Amazon today released its financial report for the first quarter of fiscal year 2024: net sales were US$143.313 billion, a year-on-year increase of 13%, excluding the impact of exchange rate changes, which was also a year-on-year increase of 13%; net profit was US$10.431 billion, a significant year-on-year increase of 229 %; diluted earnings per share were US$0.98, a significant increase compared with US$0.31 per diluted share in the same period last year.
Amazon's first-quarter revenue and diluted earnings per share both exceeded Wall Street analysts' expectations, but the company's outlook for second-quarter revenue fell short of expectations. After the earnings report was released, Amazon's stock price rose nearly 3% after hours.
Advertisement
After the financial report is released,Amazon CEO Andy Jassy, CFO Brian Olsavsky and head of investor relations Dave FildesAnswered investor questions.
The following is the transcript of the conference call:
JP Morgan analyst Doug Anmuth:Judging from past history, Amazon has been in a cycle of expanding investment – profit margin growth – expanding investment. The current situation is that the company's growth base, gross profit and operating profit are already very large. Management believes that Among the businesses of generative artificial intelligence, capital investment intensity, supermarkets, space Internet project Kuiper and medical care, which ones will have a greater impact on the company's profit margin in the future?
Brian Osavsky:Yes, we've been saying this before, companies are always swinging back and forth like a pendulum between profitability and investment. I think we're at a stage where we're doing both at the same time, so we're more inclined to talk about the specific investments that are being made and how that might impact the short-term growth outlook.
Advertisement
Over the past 18 months, Amazon's progress in operating income and free cash flow growth has been largely driven by improvements in our e-commerce business and lower service costs. We also talked about our efforts in regional distribution and our current incoming material storage and transportation services. Growth in advertising and Amazon cloud services has been very strong, with cloud service margins increasing 800 basis points sequentially from the fourth quarter, largely driven by the company's overall revenue expansion and cost structure optimization.
but,There will be a significant increase in capital expenditures, most of which will be used to support infrastructure, especially in generative artificial intelligence, and depreciation is definitely expected to increase in this part of the business..In the first quarter, our capital expenditures were US$14 billion, which is expected to be the lowest level of quarterly capital expenditures this year.As Andy said earlier, we're seeing strong signals from customers including longer service contracts, higher contract amounts, many of which incorporate generative AI elements, and those signals allow us to We are confident about expanding capital investment in this area. Andy also mentioned that we have been deeply involved in this field for 18 years, and our capital and resource investment plans are all long-term. Through careful planning to create business growth momentum for customers, the company's operating income, operating profit, free cash flow and investment return will also increase accordingly. In the near term, it is the growth in capital expenditures, including strong capital expenditure growth in the e-commerce business, most of which is used to modestly increase capital or operating capacity, as well as upgrades to the same-day delivery network and Amazon logistics fleet. However, to a large extent, investment is mainly used in cloud services.
Andy Jassy:I would like to briefly add and summarize. The short answer to this question is that we have the ability to do both. In the company's existing relatively large businesses, including e-commerce, advertising and cloud services, we have many growth opportunities, and the investments we have made are meaningful. At the same time, I don't think that the cost structure improvement work of the e-commerce business has been perfected. We still have incredibly huge growth opportunities in front of us. Although the company's cloud services revenue has reached $100 billion, more than 85% of global IT spending is still spent on local equipment.
If you believe that equation flips, then we still have a lot of growth opportunities, and that's not even counting opportunities in generative AI. Generative AI is probably the greatest growth potential that has existed in technology in a long time, say since the advent of cloud computing or the Internet. Cloud services involve a lot of work moving from on-premises to the cloud. Being able to continue to get value from it is why customers continue to optimize their infrastructure, and most of the generative artificial intelligence workloads that can completely change the user experience will be built and completed in the cloud, so We see huge opportunities and we are investing in a number of other early-stage areas as well. So overall, we have the ability to be profitable and invest at the same time.
Barclays analyst Ross Sandler:I have a few questions about the company's capital spending intensity. The CEO of Anthropic has mentioned that he thinks the cost of training the next generation model will be around $1 billion, I guess around the Cloud 4 high-end version, and then the next generation may cost up to $10 billion to train. Does management think the industry will do these trainings on Amazon Cloud? Do you think Olympus (Amazon’s large-scale language model) and some of the company’s self-developed projects can maintain the most advanced level in this field? How much impact will these training needs have on the acceleration of the company's cloud services revenue in the first quarter?
Andy Jassy:We've probably seen three macro trends that I think have had a certain impact on the performance of the company's cloud services at least in the last quarter. First, I think the work on cost optimization is mostly done. Our customers are smart and over the past few months have developed a deep understanding of how to run their infrastructure on the cloud. Customers are already starting to turn to new projects, some of what I would call infrastructure modernization at a macro level, and trying to create value through generative AI technologies. In terms of moving to the cloud, we started doing it before the epidemic. At that time, most companies were looking to migrate their work from local to the cloud to improve cost-effectiveness, accelerate innovation speed and development efficiency.
Then the epidemic broke out, people were in survival mode, and the economy was uncertain, so everyone tried to save costs as much as possible, and more customers invested less in cloud migration. But now they've resumed that process because it's the easiest way for many companies (to improve efficiency and reduce costs), so we're seeing very significant growth in this business. at the same time,We're also seeing very significant growth in how we use generative AI technology with Amazon's cloud services.I mentioned before that the company has billions of dollars in sales in artificial intelligence, and it's still relatively early days. We see some factors driving growth. First of all, there are a lot of companies that are still building their own models, including Anthropic, which you mentioned, which is the largest provider of basic model building services, and those that are building their own models every 12 to 18 months, or less. Companies that need to build new models. These models consume a lot of data and model units, and training is very important.
Many models are being built on the Amazon cloud, and I expect that to be more and more the case over time because of the performance, security and silicon advantages we have, including from Nvidia.Take Anthropic as an example. They are using our custom chip Trainium to train future models.I believe there will be more similar models running on Amazon Cloud in the future. What you may not realize sometimes is that while we are still in the stage where many companies are spending money to train models, once these models are put into production, they will lead to a significant increase in sales. Although not so many companies have put it into production yet, if it gets to this point, the situation will change a lot.
The cost of training a model is usually much higher than the cost of inference because training only needs to happen periodically, while inference needs to happen all the time. Many companies choose to do inference on Amazon Cloud. One of the main reasons is that we can provide rich services. Thousands of companies have built their applications on Amazon Cloud and on Bedrock. Bedrock has many large language models. For them to choose, it has a series of features that make it easier for customers to build high-quality, low-cost, low-latency production-grade generative artificial intelligence applications. Therefore, both training and inference needs are driving the growth of Amazon Cloud Services. Key factor.
In addition, because the company's customer models and these generative AI applications generate large amounts of sensitive assets and data, the security of these applications is particularly important. If you pay attention to the events that have occurred in the past year or two, you will find that not all cloud service providers have the same security record as Amazon. We have significant advantages in this area, so as companies begin to enter the stage of serious experimentation and deploying applications into production, people will want to run their generative AI products through the Amazon cloud.