Google conference call focus: AI investment returns, capital expenditures for next quarter and next year

Google's third-quarter performance was strong, especially the significantly better-than-expected cloud revenue, which greatly eased the market's "AI anxiety".

According to the financial report data, Alphabet, Google's parent company, saw a 15% year-on-year increase in total revenue in the third quarter, with EPS profit increasing by nearly 37%, far exceeding expectations. Among all businesses, cloud revenue growth was the most rapid:

Google Cloud revenue in the third quarter was $11.35 billion, a year-on-year increase of about 35%, surpassing analysts' expectations of $10.79 billion and a 28.8% year-on-year growth in the second quarter; Google Cloud operating profit was approximately $1.95 billion, a 632% year-on-year increase, exceeding analysts' expectation of $1.11 billion.

In terms of expenses, Google's capital expenditure in the third quarter was $13.06 billion, slightly higher than expected after a nearly 10% increase from the second quarter and then a 0.9% decrease.

Google CEO Sundar Pichai stated during the conference call that he is "very satisfied" with the development of the cloud business and introduced the company's progress in the field of AI, including more than 2 billion monthly active users across seven products and platforms accessing the Gemini model; over a quarter of the new code is generated by AI; AI Overview has been expanded to over 100 new countries and regions this week, with monthly active users reaching 1 billion.

Alphabet's stock price surged nearly 6% after hours.

Cloud Revenue Surges by 35% Google CEO: Very Satisfied

Google Cloud generated $11.4 billion in revenue in the third quarter, a 35% increase from last year, with an operating profit margin of 17%.

Google CEO Sundar Pichai stated during the conference call, "I am very pleased with our growth. This business has real momentum, and as customers embrace general AI, the overall opportunity is increasing."

Pichai revealed that 30% of existing customers are adopting AI products more deeply, primarily through five ways:

First, AI infrastructure. By using a combination of TPUs and GPUs, LG AI Research reduced the inference processing time of its multimodal model by over 50% and reduced operating costs by 72%.

Second, the enterprise AI platform—Vertex—used for building and customizing foundational models. Gemini API calls have increased nearly 14 times in the past six months.

Third, using the AI platform in conjunction with the data platform—BigQuery. This allows customers like Hiscox to make accurate, real-time decisions, reducing the time to quote complex risks from days to minutes. These customer achievements combining AI and data science have led to an 80% growth in BigQuery ML operations in the past six months

Fourth, AI-driven cybersecurity solutions - Google Threat Intelligence and Security Operations - are helping customers such as BBVA and Deloitte to prevent, detect, and respond to cybersecurity threats faster. Customers adopting our Mandiant-driven threat detection have increased fourfold over the past six quarters.

Fifth: In the third quarter, we expanded our application product portfolio by introducing new customer engagement suites. It aims to enhance customer experiences for online and mobile applications, as well as call centers, retail stores, and more. A good example is Volkswagen of America, which is using this technology to support its new myVW virtual assistant.

Pichai stated that currently, more than 2 billion monthly active users of Google's seven major products and platforms are integrating the Gemini model; AI is being used internally to improve coding processes, increase productivity and efficiency, with over a quarter of new code generated by AI; AI Overview has been expanded to over 100 new countries and regions this week, with monthly active users reaching 1 billion.

Furthermore, Pichai also revealed that Google's multimodal assistant project Astra will launch in 2025.

Q4 spending expected to remain flat, higher capital expenditures next year

The financial report released this time shows that Google's capital expenditure in the third quarter was $13.06 billion, higher than analysts' expectations of $12.88 billion, not continuing the surge seen in the second quarter. After nearly a 10% quarter-on-quarter growth in the second quarter, capital expenditures in the third quarter fell by over 0.9% quarter-on-quarter.

Google CFO Anat Ashkenazi stated that she expects spending in the fourth quarter to remain at the same level. Most of the funds are used for technical infrastructure, including servers and data center equipment to support cloud and AI products.

However, Ashkenazi mentioned that the company will have higher capital expenditures in 2025. This also aligns with Pichai's view, as he mentioned a "positive roadmap for 2025" regarding search and cloud.

Ashkenazi stated that these investments are based on customer demand, so they "will translate into revenue in a fairly short period of time."

Meanwhile, to adapt to changes in the competitive landscape and economic environment, Google has undergone reductions and internal restructuring. Ashkenazi mentioned that she and the leadership team will continue to cut costs across the company to "try to offset some of" the investments to maintain its competitive advantage and profit margins.

Below is the full transcript of Google CEO Sundar Pichai's remarks on the conference call, translated by AI:

Hello, everyone.

The third quarter was another outstanding quarter. The company's momentum is extraordinary, as you have seen in recent product launches and as you will hear in today's conference call. Our commitment to innovation, along with our long-term focus and investment in AI, is paying off and bringing success to the company and our customersOur unique positioning allows us to lead in the AI era because we have taken a differentiated full-stack approach to AI innovation, and we are now seeing this approach operating at scale. It consists of three components:

First, a powerful AI infrastructure, including data centers, chips, and a global fiber optic network.

Second, a world-class research team that is advancing our deep, technical AI research and also building models that drive our efforts.

Third, extensive global coverage through products and platforms that reach billions of people and customers worldwide, creating a virtuous cycle.

Let me quickly introduce these contents.

Full-stack AI innovation approach

We continue to invest in state-of-the-art infrastructure to support our AI efforts, from the United States to Thailand, to Uruguay. We are also making bold investments in clean energy, including the world's first corporate agreement to purchase nuclear energy from multiple small modular reactors, enabling up to five hundred megawatts of new 24/7 carbon-free power.

We are also doing important work inside data centers to improve efficiency while making significant improvements in hardware and models.

For example, we shared that since we started testing AI previews, we have significantly reduced the machine cost per query. In eighteen months, we have reduced the cost of these queries by over 90% through hardware, engineering, and technological breakthroughs, while doubling the size of our custom Gemini model.

Of course, we use—and offer to our customers—a range of AI accelerator options, including multiple categories of NVIDIA GPUs and our own custom TPUs. We are now on the sixth generation TPU—called Trillium—and continue to drive efficiency and better performance.

Turning to research, our Google DeepMind team continues to advance our leadership position.

Let me take a moment to congratulate Demis Hassabis and John Jumper for winning the Nobel Prize in Chemistry for their work on AlphaFold. This is an extraordinary achievement that highlights the incredible talent we have and the importance of our world-leading research to the modern AI revolution and our future progress. Also, congratulations to Geoff Hinton, who has been working here for over a decade, for winning the Nobel Prize in Physics.

Our research team is also advancing the capabilities of our industry-leading Gemini model, including long-context understanding, multimodal, and agent capabilities. By any measure—token volume, API calls, consumer usage, business adoption—the use of the Gemini model is experiencing rapid growth.

Our team is actively researching performance improvements and new features to adapt to our range of models. Stay tuned!

They are also building an AI experience that can see and reason about the world around it. Project Astra is a glimpse of that future. We are working hard to launch such an experience by 2025

And then we bring these advancements to consumers and businesses: today, all seven products and platforms with over 2 billion monthly active users are using the Gemini model. This includes the latest product to surpass the 2 billion user milestone, Google Maps. In addition to Google's own platforms, in response to strong demand, we are making Gemini more widely available to developers. Today, we shared that Gemini can now be used on GitHub Copilot, with more to come.

To support our investments in these three pillars, we are organizing the company to operate with speed and agility.

We recently moved the Gemini Applied Team to Google DeepMind to accelerate the deployment of new models and simplify post-training work. This is a continuation of other structural changes that have unified our research, machine learning infrastructure, and our developer teams, as well as our security efforts and our platform and device teams. All of this is helping us to act faster. For example, a small dedicated team built Notebook LM, a very popular and promising product.

We are also using AI internally to improve our coding processes, which enhances productivity and efficiency. Today, over a quarter of Google's new code is generated by AI and then reviewed and accepted by engineers. This helps our engineers do more and act faster.

I am excited about our progress and optimistic about future opportunities. We remain focused on building excellent products.

Search Progress

In search, the latest advancements including new features like AI Overview, Circle to Search, and Lens are changing the user experience, expanding what people can search for and how they search.

This has led to users coming to search more frequently to meet their information needs, driving additional search queries.

Just this week, AI Overview began rolling out to over 100 new countries and regions. It will now reach over 1 billion users monthly.

We see strong engagement, which is increasing overall search usage and user satisfaction. People are asking longer, more complex questions and exploring a wider range of websites. Particularly exciting is that as people learn that Google can answer more questions, this growth actually increases over time.

Ad integration within AI Overview is also performing well, helping people connect with businesses while searching.

Circle to Search is now available on over 150 million Android devices, with people using it for shopping, translating text, and learning more about the world around them. One-third of those who have tried Circle to Search now use it weekly, demonstrating its usefulness and potential.

Meanwhile, Lens is now used for over 20 billion visual searches monthly. Lens is one of the fastest-growing query types we see in search, as it can answer complex multimodal questions and help with product discovery and shoppingGoogle Cloud Growth

For all these AI capabilities, this is just the beginning, and you will see rapid innovation and progress here.

Next, Google Cloud.

I am very pleased with our growth. This business has real momentum, and as customers embrace general AI, overall opportunities are increasing.

We generated $11.4 billion in revenue in the third quarter, a 35% increase from last year, with an operating profit margin of 17%.

Our technology leadership and AI product portfolio are helping us attract new customers, win larger deals, and drive 30% deeper product adoption by existing customers.

Customers use our products in five different ways.

First, our AI infrastructure, which we differentiate through storage, computing and software advancements, as well as leading reliability and a leading number of accelerators. By using a combination of our TPUs and GPUs, LG AI Research has reduced the inference processing time of its multimodal models by over 50% and reduced operating costs by 72%.

Second, our enterprise AI platform - Vertex - is used to build and customize the best base models from Google and the industry. Gemini API calls have grown nearly 14 times in six months.

When Snap wanted to provide more innovative experiences for its "My AI" chatbot, they chose Gemini's powerful multimodal capabilities. Since then, Snap's interactions with My AI in the U.S. have increased by more than 2.5 times.

Third, customers use our AI platform with our data platform - BigQuery - to analyze multimodal data, accessing Gemini with ultra-low latency regardless of where it is stored. This allows customers like Hiscox, one of the flagship groups of Legal & General, to make accurate, real-time decisions, shortening the time to quote complex risks from days to minutes. These customer outcomes combining AI and data science have led to an 80% growth in BigQuery ML operations in six months.

Fourth, our AI-driven cybersecurity solutions - Google Threat Intelligence and Security Operations - are helping customers such as BBVA and Deloitte to prevent, detect, and respond to cybersecurity threats faster. We have seen a 4x increase in customers adopting our Mandiant-driven threat detection over the past six quarters.

Fifth: In the third quarter, we expanded our application product portfolio by introducing new customer engagement suites. It is designed to improve customer experiences in online and mobile applications, as well as call centers, retail stores, and more. A good example is Volkswagen of America, which is using this technology to support its new myVW virtual assistant.

Additionally, we have received excellent feedback for the employee agent provided by Gemini for Google Workspace. 75% of daily users say it has improved their work qualityTurning to YouTube: YouTube's advertising and subscription revenue exceeded $50 billion for the first time in the past four quarters.

YouTube TV, NFL Sunday Ticket, and YouTube Music Premium collectively drove platform subscription growth. We are leaning towards the living room experience, offering multi-view options, and introducing new options for creators to organize content into series and seasons, similar to traditional television.

On the YouTube creation side, we announced that Google DeepMind's most powerful video generation model, Veo, will come to YouTube Shorts later this year to assist creators.

Platforms and Devices

Next, platforms and devices. Gemini's deep integration is enhancing Android. For example, Gemini Live allows you to have smooth conversations with Gemini; people love it. It is available on Android, including Samsung Galaxy devices. We continue to work closely with them to provide innovation for their latest devices, with more to come.

In Google's manufacturing, we launched our latest Pixel 9 series devices, including advanced AI models such as Gemini Nano. We have seen strong demand for these devices, which have already received multiple awards.

Other Investments

Turning to other investments, I want to emphasize Waymo, the largest part of our product portfolio.

Waymo is now the clear technology leader in the autonomous driving industry and is creating increasingly growing business opportunities.

For years, Waymo has been integrating cutting-edge AI into its work. Now, Waymo drives over 1 million fully autonomous miles per week and provides over 150,000 paid rides - the first AV company to achieve this mainstream usage.

Through its expanded network and operational partnerships with Uber in Austin and Atlanta, along with a new multi-year partnership with Hyundai, Waymo is bringing fully autonomous driving to more people and places.

By developing a universal driver, Waymo has multiple market pathways. With its sixth-generation system, Waymo has significantly reduced unit costs without compromising safety.

Before I conclude, I am pleased to welcome our new CFO, Anat. We are excited to have her join us.

And as always, I want to express my gratitude to my global employees. Your dedication and hard work have once again led Alphabet to achieve outstanding quarterly performance