Q4 2023 Baidu Inc Earnings Call
Operator: Hello, and thank you for standing by for Baidu's fourth quarter and fiscal year 2023 Earnings Conference Call. At this time, all participants are in listen-only mode.
Hello, and thank you for standing by for Baidu is fourth quarter and fiscal year two.
2023 earnings conference call.
At this time, all participants are in listen only mode.
Operator: 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, Chowdhuryn.
After managements 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 Joe.
Lynn.
Chowdhuryn: Director of Investor Relations. Hello, everyone, and welcome to Baidu's fourth quarter and fiscal year 2023 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.
Joe Lynn: <unk> director of Investor Relations.
Joe Lynn: Hello, everyone and welcome to by the fourth quarter and fiscal year 2023 earnings Conference call earnings release was distributed earlier today and you can find a copy on our website as well as a mix of our services on the call today.
Chowdhuryn: On the call today, we have Robin Li, our co-founder and CEO, Rong Luo, our CFO, and Dou Shen, our EVP in charge of Baidu AI Cloud Group. 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. Cloud Security Mitigation Reform Act of 1995.
Joe Lynn: Oh the coupon there.
Joe Lynn: Our CFO and ocean, our EVP in charge of Baidu AI call ACG I'll call. Our prepared remarks, we will hold a training session. Please note that the discussion today will contain forward looking statements made under the safe Harbor provision of the U S Securities Litigation Reform Act with Nike.
Speaker Change: Thank you bye.
Chowdhuryn: Forward-looking statements are subject to risks and uncertainties that may cause actual results to differ significantly 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 government. Baidu does not undertake any obligation to update any forward-looking statements except as required under applicable law.
Speaker Change: Forward looking statements are subject to risks and uncertainties that may cause actual results to differ materially from our current expectations.
Speaker Change: For detailed discussions of these.
Speaker Change: It's an uncertain.
Speaker Change: Please refer to our latest annual reports and other documents filed with I think the Hong Kong stock exchange Baidu does not undertake any obligation to update any forward looking statements, except as required under applicable law.
Chowdhuryn: Our earnings press release and this call include discussion of certain unaudited non-GAAP financial measures. Our press release contains a recapitulation of the unaudited non-GAAP measures to the unaudited most directly comparable GAAP measures, and it's 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. Hi everyone.
Speaker Change: Our earnings press release, and this call includes discussions of certain unaudited non-GAAP financial measures.
Speaker Change: Press release contains a reconciliation of the unaudited non-GAAP measures to the unaudited most directly comparable GAAP measures and it's available on our IR website at IR <unk> Com as a reminder, these calls this is being recorded in addition, a webcast of this conference call will be available and by this I always talk.
Speaker Change: I will now turn the call over to our CEO Robin.
Robin: Okay and when we can.
Robin Li: We concluded 2023 with a solid performance. Baidu Core's total revenue for the fiscal year increased by 8% year-over-year, and the non-GAAP operating margin expanded from 22% to 24%. This improvement highlighted the resilience of our orbit, which serves as a strong foundation for our ventures in earning and earning bonds. 2023 marked a year of Gen-AI and foundational, a massive technological shift that will re-engineer this process and bring about a real renaissance in every sector of the economy. Baidu is well prepared to benefit greatly from this groundbreaking economic revolution. We progressed from talking about the opportunities of Gen AI and foundation models to actively applying early and early bots at scale. As the frontrunner in AI, Baidu proudly became the first public company globally to launch a GPT model, with our EB 4.0 standing tall as the most powerful foundation model in China. Ernie continues to gain market recognition, as evidenced by Ernie API calls from multiple well-known companies. Notably, Samsung uses the earnings API on its Galaxy S24 5G series; its owner uses the Ernie API. It's Magic 8.0.
Robin: To date 2023.
Robin: Tom.
Robin: And of course total revenue for the fiscal year increased about 8% year over year and the non-GAAP operating margin expanded from 22% to 24%.
Robin: This improvement.
Robin: Great.
Robin: Our core business.
Robin: Which serves as a strong foundation for our ventures.
Robin: And we're just bought.
Robin: 2000, Twenty's way, Mark a year off Jim Jong Foundation model.
Robin: A massive technology shifts.
Robin: We engineer this process and bring about Oh goodness.
Speaker Change: I understand.
Speaker Change: Each sector all state economy.
Speaker Change: At least well prepared to benefit greatly from this groundbreaking economic revolution.
Speaker Change: We progress from talking about the opportunities of gender Guy and foundation models too.
Speaker Change: Hi, Arnie and Bob at scale.
Speaker Change: As to Franco on their cars and AI Baidu.
Speaker Change: Baidu property became the first public company globally to launch a GPP model.
Speaker Change: R E B 410 standing call.
Speaker Change: The most powerful foundation model in China.
Speaker Change: Or do you continue to gain market recognition.
Speaker Change: As evidenced by earning API calls from multiple well known companies.
Speaker Change: Notably Samsung used to earning API.
Speaker Change: Okay, let's see at 24 five Tcf.
Speaker Change: On her.
Speaker Change: Our new API bad.
Speaker Change: Magic eight one theater.
Robin Li: And Autohome uses Learning API to power its multiple AIGC apps. The partnerships further solidify our leadership in Genium. Furthermore, we continue improving the efficiency of earnings. For example, the inference cost of EB3.5 is only about 1% of the March 2023 version.
Speaker Change: And also home using learning HCI policy watchful eye Juicy.
The partnership's workers, maybe fine art leadership in Shanghai.
Furthermore, we continue improving the efficiency of course.
Speaker Change: Example, the interest cost.
Speaker Change: 0.5 is only about 1% off the March 'twenty two 'twenty three version.
Robin Li: With lower inference costs, Ernie Well becomes increasingly accessible to users and enterprises. We are able to do that primarily because of our unique four-layer AI architecture and our strong ability in end-to-end optimization. Additionally, we are offering LLMs in smaller sizes. These small models can balance between performance and efficiency.
Speaker Change: With lower interest costs or any well become increasingly accessible to users and enterprises.
Speaker Change: We are able to do that primarily because of our unique four New York, Yeah architecture, and our strong ability and end to end optimization.
Speaker Change: Additionally, we are offering L. L M in smaller sizes.
Speaker Change: This small model imbalanced.
Speaker Change: Orman and efficiency.
Robin Li: Better Serving Wearer Customer Needs. Since Q2 2023, we have actively utilized Ernie to revolutionize our products and services. Creating AI Native Experiences. We believe real applications are essential to unleashing the full business potential of earning and revenue. Recently, we began to generate incremental revenues early and early. In the fourth quarter, we earned several hundred million RMB, primarily from ad technology improvement and helping enterprises build their own models. I'll provide a more detailed explanation in the bins review section.
Speaker Change: Disturbing wherein customer need.
Speaker Change: <unk> Q2, 2020 suite, we have actively utilized arnie to revolutionize our products and services.
Speaker Change: Creating AI era.
Speaker Change: Yeah.
Speaker Change: We believe real applications are essential to unleashing the full business you saw off warranty and already bought.
Speaker Change: Recently, we began to generate incremental revenues from Arnie and earn evolved.
Speaker Change: In the fourth quarter, we learned several hundred million RMB, primarily from AD technology improvement and helping enterprises build their own models.
Speaker Change: I will provide a more detailed explanation in the business review section.
Robin Li: Looking into 2024, we believe this incremental revenue will multiply to several billion RMB, primarily from advertising and the AI cloud business. While we are beginning to commercialize chain AI and foundation models, we see enormous monetization potential in this groundbreaking technology. We envision earning as the future foundation system, serving as the foundation for millions of AI native applications developed by third parties and by the. This paradigm will enable us to create an ecosystem around earning, which opens up various revenue sources. Now, the key highlights of Baidu's AI native application. Equipped with an AI copilot, Baidu Wenku has become a platform for users to create content in a wide range of formats, helping them express their ideas.
Speaker Change: Looking into 'twenty 'twenty four we believe this incremental revenue well multiply two several billion RMB.
Speaker Change: Primarily from advertising and AI cloud.
Speaker Change: Well, we are beginning to commercialize chimney I and foundation models, we see enormous monetization potential in this ground breaking technology.
Speaker Change: We envision the future Foundation system.
Speaker Change: As the foundation for millions of lithium applications.
Speaker Change: Thanks Third party and Baidu.
Speaker Change: This paradigm well enable us to create an ecosystem around which opens up with revenue sources.
Speaker Change: Mhm.
Speaker Change: The key highlights in Baidu AI application.
Speaker Change: Equipped with AI co pilot Baidu, one cool has become a platform for users to create content in a wide range of formats.
Speaker Change: Helping them express their ideas.
Robin Li: In addition to summarizing, creating, and expanding content, users particularly appreciate Baidu Wenku's features for assisting them in automatically turning their inspiration into PowerPoint. We have been consistently improving and enriching its AI features. For example, we recently introduced a new one, such as Mind Maps for organizing information. We are reconstructing Baidu search using earn. Last quarter, we talked about how we used Ernie to generate search results, improving the one-shot experience. In Q4, Baidu search introduced newsfeed-like information alongside the generated search results at scale to offer users useful and relevant content to explore. In Q4, we also began facilitating certain search queries with multi-round conversations to delve deeper into understanding users' intentions.
Speaker Change: In addition to summarizing.
Speaker Change: Aiding and expanding content users, particularly appreciate that Baidu language features for assisting them in automatically turning their inspiration into Powerpoint.
Speaker Change: We have been consistently improving and in reaching its AI features.
For example, we recently introduced a new one such as mine maps for organizing information.
Speaker Change: We are reconstructing baidu search using burden.
Speaker Change: Last quarter, we talked about how we used to Ernie to generate six results improving the wingstop experience.
Speaker Change: In Q4, Baidu search introduced a news feed like information alongside that generated the secs resolved at scale to offer users useful and relevant content to explore.
Speaker Change: In Q4, we also began facilitating certain search queries.
Speaker Change: Multi round the conversations to delve deeper into understanding users' intention.
Robin Li: Overall, GenAI-enabled search complements traditional search, expanding the time of Baidu search by serving a wider range of information needs and answering content creation related questions. We are actively encouraging SMEs in service industries to build AI chatbots as new landing pages for Baidu search. AI chatbots enable them to better serve potential buyers directed to them by us, driving transactions on our platform. At the same time, our AI chatbots can help merchants increase productivity. Currently, about 4,000 merchants are using our AI chatbot, particularly in healthcare, education, travel, legal, B2B, and car, where users often interact with merchants to learn about their offerings before making purchasing decisions. We believe the GenAI-enabled search will help increase user intent on Baidu, reinforcing our position as the primary entry point for users seeking information and services.
Speaker Change: Overall general AI enabled search complement traditional search expanding the Tam of Baidu search by serving a wider range of information and answering content creation related question.
Speaker Change: We are actively encouraging S N knees in.
Speaker Change: Service industries to build AI chat bot.
Speaker Change: New lending pages for Baidu search.
Speaker Change: Yeah chat box enable them to better serve potential buyers.
Speaker Change: Directed to them by us.
Speaker Change: Private transactions.
Speaker Change: <unk> platform.
Speaker Change: At the same time, our chatbot can help merchants increased productivity.
Speaker Change: Currently about 4000 merchants are using our AI chatbot.
Speaker Change: Particularly in health care education travel legal b to B, and auto where users often interact with merchants to learn about their offering before making purchasing decisions.
Speaker Change: We believe that Gen II enabled search well hub increased user intention on baidu reinforcing our position as the primary entry point for users seeking information and services.
Robin Li: We are also expanding the reach of earnings by enabling enterprises to easily develop AI native applications on earnings. In December, we introduced AppBuilder, a comprehensive set of tools on our public cloud to support enterprises to swiftly build AI native applications from the ground up. These tools enable enterprises to incorporate capabilities such as organizing on-ramp messages for prioritization, creating drafts, summarizing meeting notes, and facilitating joint AI-enabled data analysis. Additionally, we have included several ready-made templates, such as AI Agent, to enable enterprises to promptly create applications for validating their ideas. Such progress is in large part based on our own pioneering experience in developing AI native applications. In addition to the tools for app development, enterprises partner with us because we have the most powerful and most efficient foundation model and the most cost-effective cloud platform for running. In December, about 26,000 enterprises were actively using Ernie through API on a monthly basis, increasing 150% quarter over quarter. And Ernie is now handling more than 50 million queries every day. That's a 190% increase over the quarter.
Speaker Change: We are also expanding the reach of Ernie by enabling enterprises to easily about AI native application online.
Speaker Change: Yeah.
In December we introduced as people get in.
Speaker Change: A comprehensive set of tools on our public cloud to support enterprises to swiftly.
Speaker Change: [laughter], Yeah Ah patients from the ground up.
Speaker Change: These tools enable enterprises to incorporate capabilities such as organizing unread messages for prioritize right taxation are creating jobs summarizing meeting nodes and facilitating gen AI enabled data analysis.
Speaker Change: Additionally, we have included that were already made templates such as AI agent to enable enterprises to promptly equate application for validating their ideas.
Speaker Change: Such progress is in large part based on our own pioneering experience in developing native applications.
Speaker Change: In addition to the tools for the bathroom and enterprise card members that because we have the most powerful and most efficient foundation models.
Speaker Change: And the most cost effective cloud platform for it.
Speaker Change: Running out.
Speaker Change: Yeah.
Speaker Change: In December about 26000 enterprises are actively using or any school API on a monthly basis, increasing a 150% quarter over quarter.
Speaker Change: And Ernie is now handling more than 50 million inquiries every day.
Speaker Change: 190% quarter over quarter.
Robin Li: It's a significant rise in third-party cost. As the number of AI native applications hosted on our public cloud continues to rise, we are poised to accelerate the enhancement of our model performance by incorporating feedback from our users and customers. This will further widen the gap between early and our domestic peers. Moreover, more and more AI native applications...
Speaker Change: There is a significant rise in third party column.
Speaker Change: As the number of yeah, I need your lapping applications hosted the our public cloud continue to rise we are poised to accelerate the enhancement of our model performance.
By incorporating feedback from our users and customers.
Speaker Change: This will further widen the gap between Rd, and our domestic peers.
Speaker Change: Moreover, as more and more AI native application.
And our next question today will come from Alex Yao of J P. Morgan. Please go ahead.
Hi, Alex Your line is open are you muted.
Yes, sorry, I was on mute.
Good evening management and thank you for taking my questions I have a couple of questions on cost structure and the margin trends.
First of all how much more room do we see.
The cost cutting and the optimization, how should we look at AI related investments in the past are you guys discussed that there will be a lag between chip investments and AI revenue contribution how should we look at the modern trend into 2024, if you intend to.
Jeremy.
Okay.
Hi, Alex.
So multiple questions of a service.
I think alongside of these investments.
AI businesses.
<unk> is pretty sure there are still room, we can manage the cost and the expenses in our <unk>.
Legacy businesses.
As we look into that change tend to fall, we will continue to focus on our core businesses.
Also we will be willing.
To reduce other resource allocations to the non strategy businesses.
Additionally, we are continuously enhancing our overall organization efficiencies by removing layers simplifying the executions and flattening the organizational structures.
So heading into this year, we are very committed.
To the ongoing optimization of our operations and ensuring we have the more productive HR team.
With all these measures and plan and to key Baidu cause earnings solid.
With our mobile ecosystem continued to deliver the wonderful margins and generated strong cash flows while AI cloud means continuing this profitability.
We have managed to maintain a solid operation.
Margins, despite our investments if you still remember firmly due to straightening owned SRT to them last year.
Generally the AI and last engine models. This investment is primarily reflected in our capex.
Mainly.
<unk> Tula purchasing chips and servers.
Modern trainings in fragrances.
As you know the Capex will be depreciated over several years. So despite offer a 68% year over year increase in baidu apps cost capex into disarray.
Our non-GAAP operating profit margins still saw a two percentage of growth on a year over year basis.
Looking forward.
In the process of developing our new AI businesses.
It's both.
Million.
However, as these investments are not.
Got you excited too significantly.
Models are process.
During the early stage of the market development. We're also will now.
Already prioritize margins for our aircraft business and.
US we believe that in the long run this business is expected to your much better margins.
Additional fee.
There may be some promotional activities for our native to see products.
We will carefully manage and closely monitor our two.
Balanced with investments in growth.
We're happy to see that our effort in investing in new initiatives have begun to year early results.
As Robin has mentioned just now it's had incremental revenue our January <unk> from the.
From the tax improvement have reached several media in Hangzhou and in Q4, and the incremental AI cloud revenue generated from the <unk>.
And the financial models also comparable to four 8% of total aircraft.
I think all of these.
Promising on top of that that shipments have strengthened our confidence in our strategy. So going forward, we will remain steadfast in our commitment to development.
To the development of the <unk>.
And last night.
Thank you Alex.
Yeah.
Our next question today will come from Gary you of.
Morgan Stanley. Please go ahead.
Hi, Good evening and thank you for taking my question.
I have a question regarding AI contribution.
For AI related revenue contribution is there a way you can quantify.
Proof that to Ed revenues Baidu will be generated is purely incremental contribution from AIG.
And not cannibalizing.
From your existing search business.
AI is purely incremental.
Should we be expecting faster than average growth in <unk>.
<unk> AI, how should we think about the core search growth rate.
Thank you.
Hi, Gary this is robin.
We are the largest search engine in China, we have close to 700 million monthly active users and we have established very robust a brand presence.
The Chinese Internet and mobile users.
They rely on us for a comprehensive and reliable information.
So we have a strong and stable base.
Revenue and profit.
Okay.
But we are also theyre sensitive to macro because of our advertising business.
Very.
Very broad coverage of different verticals.
And I mentioned earlier.
Steve uncertainties with the macro environment.
But generally I and.
Unlocking new opportunities.
Both at the monetization front and the user engagement trumped.
I think it's easier to quantify the incremental revenue on the monetization front.
I mentioned earlier that to January I have already helped increased advertising E CPM.
And our.
Upgraded monetization system.
Has allowed us to improve targeting capabilities.
Thereby generating and presenting more relevant ads.
We learn a several hundred million dollars.
In Q4 from.
The kind of initiatives.
And the incremental revenue will grow two 7 billion.
This year.
It's harder to quantify the user engagement part.
Okay.
Helping us improve user experience.
We are seeing a neutral outcome in search.
And loan pool.
And we will continue.
Introducing new features going forward.
This initiative will help us improve the user mind share and time spend over time and.
Bring are even <unk>.
Larger potential so I think could be.
Purely incremental revenue will come from both the monetization side and the user engagement.
Thank you.
Our next question today will come from Lincoln Kang of Goldman Sachs. Please go ahead.
Hi, Thank you management for taking my question. So my question is regarding the comp expense.
How should we look at the incremental revenue growth that's true, but generally I. So what is the perfect mix for January.
And what exactly that offerings.
The primary growth drivers here and then when we're looking at the 2024.
How should we expect the overall kind.
Our revenue growth this year and there will be the.
Modern trend this year.
<unk>.
Okay.
Thank you for the question is the Doe.
As Robin just mentioned rather the total revenue from.
From the <unk> into the foundation model related businesses, including both internal and external revenue.
Alrighty reached 656 million RMB in Q4, and this number should grow to several billion RMB for the full year 2024.
We have seen increasing interest from Andrew presses in using genii and L. L N.
To develop new applications and features.
To achieve this goal.
We are actively building models to power their products and solutions. So.
So this is how we generate the majority of the revenue from external customers.
Meanwhile, we were seeing a significant increase in our model influencing and revenue from external customers.
So the revenue from our inferencing is still small.
We believe that over the long term it will become significant and sustainable revenue driver.
We think revenue generated by our internal customers is also quite important.
A significant portion of <unk> revenue.
Model Inferencing.
So baidu is the first company to use <unk> and L. L M to reconstruct all of the businesses and products.
So this is the number of products and features powered by <unk> and <unk> continue to.
Increase.
The API calls.
From internal customers.
Had been increasing rapidly and have reached a significant magnitude.
Such development has proven that Ernie and Ernie box can well enhance the productivity and efficiency.
In the real world applications.
And we also believe a growing number if external customers will use <unk> to develop their own apps.
And drive our external revenue in the future.
So regarding your question about our product offerings. So we have the most powerful AI infrastructure.
Tier four model cleaning and inferencing in China.
This infrastructure helps our customers build and run models.
Cost effectively.
Additionally, as Robin mentioned earlier, our months offers their risk models and a full set of tubes and in terms of the model builder and App builder.
For mono building and application development.
So in addition to that we have developed our own native solutions, such as the GPI that is the <unk>.
Generative business intelligence.
So those applications are helping.
Andrew process increased productivity and efficiency.
In addition to the incremental opportunity relates to AI Genie AI and foundation models are also bring.
New opportunities to our legacy cloud businesses.
So firstly, we continue to win customers and projects for now.
CPU centric.
Good.
Because we are being highly recognized for our AI infrastructure and CPU centric cloud offerings, especially in the.
The internet intent and education sectors.
Secondly, <unk> AD funded models have allowed us to build solutions for our customers and more efficiently than before at facilitating digital and intelligent transformations.
Traditional industries.
So both of these two.
Factors or driving growth.
For our cloud revenue and overall, we should see cloud revenue growth accelerate in 2024 over last year.
Additionally, we are pretty confidence in maintaining profitability for our AI cloud.
For enterprise cloud, we should be able to consistently improve.
Gross margins.
For the legacy and cloud businesses.
As for <unk> and <unk> businesses. The market is still at a very early stage.
Element.
So we should hold pretty dynamic pricing strategy to quickly educated the market and expand our penetration into more enterprise customers.
So we believe over the long term.
The new business should have a higher normalized margins.
Then the traditional cloud businesses.
Thank you.
Okay.
Our next question today will come from Thomas Chong of Jefferies. Please go ahead.
Hi, Good evening, Thanks management for taking my questions, Matt asked about the pace for our AI to see part of the regimen, how has the traffic growth been and any key metrics, which can be sure on the deal generated search.
Benefits search traffic time span and then should we see the explosion of traffic.
Thank you.
Okay.
Yes.
Reconstructing all of our <unk> product.
Generally these AI.
I think generally I am financial models are making all of our products more powerful.
Over the past few months, we've made significant strides in this kind of efforts and the initial.
User feedback has been encouraging.
For.
The introduction of <unk> has enabled baidu to answer a broader range of questions.
Including more complex open ended and.
Comparative queries.
With early path, we can provide a direct and clear answer.
And more into active way.
The past few months in south just.
Lending some content, providing some links more and more search result.
What generated by learning Bot.
As a result users are engaging with baidu more frequently and asking new questions.
For example.
More and more users are coming to baidu for content creation.
Yet taxed or images.
Operator: During the Chinese new year holiday Baidu.
Operator: Baidu helped users create new year greeting messages and.
Operator: Generate personalized E card for their loved ones.
Operator: This is not a typical use case for search engine.
Chowdhuryn: But we see a large number of users rely on baidu for that kind of.
Chowdhuryn: Going forward.
Chowdhuryn: We will increasingly use Ernie bought two generic answers for third quarter at the end of them.
Chowdhuryn: And that use multi around conversation to clarify user intent for.
So that complicated user needs can be.
Speaker Change: A draft.
Speaker Change: Through natural language.
Speaker Change: While this initiative.
Speaker Change: Founded in <unk> and <unk>.
Speaker Change: Enhanced search experience.
Speaker Change: We.
Speaker Change: Still in the early <unk>.
Stages are using already bought.
Chowdhuryn: To reconstruct the Baidu search.
Chowdhuryn: We will continue to test and iterate Jenny I enable features.
Chowdhuryn: Down user feedback and we will follow our typical playbook.
Chowdhuryn: Our testing and fine tuning new experiences before upgrade.
Chowdhuryn: <unk> for larger scale rollout.
Chowdhuryn: Overall.
Chowdhuryn: We believe <unk> and AI will complement to get off of traditional search.
Chowdhuryn: Ultimately increase.
Chowdhuryn: Increasing user retention.
Chowdhuryn: Engagement and time spent on Baidu.
Chowdhuryn: And the ambition to search or any bought acting as our co pilot.
Chowdhuryn: <unk> has transformed up from.
Chowdhuryn: And application for users to find tablet and documents to.
Robin: To a one stop shop for users to create content.
Robin: In a wide range of formats.
Robin: And.
Year to date, I think about 18% of new paying users were attracted by Kenny features.
Robin: Cool.
Robin: Yes.
Robin: I want to emphasize that we are in the early stages of yielding.
Robin: Using our new box to reconstruct our apps and building new ones.
Robin Li: And at the same time, we are attracting and helping enterprises.
Robin Li: Bill asked online.
Robin Li: We believe that.
Robin Li: Success of Ernie Japan is wide and active adoption.
Robin Li: Whether through Baidu app or.
With third party apps.
Speaker Change: Thank you.
Speaker Change: Okay.
Speaker Change: Our next question today will come from James Lee of Mizuho. Please go ahead.
James Lee: Great. Thanks for taking my questions can you guys, maybe talk about or a niche technology road map for 2024.
James Lee: And crew.
Multi modal feature.
James Lee: Featured they'd be similar to <unk> or maybe opening about AI store or <unk>.
Speaker Change: Potentially launching AI each end.
Speaker Change: Is there any sort of milestone or key metrics you can speak to.
Speaker Change: And the second part to that question is on the cost of running Gen. AI, how should we think about maybe puts and take them managing inferencing costs going forward. Obviously, you talk about some way to make it more efficient are there any additional levers to optimize this process.
Speaker Change: Yes.
Speaker Change: Chips, we have on hand should enable us to advance <unk> four.
Speaker Change: To the next level.
Speaker Change: As I mentioned earlier, we will taken application driven approach to intense army.
Speaker Change: And let our users and customers tell us, where we should improve and adjust.
Robin Li: For our models.
Robin Li: It could be built.
Robin Li: Building multi modalities.
Robin Li: Agent.
Robin Li: Increasing reliability and fill up.
Robin Li: It's important to note that we are.
Robin Li: Focusing on using Ernie to bring real value to users and customers rather than simply achieving high rankings downloads research applications.
Robin Li: And Youre right price is very important issue.
Robin Li: Making high performance Foundation model.
Robin Li: Affordable is critical for large scale operation.
Robin Li: I mentioned earlier that.
Robin Li: We have been continuously reducing model insurance costs.
Robin Li: Now the insurance cost of $3 five is about 1% of.
Robin Li: March of 2020.
Robin Li: Sweet version.
Robin Li: And by doing so more and more enterprises are increasingly willing to.
Robin Li: Past develop and iterate their application.
Robin Li: Sure.
Robin Li: We understand that for many customer.
Robin Li: Tend to balance efficiency with cost and speed.
Robin Li: So we also offer a smaller language model and help our customers leverage Mou.
Robin Li: Mixture of experts.
Robin Li: For our best performance.
Robin Li: With our end to end approach. We believe there are still ample room to reduce the cost of our most powerful model and make them increasingly.
Robin Li: Affordable to our customers.
Robin Li: And this will further drive the adoption of <unk>.
Robin Li: Okay.
Robin Li: Internally, we are closely monitoring the number of apps developed by.
Robin Li: Ernie.
Robin Li: As I mentioned previously Ernie.
Robin Li: Now handling over 50 million queries per day.
Robin Li: And right now, earning API costs from internal application Das deal.
Robin Li: Larger than external costs.
Robin Li: But of course.
Robin Li: For Ernie in different sizes from.
Robin Li: External apps have been increasingly.