Baidu, Inc. (BIDU) Q2 2022 Earnings Report, Transcript and Summary

Thank you for standing by, and welcome to the Baidu Second Quarter 2022 Earnings Conference Call. All participants are in a listen-only mode. There will be a presentation followed by a question-and-answer session. [Operator Instructions] I would now like to hand the conference over to Juan Lin, Head of Investor Relations for Baidu. Please go ahead.

Hello, everyone, and welcome to Baidu's second quarter 2022 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 CEO; Rong Luo, our CFO; Dou Shen, our EVP in charge of Baidu AI Cloud Group; and Zhenyu Li, our SVP in charge of Baidu Intelligent Driving. 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 provision 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. For detailed discussions of these risks and uncertainties, please refer to our latest annual report and other documents filed with the SEC and Hong Kong Exchange. Baidu does not undertake any obligation to update any forward-looking statements except as required under applicable law. Our earnings press release and this call include discussions of certain another non-GAAP financial measures. Our press release a reconciliation of the note non-GAAP measures to the analysis 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 be available on Baidu's IR website. I will now turn the call over to our CEO, Robin.

Hello, everyone. Baidu Core earned RMB23.2 billion in revenues and 5.1 billion in non-GAAP operating profit in the second quarter. The resurgence of COVID-19 led to an unfavorable macro environment in the quarter. However, Baidu AI Cloud and Intelligent Driving business maintained their rapid growth momentum as we aligned well with the government's repeated cost for technology innovation. Especially Q2 revenues from Baidu AI Cloud increased by 31% year-over-year outgrowing many of our peers, contributing 18% of Baidu Core's revenues in the quarter. That's up from up from 14% in the same period last year. Further, the operating margin for Baidu AI Cloud improved on both a year-over-year and quarter-over-quarter basis. Apollo Go completed 287,000 lines in Q2, increasing by almost 500% year-over-year. On July 20, Apollo Go's accumulative price reached 1 million. And in July, we unveiled Apollo RT6, the sixth generation Apollo robotaxi vehicle, bringing the cost of robotaxi to the price range of mass-market electric vehicles for the first time in China and globally. We aim to put a sizable amount of RT6 vehicles into operation in the year 2024. While Baidu Core ad revenue decreased by 10% over the year, in second quarter due to a challenging macro environment, our mobile ecosystem continued to generate strong cash flow as we focused on operational excellence and efficiency. Since early June, we have seen signs of recovery as the control measures were gradually lifted. For example, the year-over-year decline of our ad revenues narrowed in June and July. In particular, ad revenues from retail customers resumed year-over-year growth during the 618 Shopping Festival. Project implementation for Baidu AI Cloud has also been improving, thanks to the gradual removal of some major travel restrictions. While the recovery is likely to be at a gradual pace, and there are still be some uncertainties about the development of COVID-19, our long-term goals, strategies and capabilities for shareholder value creation remain intact. As a group, we have built a discipline of long-term thinking, which we believe is a strong asset in this fast-evolving innovative and highly competitive wins environment. Cost optimization and operational efficiencies are always top of mind for us. Our mobile ecosystem business should continue to generate strong cash flow to fuel and fund our investments in Baidu AI Cloud and Intelligent Driving over the long term. Now let's review the second quarter operational highlights. Revenues from Baidu AI Cloud grew by 31% year-over-year in the quarter, once again, offsetting our peers despite the macro challenges brought by COVID-19 resurgence. Our growth was driven by the strong performance in cloud business within enterprise business and the public sector. Not only, we managed to improve our operating margin both on a year-over-year and quarter-over-quarter basis in Q2, as we focused on healthier and more sustainable cloud revenue growth. We aim to achieve profitability as we scale up further. Baidu AI Cloud continues to benefit from the growth opportunity in China, digital and intelligent transformation. We believe this transformation is due at the early stage, and organizations are eager to leverage technologies to improve efficiency and productivity. Over the past years, we have made tremendous advances in AI technologies, made our AI capabilities widely available to enterprises and developers through our open platform, gained valuable for industry expertise, saw significant improvements in productivity in areas like manufacturing, transportation, energy and utilities and the public sector. Baidu AI Cloud's differentiating strengths are our AI capabilities and deep understanding of industry-specific pain points to help our customers to accelerate their transformation. In short, our competitiveness enables us to take full advantage of the tailwind offered by these new secular trends. Revenues from ACE Smart Transportation increased by about 40% in Q2. During the second quarter, we won several new projects for highway and in urban roads, both from our new and existing customers. By the end of second quarter, Baidu ACE Smart Transportation has been adopted by 51 cities up from 20 cities a year ago and 41 cities a quarter ago based on the contract amount over RMB20 million. Going forward, we will continue to focus on key industries and enable cross-industry replications for similar use cases by utilizing AI innovation. By doing so, we believe we will be able to reduce the operating losses and achieve profitability over time. In the meanwhile, we will continue to focus on quality and sustainable revenue growth and aim to continue outgrowing our peers. Moving to Intelligent Driving, Apollo Go continues to take a big step towards building up operations, strengthening its lead position as the world's fifth autonomous ride-hailing service provider. In Q2, Apollo Go provided 287,000 rides to the public, up from almost 500% year-over-year and 50% quarter-over-quarter. On July 20, 2022, Apollo Go accumulated rides reached 1 million. In the Yizhuang region of Beijing, according to our estimate, Apollo Go has obtained a 10% market share for ride-hailing services with pickup and drop-off both in Yizhuang. Apollo Go already becomes an important supplement for daily commuting app we launched the services less than two years ago. Now more than 100 Apollo Go vehicles are serving the Yizhuang resident on a daily basis, and each vehicle on average, completes more than 20 rides per day. In July, Apollo Go branched out to Hefei and Chengdu and is now operating in more than 10 cities charging fees in seven cities across China. We believe large-scale operation allows us to improve our level for autonomous driving technologies further. With data generated from everyday operation, we can identify problems that are not visible during the testing phase. The number of corona cases incurred during operations has far exceeded the ones found by vehicles and testing. In addition, Apollo Go has made some exciting progress since our last earnings call, marked by receiving the green light to provide fully driverless ride-hailing service systems and launching of the sixth-generation robotaxi vehicle, RT6. Earlier this month, we obtained the regulatory permits in Chongqing and Wuhan to provide fully driverless lighting services on open roads. We were also authorized to collect fees on those services. We are proud to be the first and only company in China now to offer fully driverless ride-hailing services to the public on open roads, entirely without human drivers in the car. Such achievements validate our superior Level 4 autonomous driving technology and endorse our strategy of utilizing large-scale operations to boost further technological advancements. On July 21 at Baidu World, we unrolled Apollo RT6, the sixth-generation Apollo global taxi vehicle with a targeted mass production cost of RMB250,000. This is significantly lower than the mass production cost of the fifth-generation vehicle and is in the price range of mass market electric vehicles. RT6 is designed to offer large-scale fully driverless ride-hailing services and launched at a time that both our technological and operational capabilities are ready. We believe that China's smart EV value chain has already been well established. Some of the auto parts and components for Level 4 autonomous driving solutions are already being produced in China at a much lower price than a few years ago. Moreover, RT6 is a passenger-centric vehicle, so we have removed some components, especially designed for drivers. For example, RT6 has a detachable steering wheel, which could unlock space for extra seat, luggages, desks and even some entertainment equipment in reaching the in-car experience for the passengers. Such achievements have set a strong foundation for Apollo Go to significantly reduce the two largest cost items for its operations, labor cost and the robotaxi vehicle cost. This is a critical step for Apollo Go to build a revolutionary business model and accelerate its expansion. Over the long term, we aim to make Apollo Go an alternative mobility option for millions of people, creating tremendous economic and social values. Baidu's Apollo auto solutions, including ASD and DuerOS for auto continues to gain popularity among automakers. As an increasing number of auto OEMs have publicly committed their brands to autonomous and connected vehicles, we're pleased to say that many have chosen to form alliances with Baidu to expedite their time line and plan. Recently, the total projected accumulated sales of our auto solutions have grown to RMB10.3 billion per our internal estimate. Moving to the mobile ecosystem, the user base of our mobile ecosystem continued to grow steadily. We once again delivered strong margins in the second quarter despite the negative impact on our absence caused by macro weakness. On the user side, Baidu app's MAU increased by 8% year-over-year to 628 million in June. Data search queries and the content is distributed through Baidu app grew double digits year-over-year in Q2. Our innovations and efforts in delivering a closed-loop experience has made Baidu app increasingly valuable to users. In addition to searching for information and knowledge, people come to Baidu app to look for ways to fulfill their needs. We have enabled them to order services, purchase products and interacted with industry experts and others without leaving Baidu app. Here are some metrics to share. 84% of our DAU in June were locking users, that's up from 76% a year ago. People have discovered that their user experience within Baidu app continues to improve once they are locked in. Quarterly GMV facilitated by Baidu search, [indiscernible] [LotSmart] grow by about 127% year-over-year. In particular, during the 618 Shopping Festival, GMV increased by about 260% year-over-year. In June, while 130,000 industry experts actively signed up for our instant replies future, up from about 100,000 in March. Total daily instant replies increased by about 190% year-over-year, up about 30% from March. Another highlight is our progress in short videos. Currently, about 85% of the content distributed within Baidu app feed are short video. During the quarter, we expanded the fully immersive video experience in the service home page to all the users. This feature is now available on both search and feed, helping by short video consumption on Baidu app. In Q2, short videos distributed through certain feed increased by double digits year-over-year. In addition, we are also using AI to expand our content portfolio. We give content creator AI tools, which they can use to produce video content more efficiently. AI generated content work particularly well for long and short content and breaking mils. While using our closed-loop experience, more and more users leading their footprint within our mobile ecosystem, increasing our capability to understand user needs and enabling strong conversion optimization at each step of their journey. Together with our continuous effort to optimize the advertising technologies, this has helped us to significantly improve the app conversion rate for our customers. Looking ahead, here are some key takeaways I'd like to give you. First, operationally, we have built strong new growth engines for Baidu Core, which we believe will boost Baidu Core's revenue growth in the coming years. Baidu AI Cloud uniqueness in AI capabilities continues to outgrow most of our peers. Apollo Go continues to solidify its position as the largest autonomous ride-hailing service provider in the world. Apollo Go has entered a new chapter with fully driverless ride-hailing onopen roads. While the launch of our purpose-built RT6 robotaxi vehicles for ride-hailing at really low cost, Apollo Go is set to accelerate its operation, further differentiating from our peers in both technology and operation. And the second takeaway financially is -- the mobile ecosystem will continue to generate strong cash flow to fund our investments in cloud and intelligent driving. On a separate note, we are also excited about iQIYI's continuous efforts in improving operational efficiency. iQIYI once again generated operating profit in Q2, the second consecutive quarter to report positive operating profit on both GAAP and non-GAAP financial measure. Furthermore, iQIYI generated positive net operating cash flow in the quarter for the first time. Before I turn the call to Rong, I'd like to highlight that we will be added to the Hang Seng Index from September 5. We believe with the inclusion of Hang Seng Index, one of the best known indices in Asia, the Hong Kong shares of Baidu will receive more front flow. Overall, I'm very proud of the team's strong execution in the quarter and remain excited about our future. With that, let me turn the call over to Rong to go through our financial highlights.

Thank you, Robin. Now let me walk you through the details of our second quarter financial results. Total revenue was RMB29.6 billion, decreasing 5% year-over-year. Revenue from Baidu Core was RMB23.2 billion, decreasing 4% year-over-year. Baidu Core, online marketing revenue was RMB17.1 billion, decreasing 10% year-over-year. Baidu Core non-online marketing revenue was RMB6.1 billion, up 22% year-over-year, driven by cloud and other AI power businesses. In Q2, Baidu AI Cloud increased by 31% year-over-year to RMB4.3 billion. Revenue from iQIYI was RMB6.7 billion, decreasing 13% one- year-over-year. Cost of revenues was RMB15.2 billion, decreasing 5% year-over-year, primarily due to a decrease in content cases, partially offset by the increase in personnel-related expenses and other courses related to new AI businesses. Operating expenses were RMB11.1 billion, decreasing 8% year-over-year, primarily due to an increase in channel spending and promotional marketing expenses. Operating income was RMB3.4 billion. Baidu Core operating income was RMB3.2 billion, and Baidu Core operating margin was 14%. Non-GAAP operating income was RMB5.5 billion. Non-GAAP Baidu Core operating income was RMB5.1 billion, and non-GAAP Baidu Core operating margin was 22%. Total other income net was RMB151 million compared to a total other loss of RMB2.4 billion last year, which included a fair value loss of RMB3.1 billion from long-term investments. In the second quarter of 2022, we recognized a favorable gain of RMB536 million, a significant portion of long-term investments including, but not limited to, investments in equity securities of public and private companies, private equity funds and digital assets is subject to quarterly fair value adjustments which may contribute to a net income volatility in future periods. Income tax expenses was RMB125 million, decreasing 99% year-over-year, primarily due to the reversal of certain tax expenses accrued for 2021 based on the 2021 tax return filed in the second quarter of 2022 and an increase in deduction on certain expenses that were considered not deductible in the second quarter of 2021. Net income attributable to Baidu was RMB3.6 billion and diluted earnings per ADS was RMB9.97. Net income attributable to Baidu Core was RMB3.7 billion. Non-GAAP net income attributable to Baidu was RMB5.5 billion. Non-GAAP diluted earnings per ADS were RMB15.79%. Non-GAAP net income attributable to Baidu Core was RMB5.4 billion and non-GAAP net margin for Baidu Core was 24%. Adjusted EBITDA was RMB7.1 billion, and adjusted EBITDA margin was 24%. Adjusted EBITDA for Baidu Core was RMB6.6 billion, and adjusted EBITDA margin for Baidu Core was 28%. As of June 30, 2022, cash, cash cadence restricted cash and shopping investments were RMB189.4 billion and cash, cash equivalents, restricted cash and short-term investments, excluding IT were RMB184.5 billion. Free cash flow was RMB5.5 billion and free cash flow, excluding iQIYI was RMB5.5 billion. Baidu Core had approximately 36,000 employees as of June 30, 2022. With that, operator, let's now open the call to questions.

[Operator Instructions] Your first question will come from Alex You of JPMorgan. Please go ahead.

Hi, management, good evening. Thank you for taking my question. I note you guys mentioned that you are seeing a sign of recovery in your advertising business since June. Could you management just share more color on the recovery in terms of advertising categories? Looking into the second half of 2022, how do you expect the advertising to develop? And in the medium to longer term, if we look beyond the near-term impact on COVID, how do you foresee the advertising business to develop in the medium to longer term?

Alex, this is Robin. As you know, Baidu Core ad revenues decreased about 10% year-over-year last quarter because of the challenging macro environment. In April and May, our ad revenues experienced a significant year-over-year decrease because COVID-19 impacted some of China's major cities, especially Shanghai and to a lesser extent, Beijing. In June, our ad revenue started to recover, when the situation gradually improved, and July was better than June. In terms of ad verticals, some of ad verticals got hit really hard in April and May. That includes the retail sector, travel, local services and health care. Ad spending from these verticals has been picking up quickly since June. Looking into the second half of the year, we're still facing macro uncertainties. It is hard for us to predict the divestment of COVID at this stage. For example, the resurgence in some of the hot travel destinations like Sanya, caused by the early ending of the summer travel peak season. Our search app is highly correlated with China's GDP growth and the performance of China's SMEs. So, we believe our ad revenue should recover when macro improves. We have also seen some opportunities for our mobile ecosystem around e-commerce and short videos. For example, integrating e-commerce and short videos with our search and feed, in fact, more and more short videos are available in our search results. This should help us to improve the user experience. I think in the mid- to long term, our -- in which the content in key verticals and our capabilities in facilitating transactions within Baidu App as well as our closed-loop experience for users, will help us to create more user insights for our customers. These insights will help us drive ad conversion and generate advertiser value, supporting the long-term growth of our online advertising business.

The next question comes from Kenneth Fong of Credit Suisse. Please go ahead.

Hi, good evening. Thank you for taking my questions. Congratulations on another quarter with a strong cloud revenue growth. Can you help us understand the underlying drivers that help Baidu AI Cloud to outperform your peers? What's your outlook for the cloud business for the second half of the year? And I'm wondering if you can also help us understand your strategy as well as time line for the cloud business to reach breakeven?

Thank you, Ken, for your question. This is Dou. I will address the revenue part and leave the second part to Rong. As Robin just mentioned, right, Baidu AI Cloud continues to grow pretty well in Q2 despite the challenging macro environment. And the revenue was up 31% year-over-year, and the AI Cloud now contributed to almost 20% of the total Baidu revenue. We believe what differentiates by Cloud from our peers is that we have been dedicated to integrate cloud computing with AI so that our AI capabilities can help our customers to improve efficiency. To be more specific, as a cloud vendor, we have strong AI capabilities. While on the other side, as an AI solution provider, we have strong cloud computing capabilities. So, also as we mentioned, right, so the enterprises in the public sector were the main growth driver for cloud computing for. So, we believe that in China, the traditional enterprises in the public sector are still at the early stage of digital and intelligent transformation. So clearly, this transformation means organization is no matter big or small, they want to use AI to improve their efficiency and their competitiveness, so which will further drive the demand for high-performance computing. So as a result, right, the combined revenues for -- from manufacturing, energy, utilities and the public sector almost doubled in Q2. In the smart transportation, we continue to leverage our leading AI technologies and the deep understanding of autonomous driving to grow our business. So, we see small transportation has already become one of the largest verticals for our Baidu AI Cloud. So with that, we are confident that revenues from enterprises in the public sector will continue to drive the growth of Baidu AI Cloud in the future. Also since early June, we have seen some gradual improvement in project implementation. But for sure, there are still some uncertainties around COVID-19 in the near term. But however, we believe the strong need and the trend for in China remain unchanged. And with that, we aim to continue growing faster than our peers.

Yes. Hi, Kenneth. So, I will answer the question about margins. As the AI Cloud continue to generate positive gross margins in the second quarter, same as we in the first quarter. On top of that, we managed to improve our operating margin, thanks to our strategy to focus on quality and sustainable revenue growth. So we have two parts in our AI Cloud, the first part is the personal cloud. The personal cloud contributes a small portion of the overall cloud revenue. It should continue to generate big operating profit, and its revenue is growing a little bit slower than overall cloud. And the second part which is the majority part is enterprise and public service cloud, which continue to generate positive gross profit in the second quarter. We improved operating margin growth on a year-over-year and quarter-over-quarter basis. This was because we practically focus on high-quality revenue growth and gradually will reduce the unhealthy projects. I think in the future, we will work very hard to narrow the loss margins. Within enterprise and public sector cloud, the ACE Smart Transportation, which just as mentioned by Dou just now already enjoys a very healthy gross margin because of the higher portion of the software and AI components. We think ACE's Smart transportations may achieve breakeven earlier than any other verticals in the coming quarters. and because we have focused on to provide mortalized solutions and products to for the key use cases, for example, urban roads, highways and power. Secondly, I think we are also working very hard to replicate the similar scenarios in use cases across different industries. Cross-industry replication required AI technology to find our connections across many different scenarios. We believe that our AI technology, such as the large models will help us to achieve that goal. Overall, we will continue to focus on quality and sustainable revenue growth to improve margins and reach profitability in the next few years.

The next question comes from Lincoln Kong of Goldman Sachs. Please go ahead.

My question is about robotaxi. So, we're quite impressed that RT6 is targeting a cost of RMB250,000 per car and that's significantly lower than the generation of the taxi models. So can management help us understand, what's the main driver for this cost reduction? And what's our outlook for the future cost trend? And how should we think about robotaxi's economic mechanism on this cost? And any time line, I think when to expect major commercial contribution from the robotaxi business?

Lincoln, this is Robin. I think there are three drivers for the cost reduction. First, it's our deep knowledge in the autonomous driving technology and operations. Over the past nine years, Baidu has viewed strong Level 4 technology and gained industrial industry know-how on the auto industry and how to operate L4 ride-hailing services. Also, we have become the largest autonomous ride-hailing service provider in the world. So, we fully understand the autonomous ride-hailing market and understand our passengers' needs. We know where we can simply find our in-car features and where we should allocate more resources. For example, RT6 is a passenger-centric vehicle not a driver-centric vehicle. So on one hand, we have reduced some in-car features designed for the driver to reduce the cost like I mentioned during the prepared remarks. RT6 can remove the steering well and exclude some other accessories such as wide screen, wide display screen for drivers. On the other hand, we put more efforts on safety performance and the passenger experience to make RT6 passenger centric. For example, we have used safety redundancy in seven parts of RT6, setting an industry standard. This includes redundancy for computing unit, sensors, steering systems, braking systems, communication system, power supply and auto architecture. Also RT6 is designed for more in-car space and better in-car experience. I also want to highlight that about 60% of our RT6 BOMs for intelligent features like L4 and automating capabilities. The second driver is scale. We plan to leverage third-party automakers to produce RT6 for us. So RT6 will not retrofit on a passenger vehicle. This is a key differentiator of RT6. We will pay for the BOM of RT6 instead of the retail price like what we did in the past. Once Baidu's able to move away from retrofitting vehicles to use third-party automakers to produce RT6, that's because we have the capability to offer large scale will protect the operation, supported by our leading autonomous driving technology and operational capabilities. And I think the third driver is the divestment of the value chain for China's intelligent EV industry. RT6 was launched at the time when China's intelligent EV industry is well built up with many new EV companies and auto part producers in the market. In particular, many autonomous driving-related auto parts are now being produced locally. So intelligent EVs are becoming more affordable than before and for the automotive-related auto parts, its cost is also much lower than before. For example, LiDAR. LiDAR is the most important component for autonomous driving solutions. It was charged hundreds of thousands of U.S. dollars in the beginning, but now it only costs like $1,000. So in the future, we believe the BOM for our robotaxi will continue to decrease because China's intelligent EV industry will continue to develop, and Apollo Go will continue to scale. Our approach is a gradual approach. We will continue to roll out regions in different cities. So therefore, scale up our robotaxi service, as we gain scale, as we improve continuously on technology, I think that the cost for operating such service will continue to drop. And we certainly have the first-mover advantage and the barrier to entry for this will become higher and higher.

The next question comes from Alicia Yap of Citigroup. Please go ahead.

Thank you. Good evening, management. Thank you for taking my question. I have a follow-up question on Apollo Go. So congratulations on some of the solid achievements recently. Can management provide some update on regulatory environment for autonomous driving and robotaxi in China? And how Baidu will capture this trend? And then, can management also provide updates on Baidu's goal and strategy for this fully driverless operation?

Thank you. This is Zhen. Just as Robin said, Ergo is already the largest automotive service provider in world. In quarter two, Ergo has completed 287,000 rides and were clearly ridesharing surpassed $1 million. This number is much higher than option -- our strategy is to leverage a large operation to improve level technology and growing faster and larger. This is very important in China because of the road and traffic conditions are very communicated here, and the data generated from larger steel operations will help us to improve our lower hall technology and open the problem. Today, our goal is already available in more than 10 cities including all the Tier 1 cities in China. Our operating areas are many places, including visions, office buildings and the hospitals cities. On the upside, we're also expanding of off-rate industry. With the launch of RT6 as a target market production cost or RMB250,000, we will be able to purchase grow off-rate at a faster pace. In the future, we will continue to improve our Level 4 technology, and grow our operation. Also, we will try to push and make our influence on China's regulation or autonomous drive. For technology, we started up to offer autonomous ride-hailing services on open growth in Wuhan on May 10. In less than three months, we will allow us to provide fully driverless ride-hailing services on the open road. The great progress is for our leading technologies and capability.

For operation in the Yizhuang region in as Robin mentioned, we have already acquired about 10% back share for ride-hailing services in the region. [indiscernible] our average each of activity will commit more than 25%. As a pioneer in the industry, Baidu worked closely with the regulator to set up regulations and industry standards. Recently, the Chinese Ministry for Transport issued a draft of deadline for the driving record transportation safety services. The guidelines aim to push the commercial operations for autonomous ride-hailing services in China. This is the first regulation on autonomous ride-hailing services at a central government level globally. It matched efforts in promoting high-level autonomous driving. We are proud to be the first and only company in China to offer fully driverless ride-hailing. In early August, Baidu was allowed to provide fully driverless ride-hailing services in Hefei, in Chongqing and Wuhan who owned China's major cities in the Middle East. In the future, we will continue to scale up our region and try to push fully driverless ride-hailing to [markets].

The next question comes from Eddie Leung of Bank of America Merrill Lynch. Please go ahead.

Thank you for taking my question. I have a question about your autonomous driving solution for OEMs. So you mentioned that the contract value after the end quarter was just over $10 billion. It doesn't seem to be increased -- increasing much from the last quarterly result, could you tell a little bit about the reason behind that? And then could you also remind us some of the automakers that have been working with you guys in using your autonomous driving solutions?

Ed, let me answer your questions. I think as Robin mentioned in the remarks, yes, the total projected cumulative sales of our ad solutions has been RMB10.3 billion as per our own estimates. I think later talk about the numbers we need to keep in mind. The smart EV is a trend that is irreversible. For cars, we believe that over the long term, electrification is just a part of the play and intelligence in the game changer. So in the future, EV without intelligent driving capabilities will not be competitive. And many automakers like some leading domestic or even international automakers, they are interested in using our solutions for their car models to capture this kind of new market opportunity. Based on our conversations with them, we understand the automakers like our AI capabilities, our insights of the autonomous driving and our strong brand. And we are talking with many of the automakers, both locally and globally, and some are very large names in the industry. I want to emphasize that each automaker has its own schedule for setting up the countries with the suppliers. This could result in a different pace of our backlog. This is absolutely normal in auto industry. As of today, we have partnered with many automakers, including some leading like [indiscernible] car model and so on. And in addition to our in-car infotainment solutions, more and more customers now are adopting our Apollo self-driving products, including the ANP products Apollo Navigation Pilot for highways or urban roads. The automatic AVP products and HD Maps products, I think in the future, we will continue to expand our partnership to work with more automakers. And as more and more models equate our auto solutions are set to be launched into the market in the coming three years, the auto solution should start generating many revenues.

The next question comes from Gary Yu of Morgan Stanley. Please go ahead.

Thank you for the opportunity to ask question. Congratulations on the solid results. My question is about margin and cost control and optimization. We saw that you achieved another strong quarter results with margin better than expected. Could you please elaborate the efforts to talk for such encouraging results? And what is the margin outlook for the second half of the year?

Yes, thanks so much, Gary, and I'll answer this question. I think despite of the very challenging macro environment in Q2, Baidu cost, non-GAAP operating margin is 22%, which is up five percentage points quarter-over-quarter. Over the past few months, I think we have made a lot of efforts to improve our operational efficiencies. And for example, the Baidu cost, SG&A was down by 12% year-over-year in the second quarter. I think our businesses are at a different stage of development. So we have different strategies of investment for each business. For the mobile ecosystem, which is relatively mature business, our goal is to keep its profits and margins solid so that our remote ecosystem can continue to generate a very strong cash flow to support our investments in the new AI initiatives. I think for our new AI initiatives, we will be focused on AI Cloud and intelligent driving. Both of them grew very fast in the past few quarters, which you can see there. So, we will continue to make necessary investments as needed. [indiscernible] AI now will remain focused on here and sustainable revenue growth, and we aim to continue improving our operating margins in the future.

The next question comes from James Lee of Mizuho. Please go ahead.

My question is regarding Jidum and congratulations, by the way, on your launch in June. And at the same time, you guys also launched RT6. I'm just curious about maybe the positioning of Jidu and RT6. Can you guys talk about the customer segmentation strategy of the two autonomous driving vehicles? Just curious, how much overlap in demand between the two?

Yes. This is Robin. The position for Jidu and RT6 are very different. Before I talk about Jidu, let me first talk about our strategy for intelligent driving. Our goal is to provide large-scale fully driverless ride-hailing services on open road. This is powered by our L4 autonomous driving technology. And at the same time, we use some of our L4 capabilities to put them into assisted driving features to support automakers' intelligent driving needs. And Jidu is just a perfect example of such automakers. So, if you look at our business portfolio for intelligent driving, we have RT6. RT6 represents Baidu's best autonomous driving technology. RT6 is for large-scale fully driverless ride-hailing services on open road. That is not a consumer-facing vehicle. I mean it's not selling to the consumer. Jidu's Robo-01 is a consumer facing product. We put on our Level 4 technology into Level 3 or Level 2+, if you will. Jidu Robo-01 will be equipped with AMP 3.0, our most advanced autonomous driving solution. AMP 3.0 can handle end-to-end intelligent driving and parking, urban roads and highways. Also, our smart cabin will be installed in the car. It allows Jidu to provide voice interaction, insights and outside the car. We believe these new features will help us to expand our business. For the timeline, we could see everything is well on track. We could see the first mass-produced Robo-01 finished very soon. The time line demonstrates Jidu's capability from design to mass production. Jidu plans to take orders for Robo-01 later this year and deliver the vehicles starting 2023. And Jidu has the second model. Jidu plans to launch at the end of this year, then start taking orders in 2023 and deliver in 2024 for our second model.

Our last question will come from Wei Xiong of UBS. Please go ahead.

First, I'm wondering if management can share some updates on your capital allocation. Could management talk a bit more on the implementation of your share buyback program? And how do you balance the investments into new businesses versus doing more buybacks or dividend? And second, as we recently saw the development between the U.S. and Chinese regulators regarding the audit inspections of ADRs, could management also share your thoughts on this progress? And given some of your peers have started the process to convert into dual primary listing, I was wondering would Baidu also consider such a U.S. Hong Kong due primary listing status?

Yes. Thank you for your questions. Let me to your questions. I think as of today, we have returned about $2.9 billion to shareholders under the 2020 share repurchase program. The program authorized $4.5 billion cap which means we still have around $1.6 billion left in the program. Baidu will continue to buy back our shares from open markets since, we quite strongly believing our long-term developments. If you're looking to our balance sheet, as of Q2, our cash equivalents, restricted cash and short-term investments were $28.3 billion, and we also have the other $11.1 billion in our loan term investments. And if you look into our business today, some of that was mentioned in our previous remarks. I think for mobile system, it continues to generate wide healthy profits and margins. For Baidu AI Cloud, the revenue has been growing very fast, faster than our peers in this industry. For Apollo Go, it's already the largest outside heading service provided in the world. So everything is pretty much on track. So we believe that we are on the right direction, and we will continue to invest in other steps. And about the second question about the delisting and the primary in Hong Kong, I think the most recent statement of the protocol site between the PCOB and the Chinese authorities marks the very important first step towards opening access for the PCOB to investigate complete to the rest public accounting fund in China, I think which is a very important milestone for both countries and China-based issue with the ADR trading in the U.S. Also, we can feel that the two governments are practically seeking solutions. For Baidu, we already have the Hong Kong secondary listing, and we have some major shareholders who converted their Baidu ADR into ordinary shares in Hong Kong. Also, we are also very glad to be included in the Hang Seng Index. It will take effect on September 5. We believe with the inclusion to the Hang Seng Index, our Hong Kong shares will receive more front flow. For the primary listing, as far as what we understand, the process of switching from the secondary to primary listing in Hong Kong should be straightforward. We don't see any kind of hurdles for us to achieve that. But I think what's more important for us is to remain focused on business fundamentals and drive long-term shareholder values. Also, Baidu will continue to comply with applicable loss in both China and the United States and try to maintain listing both in NASDAQ and Hong Kong exchange. Thank you so much.

That will conclude today's question-and-answer session. That does conclude our conference for today. We thank you for participating, and you may now disconnect.