Amazon: Cloud, AI, and Increased Efficiency and Reduced Costs are Three Main Themes.

This season's corresponding financial report review can be found in "Amazon: Retail with soaring profits, but can't lift the faltering "AWS"?".

1. Management Statement:

1. Overall, we are satisfied with the growth of our global store business, including the increase in revenue from the international department, which has benefited from the easing of macroeconomic pressures in Europe.

2. Customers appreciate our sensitivity to value-for-money and convenience. Nevertheless, the uncertain economic environment and continued signs of inflation are still present. We have seen a slowdown in optional consumer spending, a shift towards low-priced goods, and a demand for daily necessities.

3. Third-party sellers, including businesses that use Amazon logistics for storage and transportation services, are one of the primary contributors to the choices we offer to buyers. 3P Sellers accounted for 59% of total sales in the first quarter, up from 55% a year ago.

4. The strong participation in advertising services saw a year-on-year revenue growth of 23%, excluding exchange rates. In particular, product promotions and brand products are still key drivers of growth.

5. Regarding AWS, given the continued economic uncertainty, customers continue to evaluate methods to optimize cloud expenses to cope with the difficult economic conditions in the first quarter. We expect these optimizations to continue into the second quarter, with revenue growth in April about 500 basis points below what we saw in the first quarter. Note that we will not attempt to optimize for any single quarter or year.

6. In the first quarter, our reported operating profit was $4.8 billion, which was negatively affected by estimated employee layoff costs of about $470 million in the first quarter, including $270 million related to AWS. We have decided to cut 9,000 jobs in the AWS business, as well as Twitch, advertising, and human resources teams.

7. Inflationary pressures continue to ease on a year-over-year basis, primarily due to lower long-haul shipping rates and lower diesel and electricity costs.

8. Overall net income for the first quarter was $3.2 billion. This net income includes a pre-tax valuation loss of $467 million included in non-operating expenses - from our common stock investment in Rivian Automotive.

9. Regarding cash flow, as of March 31, our operating cash flow for the past 12 months increased by $54.3 billion year-on-year, a 38% increase. In addition to the cash benefits brought by the increasing profitability, we have also seen some relaxation in the supply chain and made progress in improving inventory procurement and payment cycles, which in turn has had a positive effect on working capital.

10. In terms of capital investment, we define capital investment as a combination of capital expenditures and equipment and financing leases. For the full year 2023, we expect capital investment in 2022 to be less than $59 billion, mainly due to an expected year-on-year decrease in network investment. We will continue to invest in infrastructure to support the demands of AWS customers, including investments in large language models and generative AI.

In terms of our warehousing business, we are very focused on reducing service costs in the fulfillment network. Due to the unexpected surge in demand during the pandemic, the scale of our fulfillment center has doubled, and over the past few months, we have redesigned dozens of processes to improve productivity. We have also rebuilt our layout methods and expanded to a larger fulfillment center, transforming our nationwide fulfillment network into a regional one.

We have created eight interconnected regions that operate to a large extent on a self-sufficient basis, but can still ship nationwide when necessary. We are very excited about the progress we have made in providing more one-day and same-day delivery services to our customers, and we expect to achieve the fastest Prime delivery speed ever in 2023.

In terms of advertising, we will continue to grow strong despite economic difficulties. First, most people are still shopping even in tough economic times, making e-commerce advertising particularly beneficial for brands. However, we are still in the early stages and we will place our ads carefully in a wider range of video, sports broadcasting, audio, and retail assets, while AWS still has a big advantage in advertising.

One important attribute of the cloud is its seamless ability to scale up or down according to demand, which local infrastructure cannot do. More than 90% of global IT spending is still local and will continue to move to the cloud. The amount of new cloud business that will be generated by upcoming machine learning is currently incalculable.

Q1: Can you talk about the level of optimization achieved in AWS and how it compares to what you think may happen in the future? When did you start making these efforts?

A1: When we talk to customers, they tend to be cautious about the economy and are looking for ways to save money, like most companies, including our own. We don't want to optimize for one quarter or one year, but to work with our customers to find sensible ways to reduce costs and be able to scale up and down.

Customers are not simply cutting costs, but rather re-evaluating business priorities and reallocating resources to use IT resources more efficiently.

Q2: Regarding capital expenditures, you said overall capital expenditures will decline in 2023. Can you help us better understand the contribution between retail and AWS? From the perspective of generative AI and large language models, what is needed in terms of capital expenditures?

A2: We spent $59 billion in capital expenditures last year, and in fact, our annual expenditures are decreasing in our core fulfillment and transportation areas, which are estimating decreases and increasing gaps with AWS and infrastructure. We are investing more money in areas such as large language models and generative AI. Therefore, we have created some space in our fulfillment and transportation and are redirecting it towards AWS. We still believe that combined capital expenditures will decline year by year. Q3: What is the progress on raising the efficiency level of the company and restoring the pre-epidemic profit margin? How to strike a balance between profitability and driving growth plans for the next few years?

A3: We have carefully reviewed all businesses in the past 6-9 months and made adjustments. But our road ahead is especially long, especially in terms of operations. Looking at operating profit margins, the North American department was 1.2% this quarter, which was roughly in the range of 4%-6% before the epidemic. This marks how much room for improvement there is.

I am very optimistic that we not only have the opportunity to return to pre-pandemic levels in terms of operating profit margins, but some of the opportunities we have identified also have additional room for growth.

Q4: Can you introduce us to 2 or 3 key differentiation points that you think AWS provides in terms of the potential of large-scale language models and AI tools compared to some competitors?

A4: When you think of machine learning, Amazon has invested quite a bit in machine learning for over 25 years. It is almost rooted in everything we do. It is obviously present in Alexa, then AWS, where we have over 25 machine learning services, and we have the widest range of machine learning capabilities and quite a large customer base.

But frankly, AI models were not so eye-catching about 6-9 months ago, and they have become bigger, better, and faster, which really provides an excellent opportunity to change almost all customer experiences, as well as previously non-existent customer experiences. But this field is still in its very early stages.

The whole company has a very large investment in AI, which can be divided into three areas. Such as the key to computing is chips. So far, dedicated GPUs are expensive and scarce. In AWS, we have been committed to building custom machine learning chips. We built a chip specifically for training, called tranium, and a chip specifically for inference or predictions from inference models. The reality is that most people spend most of their time and money on training.

Therefore, we believe that a lot of machine learning and transactional reasoning will run on AWS. You have to train the model and then build the model. Truly leading large language models take many years to build and billions of dollars to build, and only a few companies can make such investments.

While most other merchants will use one of the basic models in AWS, then customize it for their proprietary data, needs, and customer experiences, and they hope to customize it in a way that will not reveal their unique IP.

People can run the basic model from Amazon (which we call Titan) or they can run it from leading large language model providers such as AI 21 and Anthropic and Ability AI. Customize it according to your own purposes, and then run it with the same security, privacy, and all the functionality used for the rest of the applications in AWS. The third layer is actually building applications on top of these large language models. So ChatGPT is a good example of an application that is being built. We're building some of these applications ourselves. So we think one of the most compelling applications that will be built in generative AI is about making developers more effective with coding assistance. So we built something called CodeWhisperer, which we just announced the general availability of, and developers can input natural language like "I want to build a video hosting site" and CodeWhisperer will produce the code you need, and developers can use it and put it into production.

Q5: What are your thoughts on the key investment focus of Echo and Alexa for the future? How do you view the return on investment around that division?

A5: Every one of our businesses internally is built on top of the large language models to reshape our customer experience. You'll see it in every one of our businesses, stores, ads, devices, entertainment, and devices, and your specific question is a good example of that.

I think when people often ask us about Alexa, what we often share is that it would be a relatively small investment if we were just making a smart speaker. But we have a vision, and we firmly believe that we want to build the world's best personal assistant. That's a hard thing to do. It spans a lot of areas, it's a very broad surface area. However, when you think about the emergence of large language models in generative AI, it does make the underlying models more effective, and I think it really accelerates the possibility of building the world's best personal assistant.

I think we start with the PR around Alexa because we have hundreds of millions of endpoints for entertainment, shopping, smart home and information, and a lot of engagement from third-party ecosystems partners. We have a large language model, but we are building a bigger, more general and more capable language model. I think this will really accelerate our vision of becoming the world's best personal assistant. I think there's an important business model behind it.

Q6: Can you reveal more about the performance of the international market profitability?

A6: We saw acceleration in growth on a constant currency basis to 9%, and then Q4 was 5%. The overseas economies are starting to stabilize. Particularly in the mature countries in Europe, we see confidence growing and inflation receding.

This has been an advantage that wasn't expected in Q1 in the international markets. There's been some narrowing of losses on the profit side, some of it from the benefit of revenue growth, but some of it attributable to some spending cuts that we made in some investments.

Most of that is in North America, but operations efficiency in global projects is also improving, which is lowering international costs. Again, just a reminder, the international market is a collection of profitable mature countries, and while these markets may seem mature, they're actually early stage. We have added over 10 countries in the past 5 years. In the North American market, it took us 9 years to achieve breakeven. We're seeing similar curves in many international countries. In fact, we often face additional challenges, such as the lack of payment methods and established infrastructure - especially transportation and the internet. Therefore, application practices may be slower, but we feel good about the businesses being established. They have many of the same features that North America possesses. Price and convenience are at their core, and we are very pleased with the promotion, traffic, and new customer acquisitions of Prime Video in many emerging countries. We will continue to focus on optimizing cost structures and developing these businesses on a country-by-country basis.

Regarding food and groceries, we have an interesting grocery business that has been around for a while and has grown considerably. This premium food and grocery business, like 25 or 30 years ago when mass purchase buyers entered the food and grocery industry, selects temperature-uncontrolled goods such as canned goods, packaged food, paper products, pet products, personal care products, health products, beauty products, and various consumables.

To offer as much food and groceries as possible, large quantities of goods must be supplied. This is what we have been committed to for the past few years, under the brand of Amazon Fresh. We hope to go further on this point, but the form of further expansion has not yet been determined. We have a series of experiments, ideas, and concepts being tested in dozens of stores.

Q8: Can you point out any unusual projects related to competition for AWS in the April or second-quarter outlook?

A8: Currently, workload reductions are taking place among customers on AWS, who are extending contracts when they reach the contract limit, and planning for the future. Therefore, we believe that the business outlook is good and can understand the short-term work we have done to help customers save money. Therefore, there is no significant year-on-year difference between the second quarter and the first quarter, except that we have once again learned which customers are increasing or decreasing in some areas and helping them continue their new plans.

Q9: Why did you choose healthcare and satellite internet as the major investment areas in the shareholder letter? Have you considered breaking down all large investments so that we can have a clearer understanding of the retail profit margin structure?

A9: In terms of healthcare, when considering whether we should make large investments and asking ourselves a series of questions, the healthcare industry is a trillion-dollar industry. Customers have been requesting a pharmacy for many years, and this is a natural extension of the retail business. We launched Amazon Pharmacy in 2020, which is a good start.

Many customers want a broader healthcare experience. We found that the digital applications of medical institutions are very attractive. You can chat or have video conferences with doctors, or if you have to go to a physical institution, they have clinics across the country and you can make an appointment for the same day or the next day. And in all of these cities, when you need medication, you can have Amazon Pharmacy or other third-party drugstores deliver them automatically.

I just want to briefly talk about Kiper. Today, billions of households, businesses, and government entities are unable to connect to the internet. For businesses and governments to be able to have this kind of coverage so that they can operate more seamlessly in all the environments where you must be present will fundamentally change the game. So we believe it has the potential to become a very large enterprise. We have already released some information about it in some of our designs. What excites us is that many customers in all these different market segments are excited about it. So I just chose these two examples of inventions that have emerged in the investment process, which we believe are very important for the company.

Disclosure: Dolphin Investment Research Disclaimer and General Disclosure