchmark vehicle to show that how good the experience is. It also serves the purpose of better and faster and stronger feedback loop so that we can improve our self-driving technology faster. Right now, the venture is progressing very well. We have a CEO on board, and we have decided on the brand of the new vehicle. As you know, it usually takes around 3 years from the beginning to the launch of a new EV model, and we try to do the same for this new venture. On the other hand, Apollo and Baidu Brain, in general, remains to be a very open platform. We will continue to work with other OEMs to provide the technology and the products to them whatever that is needed. We will serve them to our best efforts. And in the past, it's just without someone first trying, they tend to wait. Now we have a flagship where we have a benchmark they can follow. I'm sure it will also help the revenue of the Apollo technology provider products of autonomous driving.

Our next question comes from the line of Eddie Leung from Bank of America Merrill Lynch.

Just have a follow-up question on AI cloud. I remember one of the key reasons of the high AI cloud is that it should have better margins than some of your competitors, which are focused on some of the more infrastructure cloud services. So just wondering if you can offer some not necessarily a very quantitative guidance, but some marginal outlook for your AI cloud thesis given some of your competitors already achieving a breakeven status.

Yes. Eddie, Happy New Year. And this is Herman, I'll take that. So for cloud, we have several pieces of business in our cloud. We talked about our enterprise cloud, which is IaaS, but we differentiate in the market, which -- with AI solutions such as PaaS and SaaS. And clearly, when we're able to do PaaS and SaaS, we see there's much more profitability to be made over the IaaS. But what we have to recognize is that when you're going into a software space of PaaS and IaaS, initially, it might not be very profitable. For example, the AI solution that Robin talked about, when we did our first solution a couple of years ago, that was actually a very, very poor margin because you're going into a new environment, you might have a good product, where you've got to spend a lot of effort to integrate with your existing legacy systems. So as you build that out, as you're selling to more customers and so forth, and you make your product more standardized, then the next few versions with this allows your profit margin to increase. So because we're growing so fast and we're trying a lot on the product, we would not expect profit margin at this stage to be very profitable. But when we look at our product life cycle, such as the automated AI solution, customer solution, call center, we can see that over time that, that margin trends up. So to answer your question, I think on the one part on enterprise cloud and so forth. One, our margin is going to be a function of how much is PaaS and how much is PaaS and SaaS and how much is IaaS. I think IaaS over time, because we're differentiating with SaaS and PaaS, over time when that proportions, things go bigger and bigger. I think our profitability would be better, especially when products mature. And then another big part of our cloud is, for example, a vertical cloud such as smart transportation. And again, that's a principle of product maturity. That's a principle of how -- what mix of that is software and so forth. I think, overall, all of this would be very helpful. And just to answer your question, some of our competitors might be profitable and so forth. I think it really depends on the business that you want to be in. And unfortunately, we don't have the luxury of some of our peers where they can just allocate expenses right to their other businesses and so forth because the infrastructure is the same. So I think, profitably, I think, one, is how you do allocation of expenses. Another is, are you focused more on software? Do you have a differentiation in the software sector, such as our AI solutions? And then how mature your product is over the product life cycle. I think these are determinations of future profitability.

Our next question comes from the line of Piyush Mubayi from Goldman Sachs. Please ask your question.

Congratulations, Robin, Herman, for a splendid quarter, a great outlook. Can I just go back to the intelligent EV company that you talked about, Robin? And you talked about the 3-year lead time to get a car to market under the name. So you selected a name, let's call it, Baidu car for the time being. Why would it take 3 years for that to happen, first? And second, when you launch that car, and I'm just trying to fish here also. What segments are you targeting? And are you targeting a certain market share in this explosively growing market? And we're trying to get a better sense of how this will evolve with time. And while this is happening, what happens to the revenue line as we start to see accrue from the ANP in particular? So if you could give us a sense of what pricing could be like with products like the ANP that has just gone, I suspect, live right now. And I'll stop there.

Piyush, yes, as I mentioned that the reason we're doing this smart EV that we want to bring the best of our technology to the market. So the selling point -- the main selling point of this new car model will be more on the software side or the software and hardware integration, so that has a better autonomous driving capability as well as better infotainment because of DuerOS for auto, et cetera, et cetera. We will announce the price range or demographic target when we think it's mature enough when we are really ready for that. But that means that you can count on that. We will try to speed up the process to launch this new car as soon as we can. And -- but it will be the best car on the marketplace with best autonomous driving technology and that infotainment for the smartest car. People generally call the class EV. But for us, the real point, it's not electric. The real point is intelligent. And we are best in terms of the intelligent driving. And we will fully leverage this kind of capabilities.

Yes. And let me elaborate that, Piyush, when -- just to go a little bit deeper, give you more color on the meaning of intelligence. So when you look at EVs, right, there are several components for example, there is the ASD Apollo self-driving services that we talked about. That would be HD Map. That would be valet parking. That would be navigation pilot. Stuff like that, right? And our model is different than most of EV manufacturers where we license OEM, right? So we can spend that R&D. And then because our market is not just limited to how many cars we sell, it's open platform so we sell to many OEMs. That's number one. Number two, in addition, being able to power the operating system of infotainment, that's very key because that requires a big ecosystem where you have a lot of content made. So DuerOS for auto, as we mentioned, we've already installed for over 1 million cars in 2020. And again, that's the open model we're working with several OEMs. So we have the edge in these 2 areas. In addition, we're doing autonomous driving so we can leverage that technology. And more importantly, you've got to realize we're doing this on an ecosystem basis. So on the one hand is making the car more intelligent from ASD, from connected vehicles, from autonomous driving. Another is the fact that we're doing a V2X with smart transportation. Because a lot of the key data is being gathered on our transportation cloud, whether it's on the road, whether it's through the maps or whether it's connected vehicles and so forth, we can understand the road better. So all of that data and that capability will then empower our EV. So 2 things. Number one is we think our EV can be more intelligent because we're doing other components and because R&D is centralized and it supports other manufacturers. And secondly, because of our ecosystem strategy, we were coming up from a transportation angle, and also from the different components.

Our next question comes from James Lee from Mizuho.

Congrats on an excellent quarter here. On the smart transportation side, can you guys maybe provide some framework on the government's plan in terms of building out across the country? Should we think about that build-out being maybe top 30 markets? And also maybe help us understand the bidding process for city contracts a little bit. Do they choose multiple vendors and tech standards to do testing? And what is your key advantage compared to other operating systems? Is it your experience in autonomous-driven miles or your ability to offer multiple services on the platform? And also lastly, maybe, Herman, you can address this. Can you talk about the level of investments that is needed to make for smart transportation and also for autonomous driving over the next few years? And would you be breaking out the core advertising margin separately so we can also see the organic growth. Sorry for long-winded questions.

James, in terms of smart transportation, we are the pioneer of the ACE concept, which means that we are the first company to really consider transportation as a whole, using more software-centric solutions to improve the throughput to connect the road with vehicles and to help reduce accidents and reduce carbon emission. Yes, when we talk to the potential customers of all kinds of different tier cities, they will obviously compare and invite multiple vendors to bid on those kind of projects. But in almost all cases, we're so much ahead of any competition. Essentially, if they care more about the effectiveness of those smart transportation projects, we win. So we're so much ahead because Apollo system is very open, and we've invested in this area for many, many years. And our AI capability provides a very solid foundation, and we have been doing a lot of customization for smart transportation. As Herman mentioned previously, initially that the margin could be low. But when we repeat this kind of project city after city, the margin improves in the overall gross margin for smart city, smart transportation projects keep improving. It's much better than the apps and CDN business off of a lot of cloud services. On investments:

Yes. So the good thing is, when you look at our financial results, the amount that we're spending for autonomous driving and also for smart transportation is built into our model. So smart transportation is part of our AI cloud. So what you're seeing is already built in there. And you mentioned what does our margin look like for our core business. I suppose you're asking for mobile ecosystem. So for mobile ecosystem, our margin on that on a non-GAAP basis, usually is between mid-40s to high 40s. And we've been pretty consistent in the last couple of years. Sometimes when revenue -- in a higher quarter, we might reach a little bit over 50, but usually, it's around that range. And we currently expect to be in that range right now. So you can see, as I talked about, the V-curve growth, mobile ecosystem will try to maintain within that range. And then for our new businesses and so forth, we look at it differently because that is more about are we funding right to optimize growth for the next couple of years.

Our next question comes from Jerry Liu from UBS.

So my question first is just to go back to the revenue guidance for the first quarter, definitely a bit stronger than we even expected. So wanted to look at what are some of the key growth drivers here. On the non-ad side, I would assume some of this is related to cloud and smart transportation. I'm wondering if there's any other key growth areas. And then secondarily, in the core advertising business, do we chalk this up to improving macro? Or are there some maybe different dynamics in terms of competition or products? Thank you.

Yes. Let me answer it in different components. We talked about how non-advertising is now 18% of our revenue and in Q4, it's growing robustly 52%. We expect non-advertising to continue to outgrow our overall Baidu. So meaning it's going to -- because a lot of these are new businesses, it's going to be able to pull the weight. Smart transportation is part of cloud, and we talked about it's over 60-some-percent growth in Q4. So I think that with our strong offering and so forth, I expect this to continue into 2021. But do recognize that there is seasonality, for example, when you think about AI cloud, Q4 tends to be stronger than Q1. Just so how the budget works and so forth. But overall, I think non-advertising is going to be driving us not only in cloud, for example, in our other growth initiatives, such as DuerOS and so forth. We think that these are all going to be engines of growth. Piyush talked about earlier about our Apollo self-driving services. That -- while it's a small base, but that growth also helps. So it's all of these things. And then with regards to our advertising business, I think several things. I think, number one, is the economy is coming back. So the macro helps certainly. But more importantly is Baidu App. We talked about in-app services several times already. Because we're seeing user behavior switching from browser into in-app services because we have AI building block, we're the only one at this scale in China where the app was designed as a search app and it's open. It allows worldwide web search. It allows content from all different apps and from all different websites. Because of that and because the content and services on our platform making the closed-loop native app experience, the best, we think, in China that the AI building blocks provides, we think that, that helps user behavior. And because of great growth and so forth, that helps drive our revenue. And then also because it's native app, it's on the platform. Obviously, CPM helps. So you have traffic growth, you have CPM, and then also you have the macro as a backdrop helping.

Thank you. Next question comes from Alex Yao from JPMorgan. Please ask your question.

Good morning management team. Congratulations on a fantastic quarter. So I have a question regarding the YY integration. Robin, I think you mentioned you guys plan to start integrating YY operation into your own content ecosystem in the coming quarters in your prepared remarks. Can I confirm that the deal is still going ahead with the previously announced valuation despite of the money order report and the changing regulatory environment of live broadcasting business in China? And then I think you also mentioned new business opportunities will gradually unfold as you guys integrate YY's live streaming operation capability into your content ecosystem. Can you talk us through some of the key addressable market and the business model that will potentially unfold with the integration process? And lastly, perhaps, Herman, can you talk about how should we think about the financial impact in the coming quarters as you guys consolidate business and integrate the operation into your content ecosystem? Thank you very much.

Alex, let me first answer the business model question, and I'll let Herman to answer the regulatory and financial guidance question. In terms of this model, we basically view live streaming as yet another new form of content and also new form of monetization. For content, live streaming is comparable to text images and videos including long and short form videos. Live streaming is just a new form of media. It serves the users best when you combine all kinds of different media forms and provide them the right form of media whenever necessary. And YY apparently provides a very vibrant and also mature ecosystem for live streaming content. Although they are more show-based content, but the capability can be transformed into other forms of live streaming, including knowledge or e-commerce related live streaming. And in terms of monetization, we view that the Baidu platform as a whole, mobile ecosystem with hundreds of millions of users logged on to our platform on a daily basis. And we increasingly provide more native app experiences. Super majority of users logged in so that we know who they are, and they can have a better experience going from here to there. Right now, the large user base is primarily monetized through advertising. This is sort of unusual when you compare with other large mobile ecosystems. They typically have well more than 50% of their revenue coming from non-advertising. And for Baidu, in addition to advertising or online marketing, we can monetize our user base through live streaming, through subscription, through e-commerce and many other forms of revenue stream. That's the rationale behind the YY acquisition.

Yes. I mean I just want to clarify the transactions due going ahead, despite of the Muddy Water report and the changing regulatory environment. And also the deal still going ahead with the previously announced valuation?

Yes. Yes. So you asked quite a mix of questions. With regards to the Muddy Water report, I think JOYY has come out with a statement. As you guys know, they hired a global legal team, and they also hired a Big Four auditor, and they have made their own announcement. That's a JOYY announcement, not a Baidu announcement. So I think looking at that, what would be the best. With regards to a regulatory environment, we are cognizant of the situation. And that's why we think that and wanted to navigate. As you know, you have a lot of experience looking at this. As you know, whenever there's a regulatory environment that you have to steer towards, it's more important to have a bigger leading player in China who have good relationship with the government, who knows how to do things correctly and so forth. And we needed an experienced team. And that gives us a better reason if we're going into a live streaming, we will buy a large experienced team. So this is more of an acquisition of a team rather than a particular product or just a business. It's the team talent is very important. With regards to the contribution and so forth, currently, we're moving ahead. We're planning to integrate onto our platform, as Robin has said. And we're currently dealing -- addressing with some conditions and procedures. And we still is on pace to do the closing. As we announced previously, we're moving ahead with this. And the revenue is usually between CNY 2 billion, CNY 2.5 billion a quarter. So when we can confirm all of this thing that we think we'll be able to come back with that revenue amount. And with regards to pricing and so forth, yes, our final agreement was the original price. We do have stipulations in the agreement to protect Baidu shareholders. As you mentioned, there was money order report, the Big Four, of 2016, came in did not see anything. The global attorney came in, did not see anything. But to protect us, we do have indemnification in there to get a recovery just in case if something blows up in our head as a result of situations and so forth.

Next question comes from Gregory Zhao from Barclays. Please go ahead.

So we know now Baidu has several different business lines, cloud, autonomous driving, mobile, smart devices. So in the daily operation, just want to know how the management team allocates for time spend and efforts among all the different segments and also your current priority. And a quick follow-up on your ADS system. So you mentioned the partnership with some leading automakers. So in the future, what's your best guesstimate of the pricing of the systems? Do you think the existing products such as Tesla's SSD is a good benchmark for pricing? Thank you.

Greg, in terms of time management, I would say roughly half-half. Half of our time and resources are invested in the mobile ecosystem part, and the other half in the AI cloud and intelligent driving and other growth areas. Because it's a relatively diversified portfolio of businesses, we try to set up more flexible mechanisms or incentive structures so that the team can be more independent and driven. For example, the smart EV company, we have a strategic partnership with a very separate incentive structure for the team. And we recently raised money for our smart device company. So going forward, we will continue to do this. On the Apollo smart driving services, we will continue to provide technology at a very open and flexible manner. I talked about we offer HD Maps, we offer AVP, we offer ANP, and potentially, we offer other combinations of this kind of technology and services. The pricing range could differ based on the period of the contract and products we offer. And it's growing very quickly, and it's still evolving very quickly.

Yes. And let me just add a little, Greg. I think Robin was very modest with regards to answering your talent question. Robin talked about how he spends his time. But more importantly, as you may know, over the last 2 years, we have significantly strengthened our bench strength. So when you look at, for example, the lead of mobile ecosystem, Dou, he's a PhD. And we not only have (inaudible?) from -- very young, joining us 6, 7 years ago. And also, if you look at over the last 12 months, we've been aggressively hiring good talents underneath Dou. If you look at Dr. Wang, again, he's a PhD, and we have been helping him build his team underneath too. So -- and you look at, for example, our EV company, we're looking for a CEO who's out in the industry. So in addition to the times -- Robin time and people from group management, we do spend a lot of time building talent. And I think that's why Robin is versatile, being able to manage all these businesses because a lot of it is on building that middle stream in Baidu for each of the business groups so that they can run more autonomous, as Robin had mentioned, centers and so forth.

And next question comes from Natalie Wu from Haitong International.

Thanks for taking my question. So my question is regarding the autonomous driving. So Apollo is absolute leader for the third-party service providers in that area. But just curious, what are the closest competitors you can think of regarding ASD, Robotaxi, smart city transportation projects, these 3 segments, respectively? Would love to hear management's thoughts on that. Thank you.

You mentioned the competitors from ASD, smart transportation, and MaaS.

The smart city transportation project.

Yes, sure. Yes, in terms of ASD, I think that the main competitors come from the traditional Tier 1 companies. But like I mentioned, we have unique advantages, for example, HD Maps. Basically, the non-Chinese companies are not allowed to provide high-definition maps. And for smart transportation project, there are a number of more traditional smart transportation companies domestically. Their business model is more like selling the hardware and complete the project and go away with it. But our model is more like an operating model. We would continue to optimize the throughput, reduce the traffic jam and improve safety for the overall transportation system as well as enable autonomous driving and connecting road with vehicles to give them more guidance on the road information and help them to avoid any kind of congestion. So like I mentioned, there might be some other vendors that could compete in these areas. But when you really compare about effecting of this kind of project, we are so much ahead. In terms of MaaS, there are a number of start-ups that have a trial or projects in scattered cities, but it's in such an early stage. I think the adoption will be pretty much driven by the maturity of the technology. And things we are well ahead in terms of technology development, I think we are only limited by our capability of penetrating the cities and improving the technology itself. Not really limited by any competitors.

Thank you. Our last question for today comes from Tian Hou from T.H. Capital. Please ask your question.

Okay. I have 2 quick questions. One is related to your Apollo autonomous driving. A lot of questions have been asked. So my question is very short. So on one hand, all the modules you are selling through OEMs. On the other hand, you're going to start your own EV companies. So under what kind of circumstances some of your technology, you're going to just keep yourself instead of licensing to others? So that's number one. Number two, related to the AI call center solutions. So I'm very happy to see that operators and they start to purchase. And I wonder what is the potential for the AI call center solution. Who can be the potential customers and what industry this call center can apply to? So that's the 2 questions.

Yes. For the first question, Apollo is an open platform. All of our technologies are available to our customers. We will work with them to our best knowledge and best efforts. So anything that we use for ourselves is available to third party, too. It's just because we have a better understanding of the technology itself. We can better integrate software with hardware and use the feedback loop in a more timely fashion so that we can improve our technology faster. If you can compare this with the Windows and Surface, right? Surface serves as a benchmark for laptops. That's pretty much like Apollo for third party and for our own EV. On the AI solution for call centers, it actually can be used in a lot of industries. I mentioned the airline industry, the telecom industry, the financial services industry. Wherever there is some customer service need that our AI call center solution can be apply to, it's actually being used in a number of different industries. And we are going through Phase 1, Phase 2, Phase 3 so that the experience continued to improve and the knowledge base for that certain sector or for that certain company can be viewed. And the system will become smarter and smarter. End of Q&A

All right. Thank you. So ladies and gentlemen, with that, we conclude our conference for today. Thank you for participating. You may all disconnect.