Q1 2024 Cheetah Mobile Inc Earnings Call
After todays presentation, there will be an opportunity to ask questions.
Speaker Change: To ask a question you May press Star then one on a touchtone phone.
Speaker Change: To withdraw your question. Please press Star then two.
Speaker Change: Please note this event is being recorded.
Speaker Change: I would now like to turn the conference over to Howard Investor Relations for Cheetah mobile.
Speaker Change: Thank you operator, welcome to the Cheetah mobile fourth quarter 2024 earnings conference call with US today, Our company Chairman and CEO, Mr. Fu Sheng and director and CFO, Nick just how much it fully management's prepared remark will work.
Speaker Change: In fact, the Q&A section before we begin I refer you to the Safe Harbor statement in our earnings release, which also applies to our conference call today as we will make forward looking statements. At this time I would now like to turn the conference call over to our chairman and CEO. Mr. Fu Sheng. Please go.
Speaker Change: Hi, Susan.
Speaker Change: Hello, everyone. Thank you for joining us today.
Speaker Change: This is our first in class C. Its November 20 221.
Speaker Change: And we are excited to share our progress.
Speaker Change: We assume our quality applebee's.
Speaker Change: Cheetah mobile user making changes.
Speaker Change: We are moving from a focus on placebo to beat.
Speaker Change: Our revenue from <unk> and others.
Speaker Change: Ooh enterprise focused increased by six 2% compared to last year and 36% from its approved.
Operator: Good day, and welcome to the Cheetah Mobile First Quarter 2024 Earnings Conference Call. All participants will be in a listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star and 1 on a touch-tone phone. To withdraw your question, please press star then 2. Please note, this event is being recorded. I would now like to turn the conference over to Helen, Investor Relations for Cheetah Mobile.
Good day and welcome to the Cheetah Mobile first quarter 2024 earnings conference call.
All participants will be in a listen only mode.
Speaker Change: Quarter.
Speaker Change: Nowadays revenues make up 43% of our total revenue.
Speaker Change: Should you need assistance please.
Speaker Change: Specialists by pressing the star.
Speaker Change: Zero.
Speaker Change: We expect this to grow this to grow to about 50% by the end of the year, making so can you just kind of step in our transformation.
After todays presentation there Bob.
Speaker Change: It will be an opportunity to ask questions.
Speaker Change: To ask a question you May press Star then one on a touchdown.
To withdraw your question. Please press Star then two.
Speaker Change: Oh I recently.
Speaker Change: Okay session awful leaching ore and stop the AI service provider, what's really important to move it gave us a skilled sales team.
Please note this event is being recorded.
Speaker Change: I would now like to turn the conference over to Helen Investor Relations for Cheetah mobile.
Helen: Thank you, Operator. Welcome to Cheetah Mobile's 4th Quarter 2024 Earnings Conference Call. With us today are our company's Chairman and CEO, Mr. Fu Shun, and Director and CFO, Mr. Tom McGinn. Following management's prepared remarks, we will conduct the Q&A section. Before we begin, I refer you to the safe harbor statement in our earnings treaties, which also applies to our conference call today, as we will make forward-looking statements. At this time, I would now like to turn the conference call over to Chairman and CEO, Mr. Fu Shun. Please go ahead, Mr. Fu.
Speaker Change: Thank you operator, welcome to Cheetah Mobile's fourth quarter 2024 earnings conference call weakness today, our chairman and CEO Mr. Fu Sheng.
Speaker Change: <unk> highest leads of business customers and the end to end.
Speaker Change: Probabilities for.
Speaker Change: Yes.
Speaker Change: Including MAU DAU trainee titanium developing.
Speaker Change: Developing.
Speaker Change: Based apps and enhancing.
Speaker Change: Director and CFO, Ned just how much it following management's prepared remarks.
Speaker Change: So these robots.
Speaker Change: Youll touch minutes for interactions with end users and the customers in the AI era.
Speaker Change: In fact, the Q&A session before we begin I refer you to the Safe Harbor statement in our earnings release, which also applies to our conference call today as we will make forward looking statements. At this time I would now electric to end the conference call over to our chairman and CEO. Mr. Fu Sheng. Please go.
Speaker Change: With our in stock we are now a focused full casing.
Speaker Change: Focusing on market, making customer enterprise apps Lees L. M S.
Speaker Change: Yeah.
Interest introducing M O M powered robot first.
Speaker Change: Perhaps.
Fu Shun: Hello, everyone. Thank you for joining us today. This is our first learning course since November 2021, and we are excited to share our progress as we resume our quality updates. Cheetah Mobile is making changes. We are moving from focus on 2C to 2B.
Speaker Change: Hello, Andrew.
Speaker Change: Thank you for joining us today. This is our first.
Speaker Change: The basic business needs.
Speaker Change: Since November <unk>.
Speaker Change: You say too many reasons for this focus.
Speaker Change: And we are excited to share our progress.
Speaker Change: First first market.
Speaker Change: We assume our quality our base.
Speaker Change: Market opportunity.
Speaker Change: Unlike competitive.
Speaker Change: Cheetah mobile user making catch it.
Speaker Change: Very few people to see market enterprise are increasingly choosing a M best.
Speaker Change: We are moving from a focus on policy.
Fu Shun: In Q1, our revenue from AI and others, all enterprise-focused, increased by 62% compared to last year and 36% from the previous quarter. Now these revenues make up 43% of our total revenue. We expect this to grow to about 50% by the end of the year, making significant steps in our transformation. Our recent acquisition of Beijing Orange Star, an AI service provider, was an important move. It gave us a skilled sales team, strong ties with business customers, and end-to-end compatibility for LMS, including model training, file turning, developing LLM-based apps, and enhancing service robots, a new way for interacting with end users and customers in the AI era.
Speaker Change: <unk>.
Speaker Change: On Friday, the clout due to VITAS. The current peak confess. However, however, they face challenges in developing Harris.
Speaker Change: In Q1, our revenue from AI and others.
All enterprise focused increased by six 2% compared to last year and the search is 6% from the pure <unk>.
Speaker Change: Yes.
Speaker Change: Presenting present King says.
Speaker Change: Yeah.
Speaker Change: So the potential opportunities in China's enterprise sector seconds.
Speaker Change: Nowadays revenues make up 43% of our.
Speaker Change: Our total revenue.
Speaker Change: Synergies.
Speaker Change: Expect this to grow this to grow to about 50% by the end of the year, making some bdcs tend to step in our transformation.
Speaker Change: Bringing together to achieve high and.
Speaker Change: Or is that.
Speaker Change: Allow us to combine our.
Pat: Hey, Pat.
Pat: Our App enterprise leads AI skills.
Speaker Change: Oh I raised it.
Pat: King the market opportunities and selling.
Speaker Change: Okay session awful leaching Arista.
Pat: I say D robots to business, where you can even find new ways to use the arris to improve efficiency.
Speaker Change: Service provider was another important move is it gave us a skilled sales team John highest leads of business customers and <unk>.
Pat: We are using our products driving approach.
Speaker Change: And.
Speaker Change: Probabilities for.
Pat: Our approach to enhance our comparability.
Speaker Change: Yes.
Speaker Change: The modal trainee criterion.
Pat: Compatibilities.
Speaker Change: This is why we focus on the 10 b.
Speaker Change: Developing.
Speaker Change: Based apps and enhancing.
Speaker Change: Metres SYGMA.
Speaker Change: Segments.
So these robots are new attachment for interactions with end users and the customers.
Speaker Change: Avoid large.
Speaker Change: Upfront upfront investment in Gpus, we believe without a change in parameters.
Speaker Change: The AI era.
Fu Shun: With Orange Star, we are now focusing on making customer enterprise apps with LMS and the LM Powered Robot. For specific business needs, we see two main reasons for this focus. 4th, 1st, March, Market Opportunity: Unlike competitive 2C markets, enterprises are increasingly choosing LRM-based apps on private clouds due to data security concerns. However, they face challenges in developing parallel apps and presenting a sustainable solution to substantial opportunities in China's enterprise
Speaker Change: But he's already stock we are now focused focusing focusing on market, making customer enterprise apps, Liza LMS and <unk>.
Speaker Change: Unnecessarily.
Speaker Change: Unnecessary.
Speaker Change: The enterprise can do it.
Speaker Change: And they use 10 P M.
Speaker Change: On private clouds.
Speaker Change: Inkjet introducing.
Speaker Change: Lower costs.
Speaker Change: Oh and powered the robot.
Speaker Change: Past few months, we changed our Ford cheap 14th E parameters Foundation models from scratch.
Speaker Change: So basically think is and its nice when you say to remain reasons for this focus.
Speaker Change: Which has been approved.
Speaker Change: First for Mark.
Speaker Change: Market opportunity.
Speaker Change: Also a case for a large scale rollout and they rank ranks among the top of a virus. These additionally, we are fine tuned nearly audio doing open great opening new stores.
Speaker Change: Unlike competitive compared to people to see market enterprise are increasingly choosing based.
Speaker Change: Own credit weighted clouds due to VITAS. The current peak consensus. However, however, they face challenges incubator.
Speaker Change: And the source foundation models to offer more options for our customers always out significant significantly increasing cocos.
Speaker Change: And Craig President Jing President King.
Speaker Change: So.
Speaker Change: So there's potential opportunities in China enterprise sector second.
Speaker Change: Furthermore, we have seen.
Fu Shun: Secondly, bringing together Cheetah and Orestar allows us to combine our app, our app enterprise with AI skills. After trying the market opportunities by selling robots to businesses, we can even find new ways to use LRMs to improve efficiency. We are using product driving approach to enhance our LRM compatibility. This is why we focus on the 10D, parameters, and LLM segments and avoid large upfront investments in GPUs. We believe that the training parameters of LM are unnecessary, are not necessary, and the enterprise can deploy and use 10B LMs on private clouds at a lower cost. Over the past few months, we've changed 14 beads.
Speaker Change: Seeing positive developments.
Speaker Change: Integrating RM based apps into our service robots in particular.
Speaker Change: Synergies, bringing.
Speaker Change: Bringing together to achieve that.
Speaker Change: Or is that.
Speaker Change: Our D E Berry robot can now.
Speaker Change: Allow us to combine our <unk>.
Speaker Change: Interacts better beta users leading to increased demand increased for the three months, especially in Japan and South Korea.
Speaker Change: Oh, App enterprise AI skills.
Speaker Change: The packaging market.
Speaker Change: <unk> opportunities.
Speaker Change: I say D robots to business, where you can even find new ways to use <unk>.
Speaker Change: Currently our overseas revenue has surpassed M S T. The revenues.
Speaker Change: We improved efficiency.
Speaker Change: And the continued to gorilla that did it.
Speaker Change: We are using them products driving approach.
Speaker Change: L. M. S. We believe the features of our robot service robot, where he spent even future further.
Approach to enhance our.
Speaker Change: Compatibilities this.
Speaker Change: This is why we focus on the 10 b.
I would also like to highlight how we I say assists our customers in using O M based apps efficiently.
Metres SYGMA.
Speaker Change: Pigment.
Speaker Change: A weight loss.
Speaker Change: Upfront upfront investments in Gpus, we believe will set a chain currently post.
Speaker Change: Remember, we have the Queensland University develop Oh, and East Chewy Fisher feature for our fault is app improving user experience.
Speaker Change: Unnecessarily.
Speaker Change: How necessary and the enterprise can give you.
Speaker Change: And they use 10 P M.
Speaker Change: Also developer.
Speaker Change: Private clouds.
Speaker Change: Howard customer service features.
Speaker Change: Lower costs.
Speaker Change: The past few months, we changed our four key 14 E parameters Foundation models from scratch.
Speaker Change: Another customer products, including Wechat mini programs apps and our service robots.
Fu Shun: Parameters Foundation Models from Scratch, which has been approved by authorities for a large-scale rollout and ranks among the top of various lists. Additionally, we are fighting against nearly all leading open-source foundation models to offer more options for our customers, all without significantly increasing costs. Furthermore, we have seen positive developments by integrating LM-based apps into our service robots. In particular, our delivery robot can now interact better with users, leading to increased demand, especially in Japan and South Korea. Currently, our overseas revenue has surpassed domestic revenues and continues to grow steadily.
Which has been approved.
Speaker Change: So with this is always service is now available in shop, how he low pill regimens apply for housing funds.
Our historic case for a large scale rollout and they rank <unk>.
Speaker Change: Among the path of the virus. These additionally, we are fighting nearly audio doing open open source open source foundation models to offer more options for our customers.
Speaker Change: Also we are also working with enterprise in China.
Speaker Change: France, you can see in past Chi to improve free improve management team, especially with I O M based apps.
Speaker Change: Without significant significantly increasing cost.
Speaker Change: He is the early stage of a L. M based apps developments, we closely work with our customers to understand their needs.
Speaker Change: Furthermore, we have.
Speaker Change: Seeing positive developments.
Speaker Change: And verify here's areas for improvement, let's say O M S.
Speaker Change: Integrating RM based apps into our service robots in particular.
Speaker Change: Our <unk> robot can now.
Speaker Change: Find the most appropriate H O M fine change and the developer customer apps.
Speaker Change: Interact with the users leading to increased demand increase the demands.
Speaker Change: This process how Pos so then so.
Speaker Change: Especially in Japan, and South Korea.
And then did some based apps and their capabilities.
Speaker Change: Currently our overseas revenue has surpassed <unk>, she said revenues and.
Speaker Change: How do you feel are leading customer service enterprise management and a chain.
Speaker Change: And they continue to gorilla that did it.
Fu Shun: With LMS, we believe the features of our robots, service robots, will expand even further. I would also like to highlight how we assist our customers in using LM-based apps efficiently. For example, we helped Hundun University develop a LM-based QA feature for its apps, improving user experience. We also developed LM-powered customer service features for another customer product, including WeChat mini programs, apps, and our service robots. This service is now available in Yinsha, helping local residents apply for housing funds.
Speaker Change: Which we can replicate to more customers.
Speaker Change: With the Oems, we believe the features of our robot, so basically robot where expense even future further.
Speaker Change: As a result, we are one team.
Speaker Change: Monitoring customer feedback and the certificate so dislocation.
Speaker Change: I would also like to highlight how we I'd say assists our customers in using <unk> based apps efficiently.
Speaker Change: Recently.
Speaker Change: The all the applications can be.
Speaker Change: Incorporated into our service robots.
Speaker Change: Assemble.
Speaker Change: Alberta, Queensland University develop.
Speaker Change: Our long term business model.
Speaker Change: <unk> Europe, where.
Speaker Change: East <unk> feature phone.
Speaker Change: Meanwhile savings.
Speaker Change: <unk> robots and offering them value added service.
Speaker Change: All his app improving user experience.
Speaker Change: We also developed.
Speaker Change: That's what we focus on building.
Speaker Change: Howard customer service features for another customer products, including Wechat mini programs apps and our service robots.
Speaker Change: Fab four enterprise playwear.
Speaker Change: We were shifting our resource from our agency.
Speaker Change: Internet business to AI business.
Speaker Change: Somebody improved operating margin of hour.
Speaker Change: This is <unk>.
Speaker Change: This is <unk> service is now available in Wuhan housing low pill regimens apply for housing funds.
Internet business, which.
Speaker Change: Should we use as our financial performance metric.
Speaker Change: In summary.
Fu Shun: We are also working with enterprises in China, branching the industry to improve management efficiency with LM-based apps. In the early stage of LM-based app development, we closely work with our customers to understand their needs and 35 areas for improvement with the LMS. Find the most appropriate HLM, fine-tune it, and develop a customer app.
Speaker Change: Hello Am is at once is a once in a generation opportunities with our in stock.
Speaker Change: We are also we are also working with enterprise in China.
Speaker Change: Brian GNC in past Chi to improve free improved management team, especially with the O M based apps.
Speaker Change: And our Korea strategies.
Speaker Change: We are now confident in our direction.
Speaker Change: I would like to emphasize that with boot barn to set short time revenue.
Speaker Change: In the early stage of a L M.
Speaker Change: Based apps development, we closely work with our customers to understand their needs.
Speaker Change: Revenue growth targets.
Speaker Change: We are aggressively for.
Speaker Change: And verify here.
Speaker Change: Gross.
Speaker Change: As areas for improvement.
Speaker Change: For archiving.
Speaker Change: Our customers certificate satisfaction and building lighthouse.
Find the most appropriate H O M five change and the developer customer apps. This.
Speaker Change: Jack's by doing so we believe we will establish a new growth new growth engine to Jive says says 10 dependable long term growth in both the revenue and the margins over time.
Fu Shun: This process helps us standardize some LRM-based apps and capabilities, particularly in customer service, enterprise management, and training, which we can replicate to more customers. As a result, we are monitoring, monitoring, customer feedback, and certification. Additionally, that all the applications can be incorporated into our service robot. As a long-term business model in LMS, we are involved in selling robots and offering valued-added services, as we focus on building LRM-based apps for enterprises.
Speaker Change: This process how Pos.
Speaker Change: Despite some pace to apps and our capabilities.
Particularly in customer service enterprise management and a changing.
Speaker Change: Which we can replicate to more customers.
Speaker Change: All we need is a bit of a patient with.
Speaker Change: Thank you authenticate employees for their hard work in making this happen. Thank you.
Speaker Change: As a result, we are one team.
Speaker Change: Monitoring customer feedback and the certificate so dislocation. Additionally.
Thomas: And Thomas.
Speaker Change: Thank you <unk> Hello, everyone on the call. Please note that unless stated otherwise all money amounts are in RMB terms.
Speaker Change: All the applications can be in cook incorporate into our service robots.
I am going to talk about two topics.
Speaker Change: Our long term business model in.
Speaker Change: Our continued investment in large language models are.
Speaker Change: Europe.
Speaker Change: Meanwhile savings.
Speaker Change: <unk> got bought and offering and valued added service.
Speaker Change: Resulting in operating loss for the quarter.
Speaker Change: Total revenue has resumed its increase.
That's what we focus on building apps.
Speaker Change: For enterprise, we were shifting our resource from our legacy.
Speaker Change: Second our healthy balance sheet.
Fu Shun: We will shift our resources from our agency in the internet business to the AI business. This will improve the operating margin of our... Internet business, which we use as a financial performance metric. In summary... LLM is a once-in-a-generation opportunity with Orange Star and our career strategies. We are now confident in our direction. We would like to emphasize that we don't want to set a short-term revenue growth target, but we are aggressively promoting our customer satisfaction and building lighthouse projects. By doing so, we believe we will establish a new growth engine to drive sustainable long-term growth in both revenue and margins over time. All we need is a bit of patience.
Speaker Change: First we are investing in firearms, we aim to help enterprises quickly develop.
Speaker Change: Internet business to AI business.
Speaker Change: <unk> improved the operating margin of hour.
Speaker Change: New apps I suppose Don mentioned in his speech our acquisition Orange Star has allowed service robots to become a key revenue contributor to the segment of the eye on others.
Speaker Change: Internet business, which.
Speaker Change: Should we use as our financial performance metric.
Speaker Change: In summary.
Speaker Change: Is that one.
Speaker Change: Is a once in a generation opportunities with.
The Q1 of 2024 revenues from AI, and others increased by 62% year over year, and 36% quarter over quarter to $81 million.
Speaker Change: Or is that all.
Speaker Change: And our careers package.
Speaker Change: We are now confident in our direction.
Speaker Change: I would like to emphasize that a weight boot longed to set short time revenue.
Speaker Change: Accounting for 43% of total revenue in the same period.
Speaker Change: Revenue growth targets.
Speaker Change: We are aggressively for growth.
Speaker Change: Driven by contributions from Beijing, Orange staff, our total revenue increased by two.
Speaker Change: Yeah.
Speaker Change: For archiving.
Speaker Change: <unk> percent year over year, and 14% quarter over quarter to $119 million.
Speaker Change: Our customers statistically satisfaction and building lighthouse.
This acquisition also allow the two teams from Cheetah and Orange star to work more efficiently together to better capture the opportunity in firearms.
Jack: Jack I doing so we believe we will establish a new growth new growth engine to Jive.
Jack: So tenable.
Jack: Tangible long term growth in both revenue and margins over time.
That should be how Chinese enterprises developed apps.
Speaker Change #100: Our arms to boost productivity.
Jack: All we need is a bit of a patient we think you authenticate employees for their hard work in making this happen. Thank you.
Fu Shun: We thank all the dedicated employees for their hard work in making this happen. Thank you, and Thomas. Thank you, Kudo.
Speaker Change #101: We expect this will lead to a substantial growth in revenue over time.
Speaker Change #102: In addition.
Speaker Change #102: Right.
Speaker Change #103: Enabling us to improve the product experience provided by our service robots.
Speaker Change: And Thomas.
Tom McGinn: Thank you, Fudo. Hello everyone on the call. Please note that, unless stated otherwise, all money amounts are in RMB terms. Today, I'm going to talk about two topics. First, our continued investment in large-language models, or LLMs, resulting in a widened operating loss for the company, while total revenue has resumed its increase. Second, our healthy balance sheet. First, we are investing in RRMs. We aim to help enterprises quickly develop new RRM-based apps.
Speaker Change: Thank you <unk> Hello, everyone on the call. Please note that unless stated otherwise all money amounts are in RMB terms.
Speaker Change #104: <unk> are now more capable of answering users different inquiries.
Speaker Change #105: This enhancement has strengthened our competitiveness and drive the sale of our service robots overtime.
I am going to talk about two topics.
Speaker Change: Our continued investment in large language models.
Speaker Change #105: In Q1 of 2024, our total non-GAAP costs and expenses increased 21% year over year.
Speaker Change: And resulting operating loss for the quarter, while total revenue has resumed.
Speaker Change: Chris.
Speaker Change: Second our healthy balance sheet.
Speaker Change #105: 19% quarter over quarter.
Speaker Change: First we are investing in our arms, we aim to help enterprises quickly develop.
Speaker Change #106: Our non-GAAP operating loss was $66 million in the quarter.
Speaker Change #106: <unk> from $42 million in the same period last year and $49 million in the previous quarter.
Don: New apps as Don mentioned in his speech our acquisition of Orange Star has allowed service robots to become a key revenue contributor to the segment of the eye and others.
Tom McGinn: As Puzong mentioned in his speech, our acquisition of Orange Star has allowed service robots to become a key revenue contributor to the segment of AI and others. In Q1 of 2024, revenues from AI and others increased by 62% year-over-year and 36% quarter-over-quarter to $81 million, accounting for 43% of total revenue in the same period. Driven by contributions from Beijing Aurene Star, our total revenue increased by 12% year-over-year and 14% quarter-over-quarter to $190 million.
Speaker Change #106: This is primarily due to the investments in Taiwan.
Speaker Change #106: Mentioned earlier.
Speaker Change #107: Through making our in stock we acquired many R&D talents to key sales personnel.
Don: Q1 of 2024 revenues from others increased by six 2% year over year.
Speaker Change #108: A very important for us to capitalize on the opportunity in this sector.
Don: 36% quarter over quarter to 81 million accounting for 43% of total revenue in the same period.
Speaker Change #109: As of March 31, 'twenty 'twenty four we had about 860 employees.
Up from about 720, a year ago. We are also ramping Gpus for module training on our <unk>.
Don: Even by contributions from Beijing, R&D staff, our total revenue increased 12%.
Don: <unk> year over year, and 14% quarter over quarter to $119 million.
Speaker Change #110: Excluding the impact of the aforementioned investments.
Tom McGinn: This acquisition also allowed the two teams from Cheetah and Orange Star to work more efficiently together to better capture the opportunity in LLM, as we help Chinese enterprises develop apps on LRMs to boost productivity. We expect this will lead to a substantial growth in revenue over time. In addition, LRMs are enabling us to improve the product experience provided by our service robots, which are now more capable of answering users' different inquiries.
Speaker Change #111: Our cost and expenses.
Don: This acquisition also allow the two teams from Cheetah and our in stock to work more efficiently together.
Speaker Change #111: Our margins remained stable.
Speaker Change #112: For example, <unk>.
Speaker Change #113: <unk> SBC, our operating profit for the Internet business.
Don: To capture the opportunity.
And as we help Chinese enterprises develop apps.
Speaker Change #114: Seven 9% in the quarter.
Don: To boost productivity.
Speaker Change #114: Up from three <unk> percent in the same quarter last year.
Don: We expect this will lead to a substantial growth in revenue over time.
Speaker Change #115: We continue to review our product portfolio, and then removed products that did not address user pinpoints.
Don: Alright.
Don: Are enabling us to improve the product experience provided by our service robots.
Speaker Change #116: We will continue this approach moving forward.
Speaker Change #117: At the same time, we will continue to invest in talent both in RMB specialized in <unk>.
Don: Which are now more capable of answering users different inquiries.
Tom McGinn: This enhancement has strengthened our competitiveness and should drive the sale of our service robots over time. In Q1 of 2024, our total non-GAAP costs and expenses increased 21% year over year and 19% quarter over quarter. Non-GAAP operating loss was $66 million in the quarter, up from $42 million in the same period last year and $49 million in the previous quarter.
Don: This enhancement has strengthened our competitiveness and drive the sale of our service robots overtime.
Speaker Change #118: And <unk> sales personnel.
Speaker Change #119: To help us the.
Speaker Change #120: Opportunity to fill the new growth engine for sheet.
Don: In Q1 of 2024, our total non-GAAP costs and expenses increased 21% year over year.
Speaker Change #121: How are your investments there will be backed by our strong cash reserves at the same time, we will continue to increase our operating profit for the Internet.
Don: 19% quarter over quarter.
Speaker Change #122: Secondly, Cheetah mobile has a healthy balance sheet as of March 31st some type of default.
Don: non-GAAP operating loss was $66 million in the quarter up from $42 million in the same period last year.
Speaker Change #122: We had cash cash equivalents and short term investments.
Don: $49 million in the previous quarter.
Tom McGinn: This is primarily due to the investment in our time mentioned earlier. Through Beijing Orange Star, we acquired many R&D talents and 2D sales personnel, which are very important for us to capitalize on the opportunity in this sector. As of March 31st, 2024, we had about 860 employees, up from about 720 a year ago. We are also running GPUs for model training and fine-tuning, excluding the impact of the aforementioned investment in our arms, our costs and expenses as well as our market remains stable.
Speaker Change #123: The U S dollar $215 million. In addition, we had a pound U S. Dollar 130 million of long term investments, which includes several holdings.
Don: This is primarily due to the investments in our plan.
Don: I mentioned earlier.
Don: Beijing are in Star, we acquired mainly R&D payloads to be salesperson, now, which are very important for us to capitalize on the opportunity in this sector.
Speaker Change #124: Even though known entities such as meat hustled off season.
Speaker Change #125: Lastly, you learned is the practice.
Don: As of March 31, 'twenty 'twenty four we had about 860 employees.
Speaker Change #127: Parable, China based companies listed in the U S capital market, we have decided not to provide revenue guidance going forward. Thank you.
Don: From about 720, a year ago. We are also ramping Gpus for module training on our fund fee.
Speaker Change #125:
Speaker Change #126: Everyone for today's call management will answer questions in chat room.
Don: Excluding the impact of the aforementioned investments in our.
Speaker Change #128: Yeah, and we'll translate management commerce into English and a separate line. Please note in the translation.
Don: Our cost and expenses.
Don: Our margins remained stable.
Speaker Change #128: And finally in the case of MH.
Tom McGinn: For example, excluding SBC, our operating profit for the Internet business was 7.9% in the quarter, up from 3.1% in the same quarter last year. As we continue to reveal our product portfolio and remove products that do not address user pinpoint needs, we will continue this approach moving forward.
Don: For example, excluding SBC, our operating profit for the Internet business.
Speaker Change #129: Management are failing Chinese whatsoever.
Speaker Change #130: You are able to steer the Chinese and English translation.
Don: <unk> of 9% in the quarter.
Speaker Change #131: First question in English will be available on our App.
Don: Up from three 1% in the same quarter last year.
Speaker Change #132: Si bookings tend and working day. Thank you so much operator.
Don: As we continue to review our product portfolio, and then removed products that did not address user pinpoints.
Patrick: Now Patrick Thank you.
We will continue this approach moving forward.
Tom McGinn: At the same time, we will continue to invest in talent, both in R&D specialized in LRMs, and one will be a salesperson now, to help us seize the opportunity to build a new growth engine for cheetahs. Our investment will be backed by our strong cash reserve. And at the same time, we will continue to increase our operating profit for the Internet business.
Patrick: We will now begin the question and answer session.
Don: At the same time, we will continue to invest in talent both in RMB.
You asked the question you May Press Star then one on your Touchtone phone.
Don: Utilizing our IMS and.
Don: <unk> sales personnel to.
Operator: If you are using a speakerphone please pick up your handset before pressing the team.
Don: To help us pick the.
Don: Opportunity to build a new growth engine for <unk>.
Operator: If at any time. Your question has been addressed and you would like to withdraw your question. Please press Star then two.
Don: Our investments they will be backed by our strong cash reserves at the same time, we will continue to increase our operating profit for the Internet business.
Speaker Change #135: At this time, well pause momentarily to assemble our roster.
Tom McGinn: Secondly, Cheetah Mobile has a healthy balance sheet as of March 34, 2024. We had cash and cash equivalents, and short-term investments of about U.S. dollars 250 million. In addition, we had about U.S. dollar 130 million of long-term investments, which include several holdings in well-known entities, such as metasou.dn. Lastly, in line with the practice of comparable China-based companies listed on the U.S. capital market, we have decided not to provide revenue guidance going forward. Thank you.
Don: Secondly, Cheetah mobile has a healthy balance sheet as of March 31st 2034.
Speaker Change #136: Ladies and gentlemen, please standby for the English translation of the question and answer session.
Cash and cash equivalents and short term investments of about <unk> dollars $250 million. In addition, we had about $130 million of long term investments.
Speaker Change #137: The first question.
Speaker Change #138: What are the plans ingalls, reaching down in 2024 wage areas that they plan to focus on customer base technology or products.
Don: It includes several are holding.
Don: And well known entities, such as meat hustle of CN.
Speaker Change #139: So I think our goal is to thoroughly implement our strategic transformation edits to say after several years Cheetah mobile has gradually shifted from a company with a focus on the home market to a company with a focus on the land market in some capability. Our main focus is to be in the artificial.
Don: Lastly in land is the practice.
Don: Terrible China based companies listed in the U S capital market, we have decided not to provide revenue guidance going forward. Thank you.
Operator: Everyone, on today's call, management will answer questions in Chinese, and an AI agent will translate management's comments into English on a separate line. Please note the translation is for convenience purposes only. In the case of any discrepancy, management's statement in Chinese works well. If you are unable to hear the English translation, a transcript in English will be available on our website within seven working days. Operator, please now take questions. We will now begin the question and answer session. To ask a question, you may press star and then one on your touchtone phone. If you are using a speakerphone, please pick up your handset before pressing the keys.
Don: Yes.
Speaker Change: Everyone for today's call management will answer questions in channel.
Our intelligence launch model, which is a technological ways, what we really need to focus on is taking a good job in the application of artificial intelligence and building. This direction. This direction is the core strategy of our entire company and we have nearly Santa company slogan, which is that we want to become.
Speaker Change: H M will translate management commerce into English and a separate line. Please note the translation.
Speaker Change: Tony.
Speaker Change: In the case of energy management.
Speaker Change: And Chinese works well.
Speaker Change: You are able to steer Chinese and English.
Speaker Change: Translation.
Speaker Change #140: Provider of new intelligent productivity tools in the era of artificial intelligence of course is part of our activity tool mainly refers to the to the industry. At this time regarding the specific points in Nanjing. We think it is still the product although artificial intelligence is very popular there are not many truly ROI.
Speaker Change: First question in English will be available on our at our uptime within seven working day. Thank you so much operator.
Speaker Change: Now take questions. Thank you.
Operator: We will now begin the question and answer session. To ask a question, you may press star then 1 on your touchtone phone. If you are using a speakerphone, please pick up your handset before pressing the key. If at any time your question has been answered and you would like to withdraw your question, please press star then 2. At this time, we will pause momentarily to assemble our roster.
Speaker Change: We will now begin the question and answer session.
Speaker Change #141: Key products, maybe the technical roles with large models is very powerful but there are not many cases that can be used by enterprise users now we think that now we needed to a solid job in the application of enterprise Skus artificial intelligence to help them take effect indication we will.
Speaker Change: You asked the question you May Press Star then one on your Touchtone phone.
Speaker Change: If you are using a speakerphone please pick up your handset before pressing the team.
Speaker Change: If at any time. Your question has been addressed and you would like to withdraw your question. Please press Star then two.
Speaker Change: At this time, we will pause momentarily to assemble our roster.
Speaker Change #142: Our service robots with our large months, so that our service robots can have better interaction capabilities that are self awareness and measurement capabilities and be used by our enterprise customers and more scenario I think it will be very good if we can do a good job in construction in 2024.
Speaker Change: Ladies and gentlemen, please standby for the English translation of the question and answer session.
Speaker Change #144: The second question is that Cheetah mobile is a company that focuses on QVC business now the company wants to change into two P. Two large model private deployment can make robots, where does your confidence come from to be business as different from two.
Speaker Change #143: And it may need to spend a lot of energy, maintaining customary and customer relationship how does precedent who lay out.
Operator: The first question is... What are the plans and goals for Cheetah in 2024? Which areas does it plan to focus on? Customer-based technology or products?
Speaker Change: The first question.
Speaker Change: What are the plans and goals for Ciena in 2024 wage Ariane does plan to focus on customer base technology or products.
Speaker Change #145: After that online store, which was originally invested by Chi tap also spent a lot of LNG existing Orion style game because the seamless. This means came out it was continued to be in.
Speaker Change #146: This is equivalent to the market. So in this process. They have a team and we also participated in some and then we also learned a lot of experience in the process of internal Orion Starz acquisition and I myself for that to be that you said. We also spent that is to say we spent a long time Larry.
Speaker Change #147: And this includes what you said about spending a lot of energy maintaining customers and customer relationships I think the most important thing is to build a set of organizational capabilities suitable for <unk>. We have also spent a lot of time in the past six months on various one more thing I want to say is that in a day.
Speaker Change #148: And to the internal units have a wrong time, we actually did a lot of work in <unk>, a few years ago, including a business called plan, which is to provide enterprises with cloud services from Amazon and Google. We are also Google in China, its equivalent to a not a cold metal, but a couple of them.
Speaker Change #149: Partner. So in fact at that time, we had already begun to continuously explore how to communicate with to be customers and how our organization can adapt and such and QB market. It should be said that indeed, the transformation from cedar to be has a lot of this cloud transformation pain.
Speaker Change #150: But we including me and our management has spent a lot of energy not only learning, but also packages.
Speaker Change #151: You said that maintaining customer relationships and customer relationships. This is indeed time consuming to do to be of course now that our organization has been built there is a similar to a wild wings iron triangle and we have a dedicated they are positioned to serve our customers online.
Speaker Change #152: How fun time, maintaining NASA customer relationship, but more communicating with customers to obtain Danny because only when I use the top theater to understand the needs of customers can I do a great job in the tubing business, which is also what we have found in the past few years for this.
Speaker Change #153: Play out I think our idea today is first of all we have to build this benchmark customer and we now have several top customers in this industry who are delivering.
Speaker Change #154: Delivering and crucial July although we aren't going to be the role of QC is to attach importance to the user experience, which is still our lifeblood and we will definitely make our customers feel that this is enough on that is the services and products, we provide to them right now.
Satisfy them after building a benchmark customer as I just said, we can standardize some of our own standard parts and then replicate them second lien. Another thing is that due to the construction of R&D organization. Many of our customer relationships can be borrowed against this kind of business up large.
Models and can also be used in the business. Our profile here. There are many cross case scenarios when customers have large models and machines purchased together. So I think that Cheetah mobile is now in a startup period again, we can't say that there is any particularly take lay out the more important.
Speaker Change #155: I had to choose the artificial intelligence frame and then charge the customer as well under the tide of artificial intelligence, we really do applications in a downturn manner to make customers feel satisfied.
Speaker Change #156: The third question is that the company's accounts receivable prepayment and accounts payable are relatively large can you explain why business. This is caused by how well the company manage receivables and payables.
Speaker Change #157: Thank you for your question the large amount can be several accounts are all related to one of our advertising agency thicknesses or cheetahs advertising agency business is to help many Chinese advertiser several relatively large overseas platforms online broadcasting plat.
Speaker Change #158: Funds to purchase advertisement since the amount of revenue we recognize only the advertising agency fee. The full amount of the customers purchase of advertisements and our payment to the advertising platform is recorded in the two account for payment and payable that you. Just mentioned this business is actually also a <unk> business and we have been.
Speaker Change #159: Operating it for nearly 10 years. However, during these 10 years, we have actually formed a very strict mechanism to evaluate the credit performance of advertisers and manage the accounts receivable and payable periods. We are still very confident in the cash management of this business.
Speaker Change #160: This question is about how the company plans to make Orion unlocks value for Cheetah shareholders and whether it will consider lifting all lines separately.
Speaker Change #161: You mentioned Orion in fact, after our acquisition the focus of Orion and China businesses that most of the welfare business somebody group has placed in this entity of Orion. However, a domestic company of Cheetah mobile we are always committed to creating the greatest value for sure.
Speaker Change #162: Shareholders of Cheetah mobile guarding the planning of Orion, we will comprehensively evaluate various capital operation opportunities, including the possibility of lifting the subsidiary separately are conducting independent financing. Our goal is to enhance the market value of orion's business performance through effective ways, thereby further.
Speaker Change #163: Enhancing the stock price of the entire company and every room, we will fully consider the market environment company's strategy and the long term interests of shareholders of Cheetah mobile to ensure that every step we take and bring the greatest return to shareholders.
The first question is that many cloud centers did project base work in the previous years, which was widely criticized but innovation does cheetah has in the private deployment of large model. What is the level of project revenue and profit margin that we are doing like Atlanta priced applications of large models with.
Speaker Change #165: A private deployment many cloud vendors provide standardized model fine tuning tools and application development tools for enterprises to use and to a certain scale and location volume has been for them at the same time the insurance cost of large models of cloud vendors is constantly reducing and even some models correctly.
Speaker Change #164: Free in this case why do enterprises still need private deployment what types of characteristics of enterprises are suitable for private deployment of large models.
Speaker Change #164: Aye.
Speaker Change #166: We will answer briefly first regarding the project based system with cloud vendors to be honest I don't know much about it because as far as I know some cloud vendors contracts are particularly large huge private deployment in the cloud is actually not the same concept as the planning we are talking about.
Speaker Change #167: Today, but in fact cloud vendors have been providing their customers with good enough deployment services, just that companies like Amazon their deployment and personnel costs are very high so they let partners completed just like we went there and undertook many such projects and because such partners have lower costs. There's also a lot of <unk>.
Speaker Change #168: <unk> regarding the revenue and profit margin of our projects today, frankly speaking the real thing we do to help enterprises with private deployment is still in the stage of benchmarking as I just mentioned and the manuscript we are not considering too much about this because of the use of consideration. However.
Speaker Change #168: The model we are launching now it's more about being able to work with partners since its equivalent to a fill in model. When we have a particularly clear time, we will talk about it again and why to large enterprises, meaning private deployment of large model because the larger they enter.
Speaker Change #169: Price the more consideration it will get to data security and today when you really use a large model what you're really past two it is a lot of internal documents of the enterprise, especially some sensitive documents in fact, the vast majority of enterprises are very concerned about there.
Speaker Change #169: Is it because the injunction instability of large models comes from data after the data on the Internet and exhausted the data inside the enterprise is also a very important data source. So at least at the customer level, we see quite a lot of concerns and basically companies all of a sudden size or.
Speaker Change #170: Require private deployment of large models, a difference between us and previous cloud such that when we do private deployment first the cost of private appointment of large models itself. It's not high it is not a complex deployment system. In fact today you can basically deploy large.
Speaker Change #171: Models into it by using some servers to open servers. So the cost of this deployment itself is very low secondly, when doing project base work because today, we are doing AI implementation. When we deploy large models large models have a different ability.
Speaker Change #172: From previous businesses because of their own reasoning and comprehension abilities are relatively strong. This makes it easier for us to approach customers in other industries and in the previous era, including not only the cloud, but also fast in fact, our ability to cross domain has been greatly enhanced compared to before that.
Speaker Change #173: Is to say in the past if I didn't know enough about this industry. It was actually very difficult for me to do true, but because large models understand themselves I don't know if you understand what I mean that is they will understand a lot of professional knowledge by themselves. So our deployment work.
Speaker Change #174: It's not called appointment, but the workload of helping them to applications will be much less than before and once this strong ties forums. It's replicability will also be much stronger for example, we just mentioned that we did this for a government project ticket Providence findings and it took us a long time.
Speaker Change #175: And to do it when the second customer came our deployment might only take two or three weeks to completing therapy. You asked that the inference cost of large models of cloud vendors is constantly decreasing and even some models directly free now most of the free model with our open source models and when using <unk>.
Speaker Change #176: Can source models the vast majority of customers we encounter require private deployment by private deployment is divided into two aspects. One is called private deployment within the internal network, which is required for enterprises with high security and the other is called private deployment in the.
Speaker Change #177: Loud that is I deploy a model in the cloud, but the model can only be used by Nielsen data cannot be crushed and this kind of deployment is also a private deployment. So what you. Just asked is actually mainly due to data security considerations of course, if you use the large model a lot it's not true.
Speaker Change #177: One machine.
Speaker Change #178: Cost of this definition is constantly decreasing but now talk he is also a challenge that this actually still has some advantage in the third and what kind of enterprises and what are suitable we think that the larger the enterprise the more basically needs data security so.
Speaker Change #179: Conversely. This is also a good thing for us more of a fee enterprises needs come from large enterprises with strong payment capabilities.
This question is about how you think about the relationship between the robotics business and large scale models what are the promotional effects of the company's all in large scale model enterprise applications on the robotics business today.
Speaker Change #180: The customer base of our robot is enterprise users and after our acquisition the store was cleared out and they not only have more of an agency system that is their agents, but also to a lot of enterprise information implementation and deployment services.
Speaker Change #181: So just from the channel perspective, a large part of that can be reused secondly from a technical point of view. This large model is the brain of the robot in the past except for industrial machinery. Other robots in the robotics industry had not developed well one of the importantly, since in fact, the brainpower of them.
The machine is limited and now why is the robotics industries. So popular in fact after the breakthrough with the large scale model, which brings enhanced decision, making and judgment capabilities now everyone believes that the robotics industry, whether it is service robots or.
Speaker Change #182: Robots or even humanoid robots will have a bright future. So what is the relationship between us on the one hand as we just mentioned the customer relationship can be reused and a large part of that can be reused or enterprise customers in there.
Speaker Change #183: <unk> segment. Once you establish a connection you will find that those who sell robots are also interested in a large model when you talk to them about it. They will feel that you can also help them improve many positions before secondly, more importantly, if we do not develop the capabilities of the large.
Speaker Change #184: Modern well our robots will lose its competitiveness in the long run right. Because we are not just a hardware manufacturer, but really focused on its autonomous decision, making capabilities now through the training fine tuning and application of the large scale model, we have a clearer understanding of how to apply the capabilities.
The large model to robots, we have already started some training in this area internally, it's no longer just about training the large scale model, but also combining the robotics with the large scale model and the intelligent language model its capabilities will continue to expand.
Speaker Change #185: At this stage, what we can see is that in the past few years, we have been doing a lot of voice interaction, but the growth has not been good enough because they use a base, it's not large enough and when the questions go beyond the scope. It cannot answer now with the large scale language model the smoothness.
Speaker Change #186: In satisfaction of the communication has been greatly improved we have also disclosed a data to the market.
Speaker Change #187: At that time, we helped customers to the Samsung T O N Dot Hill, Colin Die and we were able to achieve an accuracy rate of about 97%. This data shows that when a robot cancer with such a high accuracy rate during the explanation. It's practicality is equivalent to that of a.
Speaker Change #188: Human and this increase in accuracy is not achieved by a large amount of manual work in the past, but by and putting some documents into it and it can achieve a high level of accuracy. Therefore at this stage. It is obvious that the demand for robots and this young take care shop area is increasing.
Speaker Change #189: Especially since a large scale motto is itself diverse in the past our robots overseas, where mostly low speed because crossing a language with a lot of work for us, but now due to the large scale model. It is naturally multilingual. So we will also launch this in overseas snacks that language interactive robots.
Speaker Change #190: Of course in the long run we are also doing some training on robotic arms to enable robots to do some work, but they still need some time.
Speaker Change #190: Thank you.
Speaker Change #191: My question is about chips under the background of high end chips being restricted in China, well Cheetah continue to train its own large model when fine tuning and Iterating model for enterprise customers, how does cheetah salt the chip problem.
Planning, we are talking about today, but in fact cloud vendors have been providing their customers with good enough deployment services. It's just that companies like Amazon their deployment and personnel costs are very high so they let partners complete it just like we went there and undertook many such projects and because such partners have lower cost.
Speaker Change #192: This was founded in 17 16, and they were already doing artificial intelligence back then and in 2017 Cheetah mobile has even shout. It out the slogan of Oi and also collaborated a lot with six star in the aspect of AI, So I experienced and start from last year, although the large.
There is also a lot of profit regarding the revenue and profit margin of our projects today frankly speaking the real thing we do to help enterprises with private deployment is still in the stage of benchmarking as I just mentioned and the menu script, we are not considering too much about this because of user can.
Speaker Change #193: Language model has some different characteristics from the previous models the underlying network the underlying neural network. The transformer we have already used in the earliest T and some of the go to child was used in the speech model. So our entire team's understanding of the transformer had been through a long term precipitation well that's fine.
So durations. However, the model we are launching now it's more about being able to work with partners. This is equivalent to a fill in model. When we have a particularly clear time, we will talk about it again and why to large enterprises need private deployment of large model.
Speaker Change #193: Tuning thing you said the competitive advantage I think it comes more from the granularity that is being able to do is find enough because fine tuning itself is the preparation of about hundreds of thousands of pieces of corpus and also the refinement of this data. According to this scenario. It also requires.
Because the larger the enterprise the more consideration it will give to data security and today. When you really use a large model what you're really past two it is a lot of internal documents of the enterprise, especially some sensitive documents in fact, the vast majority of enterprises.
Speaker Change #194: A lot of careful and detailed management as well as communication with the needs of users here I think if we talk about competitive advantages or who has any competitive advantages in such a fierce market. It is very difficult for a company to say that it has any unique and insurmountable advantages in technology.
Are very concerned about this because the endogenous ability of large models comes from data after the data on the Internet is exhausted the data inside the enterprises also a very important data source. So at least at the customer level, we see quite a lot of concerns and basically companies.
Speaker Change #195: Do you think are more advantages come from the combination with customers in a market that is what we really focus on is the process of rapid iteration, rather than a certain point that you can do and others cant. So we constantly emphasize the importance of users word of mouth and the implementation of some projects and then you.
All of a sudden size or require a private deployment of large models a difference between us and previous clouds is that when we do private deployment first the cost of private deployment of large models itself is not high it is not a complex deployment system in fact today you.
You talked about private deployment itself in fact, the private deployment of large models is not difficult at all what we really do is not the private deployment of large models, but after deploying into the user's network. According to the user's business characteristics and business needs to do the corresponding.
Basically deploy large models into it by using some servers to open servers. So the cost of this deployment itself is very low secondly, when doing project base work because today, we are doing AI implementation when we deploy large models large models.
And the difficulty lies in that today's model capabilities have not reached the level of a universal AI also will not someone asked before.
Speaker Change #195: If the model was saying right I'm, sorry that is the real ability of today's launch languish motto has certain leasing but it is a large cap compared to the needs of the enterprise scenario. What is needed here is to do the application the competitive advantage of doing the application.
Have a different ability from previous businesses because their own reasoning and comprehension abilities are relatively strong. This makes it easier for us to approach customers in other industries than in the previous era, including not only the cloud, but also SAS in fact, our ability to cross domain has been greatly enhanced compare.
Speaker Change #196: Your insight into the customers' needs and they use up a whole set of technical means to help them provide a solution within this demand because the customer only cares about whether this thing is satisfactory to me not whether it is solved by the model.
Two before that is to say in the past if I didn't know enough about this industry. It was actually very difficult for me to do true, but because large models understand themselves I don't know if you understand what I mean that is they will understand a lot of professional knowledge by themselves.
Speaker Change #197: Is it falls by the model or other technologies in the application content. We have found through practice that a downward GDP I model when really used in many enterprises is just to do a professional knowledge Q and H and the satisfaction of customers is not satisfactory. This is our own practice so.
So our deployment workload has not called deployment, but the workload of helping them do applications will be much less than before and once the strong ties formed it's replicable and he will also be much stronger for example, we just mentioned that we did this for a government project to do Provident funds.
Speaker Change #198: It is necessary to customize some collagen according to the customer's needs and let the tortilla and work with the large model only after this collaborative work can be used to really reach the so called central employee role in fact, it seems that there is a lack of such a real solution in the market that can truly provide customers with satisfaction.
And it took us a long time to do it when the second customer came our deployment might only take two or three weeks to complete or the U S that the insurance cost of large models of cloud vendors is constantly decreasing and even some models are directly free now most of the free models are open source small.
Speaker Change #199: This is our understanding of the market. So when you ask about the competitive advantage. In fact, we are exploring the depths of the customers' needs and doing the detail as well as for the relevant talents on this 0.1st of all our leader because it involves us.
And when using open source models, the vast majority of customers, we encounter require private deployment, but private deployment is divided into two aspects. One is called private deployment within the internal network, which is required for enterprises with high security and the other is <unk>.
Speaker Change #200: We will not talk about it but he has also published papers and has sufficient academic and industrial foresight than in terms of the specific implementation. Some algorithm engineers. There is a considerable reserve of talent in China. At this point it is not too difficult to recruit such people in them.
Called private deployment in the cloud that is I deploy a model in the cloud, but the model can only be used by E data cannot be crushed and this kind of deployment is also our private deployment. So what you just asked is actually mainly due to data security considerations of course, if you use the large mall.
Speaker Change #201: So we are not going to compete in the large parameters and teaching on me of large model. So our demand for so called top talents is not that high we are more about combining the already oversupplied capabilities of large models to provide our enterprise customers with a set of solutions.
A lot it's not just one machine.
The cost of this definition is constantly decreasing but now I'll talk he is also a challenge and this actually still has some advantage and the third is what kind of enterprises and what are suitable we think that the larger the enterprise the more basically need state of security so.
Speaker Change #201: Is our focus.
Speaker Change #202: Under private deployment, how to solve the problem of continuous model iteration. When we selected to no confirmed slosh model based on the enterprise scenario for customers to do fine tuning deployment and application development and the enterprise started to use it but now the base model is evolving.
Conversely. This is also a good thing for us more of the enterprise's needs come from large enterprises with strong payment capabilities.
Speaker Change #203: And very quickly when the base model is updated with a completely over ranked the capabilities of the large model that'd be fine tuned for the enterprise.
Okay.
Yeah.
Operator: Okay.
Speaker Change #203: Yes.
Operator: Okay.
Speaker Change #204: Thank you very much for your question. There are a few concepts here that I would like to explain first namely fine tuning and application. These are two different concept in fact in most enterprise scenarios. There is no need to do large scale fine tuning specifically for the <unk>.
Speaker Change: This question is about how you think about the relationship between the robotics business and large scale models what are the promotional effects of the company's all in large scale.
Model enterprise applications on the robotics business today.
Speaker Change #205: Enterprise because the basic capabilities that the current model and the enterprise above 10 billion parameters. We now believe that the basic capabilities of a 300 billion parameter model can basically meet the requirements of most enterprise application framework. Moreover, most enterprises rarely have.
Speaker Change: The customer base of our robot is enterprise users.
So much data to provide that fine tuning to large models will bring more changes. Our current approach is to use what is called an application suite to combine the model with the needs at the enterprise well when the model is updated the application suite will not become obsolete because more of it is combined with some internal systems at the enterprise.
Speaker Change: And after our acquisition the store was cleared out and they not only have more of an agency system that is their agents, but also to a lot of enterprise information implementation and deployment services. So just from the channel perspective, a large part of that can be reused secondly from a.
Speaker Change #206: <unk> such as calling it you asked the large model questions such as how do I go about handling the document today. He will say what documents to you need to provide right you tell me or I D number and after you tell it he will go to check the interface with the I D number. If this is part of the.
Speaker Change: Cove point of view. This large model is the brain of the robot in the past except for industrial machinery. Other robots in the robotics industry have not developed well one of the important recency effect. The brainpower of the machine is limited and now why is the robotics industry. So popular.
Speaker Change #206: The application. After this application interfaces written your model will be updated again it doesn't.
Fu Shun: In fact after the breakthrough with the large scale model, which brings enhanced decision, making and judgment capabilities now everyone believes that the robotics industry, whether it is service robots or industrial robots or even humanoid robots will have a bright future. So one is.
Speaker Change #207: In fact him at all that's the first one secondly in fact after the model's capabilities are enhanced the smoothness of the application that is the accuracy rate and various aspects of the user experience will be improved I don't think this is.
Speaker Change: The relationship between us on the one hand as we just mentioned the customer relationship can be reused and a large part of that can be reused or enterprise customers in the <unk> segment. Once you establish a connection you will find that those who sell robots are also.
Speaker Change #207: Conflicting at all but now it seems that no matter how much the models capabilities are improved it cannot know the needs of every enterprise that is if you look closely at the various needs of the enterprise. They are different today's models will train based on this internet data and he for example.
Speaker Change: We're interested in a large model when you talk to them about this they will feel that you can also help them improve many positions before secondly, more importantly, if we do not develop the capabilities of the large model well our robot will lose its competitiveness in the long run Brian because we are not just.
Speaker Change #208: Has no idea what their administrative needs of Cheetah mobile are what are its internal documents and what are the needs of this employee is are all things that need to be solved by applications now it seems that the upgrade of this.
Hardware manufacturer, but really focused on its autonomous decision, making capabilities now through the training fine tuning and application of the large scale model, we have a clearer understanding of how to apply the capabilities of the large model to robots. We have already started some training in this area.
Speaker Change #209: Model and recently there are several startup models that we especially welcome because in this way some of our applications can be better done on it which is different from some of the previous one because the upgrade at the model. The first C. P. I interface will not change much. The second we also have a lot of interactions with him too.
Internally it is no longer just about training the large scale model, but also combining the robotics with the large scale model and the intelligent language model its capabilities will continue to expand at this stage. What we can see is that in the past few years, we have been doing a lot of voice into.
Speaker Change #210: What is called a prompt where these will not be affected after at the model is angry the model and the application suite or a complementary relationship and currently it seems that for a long time or I almost think that within the Hong Kong University. It is unlikely that our model can be deployed online and users.
Fu Shun: Our action, but the growth has not been good enough because the user base is not large enough and when the questions go beyond the scope. It cannot answer now with the large scale language model the smoothness and satisfaction of the communication has been greatly improved we have also.
Speaker Change #211: Can use that in fact, there are quite a lot of opportunities in great demand for enterprise level applications here.
My question is about chips under the background of pie and chips being restricted in China will Cheetah continue to train its own large model went fine tuning and Iterating models for enterprise customers, how does cheetah solve the chip problem.
Speaker Change #212: For the large model the number of parameters. We have is in the scale of just over 10 billion parameters as we just mentioned and we are still learning among this one but the number of parameters will not be much larger because we are more considering the actual cost of the enterprise.
Speaker Change #212: Because such a model with a certain number of parameters if really used probably only requires one a higher end server can be used. So these are combined with our needs. The users' needs. This is the first point so our demand for chips is not that great compared to many companies that do this change.
Speaker Change #212: Potential or larger parameter large models, our chip demand is not that great. Secondly, we think that today's training of this large model is indeed very nation, and we don't need to do repetitive large scale construction on this in fact, we thought earlier that the open source community would be very prosperous more and more.
Your next step language interactive robots of course in the long run. We are also doing some training on robotic arms to enable robot to do some work, but they still need some time.
Thank you.
Speaker Change #213: More open source models with good enough performance will emerge today. It seems that the situation is the same and it's also makes us when choosing a model not only our own model, but also many open source models that can be directly provided to customers. So we are not so worried about the chip issue.
My question is about chips under.
Under the background, the pie and chips being restricted in China or Cheetah continue to trade at some large model when fine tuning and Iterating model for enterprise customers, how to cheat us off the chip problem.
Speaker Change #214: And then we will now focus more on our own training. In addition to the large scale model that you won't gun asked about we may focus more on upgrading skills and mechanical collaboration because I think the supply of language models is already sufficient to meet market demand and we are also doing more applications on it we are not nature.
Speaker Change #216: And I changed the ability at the original model itself. So we are not so worried about the chip issue.
This was founded in 17 16, and they were already doing artificial intelligence back then.
Okay.
Speaker Change #214: Yeah.
Speaker Change #215: So thank you so much business halloran from Cheetah Mobiles IR teams as this is.
And in 2017, Cheetah mobile even shouted out the swelling of Orange and also collaborated along with fixed R&D aspect Guy So I experienced and start from last year, although the large language model has some different characteristics from the previous model the underlying neural network the underlying neural network.
Speaker Change #217: A translation line and all the translation Howard.
business halloran: That has been developed by our team itself. So thank you. So much for you our accommodations and we'll have an English a transcript to be available as soon as we can we'll put that on our IR website. I believe we think seven working day again. Thank you so much.
A transformer, which we have already used and the earliest T and some of the coach average use in the speech model. So our entire team's understanding of the transformer had been to a long term precipitation while it's fine tuning and he said the competitive advantage I think it comes more from the granularity that is being able to do.
Speaker Change #219: If you have any further questions. Please let us like there were actually no. Thank you so much bye bye.
Speaker Change #220: The conference has now concluded we thank you for attending today's presentation.
Two is buying enough because fine tuning itself is the preparation of about hundreds of thousands of pieces of content and also the refinement of this data. According to this scenario. It also requires a lot of careful and detailed management as well as communication with the needs of users here I think if we talk about <unk>.
Speaker Change #221: You may now disconnect your lines and have a nice day.
Additive advantages or who has any competitive advantages in such a fierce market. It is very difficult for a company to say that he has any unique and insurmountable advantages in technology. We think are more advantages come from the combination with customers in the market that is while we focus on in the process of rapid.
Okay.
Speaker Change #221: Okay.
Iteration, rather than a certain point that you can do and others cant. So we constantly emphasize the importance of users word of mouth and the implementation of some projects and then you talk about private deployment itself. In fact, the private deployment of large models is not difficult at.
What we really do is not the private deployment of large models, but after deploying it to the users network. According to the user's business characteristics and business needs to do the corresponding tolzien thing. The difficulty lies in that today's model capabilities have not reached the level of a universal AI also.
We'll not someone asked before if.
If the model.
Ray I'm sorry.
The reality of today's launch Languish model has certain reasoning, but it is a large cap compared to the needs of the enterprise scenario. What is needed here is to do the application the competitive advantage of during the application Miss your insight into the customers' needs and the use of a whole.
Set of technical means to help them provide a solution within this demand because the customer only cares about whether this thing is satisfactory to me not whether it is solved by the model.
Is it felt by the model or other technologies in the application content. We have found through practice that I downward GPI model when really used in many enterprises is just to do a professional knowledge Q&A and the satisfaction of customers is not satisfactory. This is our own practice so it.
As necessary to customize on power Gen. According to the customer's needs and let the culture and work with the large model only after this collaborative work can be user really reached to sell our digital employee role in fact, it seems that there is a lack of actual real solution in the market that can truly provide customers with satisfaction.
It is our understanding of the market. So when you ask about the competitive advantage. In fact, we are exploring the depth of the customers' needs and doing the detail as well as for the relevant talent on this 0.1st of all our leader because it involves us.
We will not talk about it but he has also published papers and has sufficient academic and industrial foresight Ben in terms of the specific implementation from algorithm engineers. There is a considerable reserve of talent in China. At this point it is not too difficult to recruit such people in the March.
So we are not going to compete in the large parameters and <unk> of large model. So our demand for so called top talent is not that high we are more about combining the already oversupplied capabilities of large model to provide our enterprise customers with a set of solution based.
Speaker Change: Is our focus.
Speaker Change: Okay.
Speaker Change: Under private deployment, how to solve the problem of continuous model.
Speaker Change: Ration when we selected an open source launch model based on the enterprise scenario for customers to do fine tuning deployment and application development and the enterprise started to use it but now the base model is evolving very quickly when the base model is updated with a completely over ranked the capability.
<unk> of the large model that will be fine tuned for the enterprise.
Fu Shun: Yes.
Speaker Change: Thank you very much for your question. There are a few concepts here that I would like to explain first namely contributing an application. These are two different concepts in fact in most enterprise scenarios. There is no need to do large scale fine tuning specifically.
Speaker Change: Or the enterprise because the basic capabilities of the current model and the enterprise above 10 billion parameters. We now believe that the basic capabilities of a $300 billion parameter model and basically meet the requirements of most enterprise application framework. Moreover, most enterprises rarely.
Speaker Change: We have so much data to provide that fine tuning to large models will bring more changes our current approaches to use what is called an application suite to combine the model with the needs of the enterprise well when the model is updated the application suite will not become obsolete because more of it is combined with some internal systems at the end.
Speaker Change: Surprise, such as calling it you asked the large model questions such as how do I go about handling the document today. He will say what documents to you need to provide ray is tell me are ideal for an FTE tell it he will go to check the interface with the I D. Number this is part of.
Speaker Change: The application. After this application interfaces ranking your model will be updated again it doesn't affect him at all that's the first one secondly in fact after the modeling capabilities are enhanced the smoothness of the application.
Fu Shun: That is the accuracy rate and various aspects of the user experience will be improved I don't think this is conflicting at all but now it seems that no matter how much the model's capabilities are improved it cannot know the needs of every enterprise that is if you look closely at the various.
Fu Shun: Needs of the enterprise they are different today's model told trained based on this internet data and heat for example has no idea what the administrative needs of Cheetah mobile are what our internal documents and what are the needs of this.
Fu Shun: Employees are all things that need to be solved by application now it seems that the upgrade of this model and recently there are several startup models that we especially welcome because in this way some of our applications can be better done on it which is different from some of the previous one.
Speaker Change: Because the upgrade of the model the first API interface will not change much. The second we also have a lot of interactions with him to call the prompt where east will not be affected after it the model is angry at the mall.
Model and the application suite or a complementary relationship and currently it seems that for a long time or I would think that within the Hong Kong University. It is unlikely that our model can be deployed online and users can use. It in fact, there are quite a lot of opportunities in great demand for enterprise level applications here.
Okay.
Okay.
Speaker Change: Okay.
Speaker Change: My question is about chips under the background of pie and chips being restricted in China will Cheetah continue to train its own large model.
Speaker Change: When fine tuning and Iterating model for enterprise customers habits, Cheetos salt the chip problem.
Speaker Change: For the large model the number of parameters. We have is in the scale of just over 10 billion.
Speaker Change: Parameters as we just mentioned and we are still learning alone.
Speaker Change: But the number of parameters you will not be much larger because we are more considering the actual cost of the enterprise because such a model with a certain number of parameters if really you'd probably only requires one a higher end server can be used. So these are combined with our niche the users' needs. This is the <unk>.
Speaker Change: First point, so our demand for chips is not that great compared to many companies that do this chain essential or larger parameter large motto. Our chip demand is not that great. Secondly, we think that today's training of this large model is indeed, purion H N and we don't need to do repetitive large scale.
Speaker Change: <unk> on that in fact, we thought earlier that the open source community would be very prosperous more and more open source models with good enough performance will emerge today. It seems that the situation is the same and this also makes us when choosing a model not only our own model, but also many.
Speaker Change: Open source model that can be directly provided to customers. So we are not so worried about the chip issue and then we will now focus more on our own training. In addition to the large scale model. Thank you won't Gung asked about we may focus more on upgrading skills and mechanical collaboration because I think the supply of language modeled this out.
Speaker Change: Already sufficient to meet market demand and we are also doing more applications on it we are now in nature and I changed the ability and the original model itself. So we are not so worried about the chip issue.
Fu Shun: Okay.
Speaker Change: So thank you so much visit tallow from Cheetah Mobiles IR team addresses.
Speaker Change: A translation line and all the translation power buying ETP that has that developed by our team itself.
So thank you so much for your accommodation and Wawa.
Speaker Change: English try script to be a vulnerable as soon as we can we'll put that out.
Our IRS sign ethylene within seven working day.
Fu Shun: Okay.
Fu Shun: Okay.
Speaker Change: Further questions. Please let us like hurricane no. Thank you so much bye bye.
Speaker Change: The conference has now concluded we thank you for attending today's presentation.
Speaker Change: You may now disconnect your lines and have a nice day.
[music].
Speaker Change: Okay.
Speaker Change: [music].
Tom McGinn: Okay.
Tom McGinn: [music].
Tom McGinn: Yes.
Tom McGinn: [music].