The following is the summary of the third quarter financial report conference call for $Alphabet(GOOGL.US)$Alphabet-C(GOOG.US) in 2024. For financial report analysis, please refer to "Google: Little Ghost Troubles? AI Solves a Thousand Worries" here .
I. Review of Core Financial Information:
II. Detailed Content of the Financial Report Conference Call
2.1. Key Points from Executive Statements:
1) AI
① Infrastructure: With a differentiated full-stack artificial intelligence innovation approach, it holds a unique leading position in the era of artificial intelligence.
a. Clean energy investments in the United States, Thailand, and Uruguay will provide 500 megawatts of round-the-clock clean energy electricity.
b. Through hardware, engineering, and technological breakthroughs, the machine cost per query is reduced by over 90%, and the customized Gemini model scale is doubled.
c. Providing customers with various accelerator options, including NVIDIA's GPU and our own TPU 6th generation.
② AI Research and Development: Driving Gemini model updates, including long-context understanding, multimodal, and Agents. From API call situations and adoption rates, the usage of the Gemini model is steadily increasing.
③ AI Product Applications: AR smart glasses Project Astra, planned for launch in 2025. Committed to using AI models on existing platforms and products, currently used in 7 products and platforms.
With over 2 billion monthly users, including the latest product Google Maps, which has over 2 billion users.
a. AI Overviews have been launched in over 100 new countries and regions, with monthly users exceeding 1 billion.
b. Circle to Search has been launched on over 150 million Android devices, and one-third of those who have used Circle to Search now use it every week
c. Lens is used for visual search over 20 billion times per month.
2) Search Advertising
This quarter, growth in financial services (insurance) and retail advertising has been strong.
AI brings two new advertising experiences to search.
One is the use of Lens nearly 20 billion times per month, with 1/4 being commercialized. In early October, we announced the product search function on Google Lens. During testing, we found that shoppers have high engagement with this new form.
Moreover, people are increasingly using Lens for complex multimodal searches. In addition to visual search, there is also voice search. Therefore, in early October, we launched shopping ads above or next to relevant visual search results to help business clients reach users.
Another change in the search experience is AI Overview. We have placed ad slots both above and below AI Overview responses. We have even found that ads can now be displayed directly in AI Overview responses, which is helpful to users. This helps businesses quickly connect with users.
For advertising clients, using our Gemini model to generate AI content can increase traffic to their official websites. For example, Audi used Gemini to generate video content, resulting in an 80% increase in traffic, a 2.7 times increase in clicks, and increased sales. Last week, we updated the Imagen3 model with various industry data adjustments to help clients generate high-quality images. Advertisers can now use Pmax and Deman Gen to create higher-performance marketing content and campaigns.
This quarter, we introduced the open-source marketing portfolio model Meridian to more clients to help expand the measurement range of cross-channel budgets.
2) Google Cloud
Google Cloud continues to show strong growth, with Q3 revenue of $11.4 billion, a 35% year-on-year increase, and an operating profit margin of 17%. Our leading technology and AI product portfolio are helping us attract new customers and drive a 30% increase in product adoption by customers.
The high growth in the third quarter reflects accelerated growth in GCP in AI infrastructure, generative AI solutions, and core GCP products. GCP's growth rate once again exceeded the overall growth rate of cloud computing. We also saw strong growth in Google Workspace, mainly due to the increase in average revenue per seat.
Currently, cloud customers primarily use our products in five ways:
a. AI infrastructure, using our TPUs and GPUs, LG AI Research has reduced the inference time of multimodal models by over 50% and reduced operating costs by 72%.
b. Enterprise AI platform Vertex, used to build and customize the best foundational models from Google and the industry. Gemini API calls have grown nearly 14 times in six months. For example, Snap chose Gemini's multimodal feature for their My AI chatbot
c. By combining AI platform with BigQuery, regardless of where multimodal data is stored, we can analyze it through Gemini with ultra-low latency access. The combination of artificial intelligence and data science has led to an almost 80% growth in BigQuery ML business in six months.
d. AI network security solutions, Google threat intelligence, and security operations are helping BBVA and Deloitte to prevent, deduct, and respond to cybersecurity threats faster. The adoption rate of Mandiant power threat detection has quadrupled in the past six quarters.
e. The launch of the all-new Customer Engagement Suite expands the application portfolio, improving customer experience in call centers and retail stores through the introduction of virtual assistants.
f. Additionally, we empower Google Workspace through Gemini, with 75% of users reporting improved work efficiency.
3) YouTube
The main growth driver in the third quarter was brand advertising, with increased user engagement on the platform.
Monetization of Shorts continues to increase, narrowing the gap with long-form video streaming, especially in the United States.
70% of videos uploaded monthly are Shorts, and currently advertisers can book the first exposure slot for Shorts short videos in nearly 40 regional markets.
In the living room consumption scenario, according to Nielsen data, we continue to maintain the top position in the U.S. streaming media market. This quarter, the number of revenue-generating creators on YouTube increased by 30% year-on-year, making YouTube the preferred platform for watching sports events.
The viewership of Olympic content on YouTube in the third quarter exceeded 12 billion views. Over 850 million unique viewers watched over 40 billion minutes of content on TV screens, with 35% of viewers watching this content. Recently, we launched the second season of NFL Sunday Ticket on YouTube TV, receiving positive feedback from advertisers, NFL partners, and fans.
In the past four quarters, YouTube's total advertising and subscription revenue exceeded $50 billion for the first time, with YouTube TV, NFL Sunday Ticket, and YouTube Music Premium driving subscription revenue growth on the platform.
Through Gemini, we have gained a deeper understanding of user preferences, significantly improving content recommendations on YouTube.
Google DeepMind's most powerful video generation VO model will be launched on YouTube Shorts by the end of this year.
4) Android Devices
Gemini has been integrated into the Android system, for example, Gemini Live allows free conversation between Gemini and users. Currently launched on Android systems such as Samsung, we will continue to work closely with manufacturers like Samsung to provide more innovative features At the same time, we released the Pixel 9 series in the third quarter, integrating large models such as Gemini Nano.
5) Innovation Projects
Let's focus on Waymo. Currently, Waymo's fully autonomous driving mileage exceeds 1 million miles per week, with over 150,000 paid services.
By expanding operations in partnership with Uber in Austin and Atlanta, and having a long-standing collaboration with Hyundai, we can bring autonomous driving to more areas and reach more users.
With its sixth-generation system, Waymo has significantly reduced unit costs without compromising safety.
Furthermore, drone delivery company Wing has been in partnership with Walmart for over a year, currently operating in 11 stores and serving 26 different towns.
6) Incremental Financial Indicators (Key Financial Performance Insights)
Cost Optimization: Further global office space optimization measures have been taken, with a growth of 11% in R&D expenses due to increased R&D personnel salaries and depreciation expenses, and a 5% increase in sales expenses mainly related to Gemini promotions.
Free Cash Flow: The reason for the year-on-year decline in the third quarter is mainly due to the deferral of tax payments from the second and third quarters of last year to the fourth quarter, resulting in a high base. Free cash flow reached $55.8 billion in the past 12 months.
7) Outlook
Advertising Revenue: The growth rate of advertising revenue will continue to slow down due to the high base effect from last year (Asia-Pacific retailers' placements).
Other Revenue: Google Other is affected by the early sales of hardware in the third quarter, and the growth rate in the fourth quarter is expected to slow down.
Capital Expenditure: $13 billion in the third quarter, with an outlook expecting it to be flat compared to the third quarter in the fourth quarter. In the third quarter, there was a plan to invest over $7 billion in data centers, with nearly $6 billion in the U.S.
Operating Expenses: Depreciation has increased, but hardware costs have decreased, partially offsetting the former increase.
2.2, Q&A Analyst Q&A
Q: What are the 1 to 2 most exciting products in the next 2 to 3 years that, once they scale, will truly bring sustained multi-year search growth? What are the key limiting factors for promoting these products among 2 to 3 billion people?
A: I mentioned "Google Lens" in my speech. Video searches are now close to 20 billion times per month, and we have introduced AI search to over 1 billion users, so I expect search products and Gemini to continue to grow significantly in 2025. We now have 1 billion users, and I don't see any limitations.
There is a latency issue due to inference costs with each query, but substantial progress has been made in the past 18 months. I believe search will continue to evolve in the next 12 months, and we are at the forefront of innovation.
Q: Can you further discuss the infrastructure advantages and capital expenditure efficiency brought by Google's own TPU? How does this affect your future capital expenditure compared to peers and other leading cloud service providers?
A: We have all the leading artificial intelligence accelerators, GPUs, TPUs, and CPUs, a comprehensive solution, and continue to increase investment.
On the GPU side, we have established a good partnership with NVIDIA. We are looking forward to the GB 200 and will be one of the first companies to offer the GB 200 on a large scale.
Regarding TPUs, we have already developed to the sixth generation and are confident in the future blueprint. Because of this, we can have higher efficiency compared to peers, not only improving efficiency internally at Google but also through some services provided by cloud computing.
I want to discuss Capex. In the third quarter, we invested 13 billion, mainly used for two areas. One is technical infrastructure, which accounts for the majority of the 13 billion, and the other is investments in general facilities, innovative projects, and other areas of the company.
Among them, the former server investment, including TPU and GPU inputs, followed by data centers and network equipment. In the third quarter, 60% of the technical infrastructure investment was used for servers, and 40% was used for data centers and network equipment.
Considering the situation for the next quarter and next year, the fourth-quarter Capex is roughly the same level as the third quarter, about 13 billion US dollars. In 2025, we expect further increases. More information will be provided in the fourth-quarter conference call, but the increase will not be as significant as in 2024 compared to 2023.
One more point I would like to add is that the pricing of Gemini Flash is more attractive. I believe this well illustrates the efficiency our architecture can generate. Therefore, we are also doing the same work for internal use. Although the functionality of search models continues to improve, we have been able to truly optimize their underlying architecture, which is where we see a significant increase in efficiency.
Q: In the process of Waymo expanding to more cities, in terms of consumer adoption of the product, and how do you view Waymo's market strategy, what are some important experiences? Looking at the trend of YouTube from the perspective of long videos and short videos. How do you view the monetization trend of YouTube?
A: Regarding Waymo, it is clear that whether in the Phoenix market or in San Francisco, this has been an exciting year. We have certainly expanded in scale, especially in terms of paid rights. From a safety, privacy, reliability perspective, consumers' love for this experience will definitely bring us surprises. So, I think all of these are positive aspects. Obviously, the product will continue to improve.
For us, we mainly focus on each city. As we develop, we are able to do more cities now, and at a faster pace. We are also building partnerships in a unique way, so we have partnered with Uber and expanded operations to Austin and Atlanta. We have more options, and we are researching Waymo driving modes with other network partners, fleet managers, and so on. So, this is an exciting time, but we still prioritize safety, testing various modes, which will help us plan for 2025 and beyond Regarding the issue you raised about YouTube short video consumption and monetization, perhaps we can look at it from the perspective of user viewing time. In fact, the number of Short videos played on the mobile and living room TV platforms on YouTube continues to grow.
The daily viewing volume of YouTube short videos exceeds 70 billion times. In terms of monetization, the monetization rate of short videos continues to grow healthily compared to the viewing rate. We will continue to work closely with advertisers as Shorts remains a platform with continuously growing traffic.
Currently, advertisers can book the first exposure slot on Shorts. In addition, Shorts has been integrated into the YouTube Select video marketing campaign. I am very satisfied with our progress at the moment.
Q: Considering the increasing adoption rate of native AI products and compared to the current 250 million active users seen by ChatGPT, are there any milestone developments in Gemini's adoption rate? Looking at Google's search lawsuit with the Justice Department, Apple's ISA contracts and some Android pre-installation contracts are likely to expire at some point in the future. Therefore, my question is, what plans do we have to regain the usage share of these search access points that may disappear? How do we gain a share of iOS queries if there are changes in Safari toolbar access points?
A: We provide Gemini services in various products, and now more than 1 billion people search and access it in our products. The momentum of user growth for Gemini itself is very strong. The number of our Application Programming Interfaces (APIs) has grown 14 times in the past six months. On the Android system, the integration progress of Gemini with Google Assistant is very smooth, and the user feedback is very positive. Therefore, we will continue to launch more products, including the next generation of large model versions that we are investing in.
Regarding legal lawsuits, it is obvious that I cannot discuss too much given the ongoing litigation. But what I want to say is, taking a step back, we have always, even as recognized by the courts, been successful because of our deep innovation, and we continue to do so. People choose us because they believe our products are the best, whether they are consumers or partners. Over the years, we have been working hard to ensure that our products are as user-friendly as possible on all platforms.
Therefore, I believe that all these methods and all the lessons learned over the years will lay a solid foundation for us. In response to these lawsuits, we plan to actively defend ourselves, and some of the early suggestions made by departments such as the Justice Department have a significant impact. I think their implementation may have unexpected consequences, especially for the dynamic tech industry and America's leadership position in this field.
Q: Regarding the issue of AI search (AI Overview). You mentioned that you have seen more and more queries or activities, have you seen any changes in user engagement for the 1 billion users who are already using the product? Then, in terms of monetization, is it possible for AI search (AI Overview) to generate additional revenue for information query behaviors that cannot be monetized in traditional searches? **
A: With our launch, we are clearly expanding our scale. We have just expanded to 100 new countries and regions, bringing our AI user base to 1 billion. Among the users we have already launched to, we can clearly see very high engagement. This is one of the highest-rated products we have launched in the search field, and it is increasing overall search usage, with people asking more complex and diverse questions and exploring a wider range of websites. What's particularly exciting is that as people learn to adapt to this new behavior, this growth will actually increase over time.
In terms of monetization, additional revenue streams, such as AR video search ads recently launched on our mobile platform in the United States, have been introduced. Previously, we also introduced ads above and below the AI Overview. Overall, we see that the monetization rate of artificial intelligence overview is roughly the same as traditional search, providing us with a very solid foundation for further innovation.
Q: Some even believe that Alphabet has not been innovative in artificial intelligence, which is clearly not true. In fact, you are developing rapidly and have launched many innovative products. Can you talk about how you will change the existing product structure? I know you have merged some assets, but can you discuss how you are rethinking how to bring some innovative products to market? With the development of artificial intelligence, what changes are you making in operations?
A: I absolutely think this is an extraordinary opportunity. I believe that at this critical moment, we must build models from scratch and achieve scalable production in our architecture, which is the essence of Gemini. Therefore, we need fixed costs to complete all the preparations and launch Gemini.
But now, I think we are in a virtuous cycle, with the rapid development of underlying models. We already have two generations of Gemini and are developing the third generation, making good progress. Now, the internal team is more complete, able to better leverage innovation in underlying models and translate it into product innovation. All seven products with 2 billion users have completed the integration of Gemini 1.0 and have set a positive development roadmap for 2025.
As I mentioned earlier about search, I think there is still a lot we can do. In addition, we are enabling smaller teams to provide updated experiences, with Notebook LM being the first example of such efforts.
Through all of this, the transformation from the company's development from PC to mobile devices is something we must do.
We are restructuring the company, in fact, if Google is seen as a neural network, we are forming new synapses that can better adapt to the current situation. I believe this lays a good foundation for our development in the coming year, and we will also bring all these innovations to the external world through cloud technology.
Q: Firstly, the profit margin of cloud computing has indeed started to show good growth, but the operating profit margins of some competitors in the industry are still much higher. I think this is a matter of scale and the goal of trying to catch up But how do you view the trend of profit margins you see? How confident are you that these profit margins can compete with other industry participants with profit margins close to 30%?
Secondly, overall, with an operating profit margin of 32% in the third quarter, from the perspective of new business development, is there still room for cost optimization in the future? In your opinion, what is the most likely driving force?
A: Let me first talk about the profit margin of cloud computing. Of course, I am pleased to see not only the growth in sales, but also the expansion of profit margins to 17%, which is indeed outstanding work done by our cloud computing team in driving continuous customer benefits. When you consider the expansion of profit margins, in fact, there are several aspects (you have mentioned one of them)—
First is scale. Obviously, as the business scales up, we have more opportunities to expand profit margins.
Secondly, leading teams can improve the efficiency of the entire cloud computing business, and we see these improvements achieved through personnel management, facility management, and other process efficiency enhancements, thereby affecting profit margins. But obviously, our performance cannot be directly compared with our peers, as this is a different business, but there will be many other things in the future.
It needs to be clear that AI cloud is an area that requires investment, such as servers, but it is also based on customer demand. Therefore, this will translate into revenue in the short term. This also means that investments will bring some depreciation costs, increased expenses that are not capitalized, and so on. Therefore, we will continue to improve business efficiency and strive to offset some of these impacts. That's my view on the dynamics of cloud computing.
For the company, by examining the entire organization, I would prioritize what to do to further improve efficiency. Leaders like Ruth and Sundar have done a lot of work in redesigning and optimizing costs, and I believe any organization can be further streamlined. As Sundar just mentioned, when we use AI in our processes and workflows, there will be efficiency improvements. And this is contrary to the significant impact of increased capital expenditures on profit margins. For the situation in 2025, I will disclose more details in the Q4 call. In short, we hope to improve efficiency to offset the expenditure growth brought about by investments.
Q: For search alone, why have two completely different search experiences, one being an engine agent similar to answers, and the other being a traditional link-based search engine? You can innovate in these two aspects and let consumers decide. This may be seen as the ultimate A/B test. So I'm very curious about your thoughts. Secondly, can you talk about the consumer environment in the fourth quarter? You should have a better understanding of consumer situations in multiple verticals. Thirdly, can you discuss the issue of slowing down ad spending by Asian e-commerce companies this year, as well as the trends during the election and holiday season?
A: Regarding the first question, people now use many things like answer engines. I would say that Google started using featured snippets to answer questions in our search products about 10 years ago. So, I think, ultimately you are serving users, and user expectations are constantly changing. So we have to work hard to stay ahead, predict, and stay ahead. That's why we have truly achieved native multimodality at both the input and output ends of search Therefore, we will continue to expand innovation.
I believe that having two search modes allows us to conduct more experiments. Now is the time for new use cases to emerge, enabling us to do things we couldn't do before. With flexibility and product services, we can act very quickly. I think this is very helpful. So, we are embracing it and will invest in this moment as we did in the past year, which I believe will bring great results for users.
Regarding trends in vertical industries, I mentioned that the growth in search and other revenues mainly comes from financial services, as the economic conditions in the insurance industry have improved, followed by the retail industry. However, overall, we see broad advantages in all verticals. Perhaps specifically in terms of advertising spending related to elections, we saw some momentum from election-related ad spending in the third quarter, which was more pronounced in YouTube ads.
Q: I think in two separate blog posts on Google Cloud, when discussing real-life use cases of Gen AI, you emphasized that many companies are starting to turn ideas into products. I believe that in six months, the growth rate of actual product use cases for Gen AI has reached 80%. I think you also released some survey data stating that your customers are generating tangible return on investment. So, in terms of the sales cycle, are you seeing acceleration? How much work does the cloud computing team need to do to help customers turn ideas into reality faster?
Philipp said, I think one feedback we received from advertisers is that although PMax's initial use case was in the search field, they are increasingly using it for mid-funnel and upper-funnel marketing activities. So, can you talk about whether this is a case or a trend you have seen across all advertising clients?
A: In terms of cloud computing, I think your question already implies this. Our customers rely heavily on us (for the underlying technology), so we do see a lot of data support for product user experience and return on investment.
Whether it's deploying internally within Google or bringing these experiences to external customers, we are constantly learning. As we see common patterns across the industries we serve, I believe we are summarizing these experiences. Therefore, I think as time goes on, companies will understand more and favor us more.
More customer choices will drive our large models to perform better. We are building more comprehensive solutions on this basis, so I believe we are well prepared for 2025, and I think the momentum in this area will continue.
In terms of PMax, we continue to see success with PMax. We have seen these success stories from large advertisers, ad agencies, small and medium-sized enterprises with different marketing objectives and vertical industries. Therefore, PMax is very cost-effective and can reach customers through all different Google channels wherever they are. With the launch of Gemini, we have added many new features to PMax, such as providing more powerful performance, helping advertisers scale, and building high-quality creative assets However, when answering your question about the funnel directly, please also remember that we have a great product called Demand Gen, which aims to inspire consumers to go beyond initial awareness and take action. We believe that Demand Gen is actually a very powerful tool for marketers to win in today's market, and we can't wait to see it bring more value.
Risk Disclosure and Statement of this article: Dolphin Research Disclaimer and General Disclosure
