Report Introduction
This article reviews the development logic of Amazon's retail and cloud businesses and finds a high degree of consistency between them. Therefore, we believe that the company's generative AI business is expected to rival advertising business and become a new growth point with high barriers and high profits.
Profit forecast and investment suggestions: Amazon, as a leading cloud computing company, will accelerate the realization of AIGC business. We expect the company's operating revenue to be 570.3/660.2/759.7 billion US dollars, an increase of 11%/16%/15%, and the GAAP net profit of 2023E-2025E to be 266/373/55 billion US dollars, an increase of 1078%/40%/48%, respectively. We use the segment valuation method and DCF valuation method to calculate the average and obtain a target price of $154.8 for 2024, with a growth potential of 28%, giving us a "overweight" rating.
The high barrier and high profit AIGC business will drive the profit level of cloud computing to increase in the long term. The market believes that Amazon's main source of profit is cloud computing, and the significant capital expenditure required by AIGC's business will continue to drag down cloud computing profit margins and affect overall valuation. We believe that the AIGC business will become the Profit Generator of Amazon Cloud, continuously driving profit margins. This is due to the following reasons: 1) the high level of standardization of the large model platform, which has cost dilution benefits; 2) AIGC business will form a higher quality subscription based business model; 3) The company provides huge room for price increases throughout the entire lifecycle of the model.
Core Information and Logic: Cloud computing will reinterpret the logic of retail development, incubating generative AI businesses with higher barriers and profits. Through a review, we found that the development logic of retail and cloud businesses is highly consistent, mainly reflected in: 1) the starting point of the business is low profit, heavy asset bitter business (self operated e-commerce&IaaS cloud), but the company is not eager to make short-term profits, but rather to seize the market at low prices and accumulate cash flow to create infrastructure barriers (logistics system&cloud data center); 2) After establishing scale, explore better business models to achieve profitability (platform e-commerce&PaaS cloud), while expanding the flow scale, cash flow is used to create barriers to low prices and richness; 3) After establishing a monopoly position, the company relies on its ecological resources and cash flow advantages to further create high profit and high barrier businesses (advertising&AIGC), achieving ecological improvement and long-term profitability. Based on the current AIGC business of the company, which also has cloud ecological resources and cash flow endowment, we are optimistic that it will become a new growth point for peer-to-peer advertising.
Catalyst: Acquisition of Antiopic; AWS joins the Meta model; Bedlock is officially available.
Risk reminder: AI application is not as expected; Geopolitical conflicts affect product sales; The research and development progress is not as expected.
Body
1
Investment advice
Core Data Forecast: We estimate that Amazon's operating revenue for 2023E-2025E will be 570.3/660.2/759.7 billion US dollars, with an increase of 11%/16%/15% respectively.
Core assumptions:
1) Retail business growth tends to flatten out;
2) The recovery of advertising demand from advertisers in the past 24 and 25 years has led to an increase in advertising revenue growth;
3) AWS cloud computing business will gradually benefit from AIGC monetization and release growth momentum;
4) Due to the continuous increase in the proportion of cloud business revenue and the prominent scale effect of IaaS cloud, the company's profit margin level will be improved.
Valuation 1: Considering that Amazon can continuously and stably generate profits and cash flows, we use the DCF valuation method to measure the company's equity value. Based on the following assumptions, we calculate that the company's target market value for 2024 is 1545.7 billion US dollars, corresponding to a reasonable valuation of 151.7 US dollars.
Core assumptions:
1) The US 30-year treasury bond bond yield is adopted as risk-free yield Rf=5.2%;
2) Assuming that the return required by investors exceeds 6.0% of the risk-free interest rate, the expected market return is 11.2%;
3) β= 1.19, based on core hypothesis 1 and core hypothesis 2, the weighted average cost of funds (WACC) is 12.3%;
4) The sustainable growth rate is 2.5%.
Valuation 2: Considering that Amazon's main businesses are core retail and cloud computing AWS, we also adopt the segment valuation SOTP method: 1) Core retail is expected to generate a net profit of 9.65 billion yuan in 2024, and Amazon, as a global retail leader, should enjoy a valuation premium. Therefore, a 24 year P/E of 26X is given, corresponding to a valuation of 250.8 billion dollars. 2) Cloud computing is expected to generate revenue of 123.1 billion. As we are optimistic that Amazon, as the global cloud computing leader with a market share of 40%, will accelerate the realization of generative artificial intelligence business monetization, we should enjoy a valuation premium of 24 years P/S 10X, corresponding to a valuation of 1230.8 billion US dollars. Finally, we calculated that the company's target market value for 2024 is $1608.4 billion, which corresponds to a reasonable valuation of 157.9 USD for American stock Amazon (AMZN. US).
Valuation Conclusion: By combining the two valuation methods, we can obtain an average of Amazon's target price of 154.8 USD and give it a "overweight" rating.
2
Core information and logic:
Cloud computing will reinterpret the development logic of retail for the second time,
Incubate generative AI businesses with higher barriers and profits
By reviewing the development logic of Amazon's two major growth curves of retail and cloud business, we found a high degree of consistency between the two:
1) Retail stage: Self operated e-commerce serves as the starting point of the business, and the company positions it as a Cash Generator. The rolling cash flow snowball is successively invested in logistics construction and Prime membership rights, strengthening the retail moat while extending the development of platform e-commerce models with better business models and higher profits (with a commission of 8-15%) to achieve profitability. After reaching a monopoly position in retail, the company then utilizes retail ecological resources (data+channels+traffic) and cash flow to incubate a higher quality advertising business as a profit generator for the retail business.
2) Cloud computing stage: The continuously increasing profits and cash flow in the retail stage are invested in the second growth curve cloud business, completing the construction of IaaS cloud data centers and market monopolies, and then incubating higher profit PaaS cloud services to enhance profitability. Based on the fact that the current generative AI business of the company is also built on a monopolistic cloud ecological resource (algorithm+computing power+data+community customer base&traffic) and huge cash flow, we are optimistic that this business will present a development logic similar to retail advertising, becoming a new growth point with better business models and higher barriers.
As shown in Figure 6, we summarize the similarities between retail and cloud business as follows:
1、 Business level
1) Starting from the low profit and heavy asset business: 1P self operated e-commerce corresponds to IaaS infrastructure cloud services;
2) High profit and light asset businesses achieve profitability: 3P platform e-commerce corresponds to PaaS platform cloud services;
3) High profit and high barrier businesses boost profitability: Advertising services correspond to generative AI clouds.
2、 Competitive advantage level
1) The construction of heavy assets forms a scale effect: the logistics center network corresponds to the IaaS cloud data center network;
2) Rich service categories enhance user stickiness: Prime membership benefits (streaming/cloud storage/shopping discounts/logistics services) correspond to the rich tool set of Paas layer (aPaaS/IPaaS/AIPaaS);
3) Long term low price leader: The low price of retail products corresponds to the low price of cloud services.
Introduction: We will review the development logic of the retail business (left side of Figure 6) in the third section and the cloud business (right side of Figure 6) in the fourth section, in order to present the consistency of the two stages and answer the core question of this article: "Why are we optimistic about the future development of the company's generative AI business.
3
The Development Logic of Retail Business from a Dual Perspective of History and Finance
The development logic of retail: As a starting point for its business, self operated e-commerce is positioned as a Cash Generator by the company, and rolling cash flow snowballs are successively invested in logistics construction and Prime membership rights, strengthening the retail moat while extending the development of platform e-commerce models with better business models and higher profits (with a commission of 8-15%) to achieve profitability. After reaching a monopoly position in retail, the company then utilizes retail ecological resources (data+channels+traffic) and cash flow to incubate a higher quality advertising business as a profit generator for the retail business.
It can be divided into three stages:
1) Starting from self operated e-commerce: The company attracts users at low prices, trades losses for growth, and creates this business as a core source of cash flow, utilizing cash flow to complete category/regional expansion and logistics construction.
2) Platform e-commerce attraction: In order to further expand the category and improve user experience, the company has launched Marketplace, Prime, and FBA services, which form a flywheel to quickly expand the traffic scale. The company takes advantage of the situation and collects "toll fees" (merchant commissions of about 8% -15%) to achieve the first cash flow of traffic.
3) Advertising platform monetization: With the accumulated retail data and ecological resources such as multiple investment channels (streaming media&e-commerce platforms) accumulated over the years of e-commerce development, the company has once again incubated high conversion and profit rate advertising business, further improving the overall retail profit and cash flow, and laying a cash flow foundation for subsequent investment and construction of cloud business.
3.1. Reviewing history and seeing how Amazon iterates its business model
1994-1999: Relying on a purely self operated model, it became the world's largest e-commerce company. In 1994, Besos chose books as a cold start product based on the high growth rate and dispersed pattern of the book market. The advantages of rich variety, low price, and high turnover led to its rapid expansion in scale. In 1997, the revenue increased by more than 8 times and it was listed in May of the same year. After going public, the company expanded its territory to Europe and expanded its categories to online music, gift toys, etc. In 1999, Wright, the former vice president of logistics of Wal Mart, began to build warehousing logistics. By 1999, Amazon had become the world's largest e-commerce company.
2000-2011: Incorporated third-party merchants to create platform flywheels. In November 2000, in response to eBay's competition, Amazon launched Marketplace (3P e-commerce platform) to support third-party merchants in sales competition with self operated businesses, enriching product offerings while lowering prices. In 2005, Prime membership services were released, allowing members to enjoy free next-day delivery, streaming, and other benefits. The company eliminated shipping barriers and increased user stickiness. In 2007, the company released FBA logistics services to assist third-party merchants in sorting, packaging, delivery, and after-sales work after placing orders, thereby saving costs for merchants and improving stickiness. The synergy between Marketplace, Prime, and FBA forms a flywheel effect, which accelerates the development of platform e-commerce.
2012-2019: Exploring advertising monetization and offline retail. In 2012, the company established Amazon Media Group, Amazon Marketing Services, and Amazon Advertising Platform, providing merchants with advertising tools such as ppc (pay per click) for the first time. In the following years, the company utilized on-site retail data and recommendation algorithms to achieve higher advertising conversion rates. In 2018, the company integrated the three major advertising platforms into Amazon Advertising, streamlined service processes and product names, and improved the experience of advertisers (advertising revenue data began to be disclosed in 2019). In 2015, the company opened its first physical bookstore for offline retail layout; In 2017, it acquired Whole Foods Supermarket with 450 stores for 13.7 billion US dollars to expand its grocery category, making it the largest acquisition in history; The following year, Amazon Go opened an unmanned cash register convenience store and implemented a strategy of optimizing the shopping experience through technology.
2020 present: Advertising has become a core profit point. Axios data shows that Amazon's advertising operating profit margin may far exceed 50%, indicating that it contributes as much operating profit as AWS; Based on the high profit margin of advertising, Amazon has started to consider it as the core profit point of retail. According to Marketplace Pulse data, if the retail business divests the advertising part from 2017, its operating profit will shift from positive to negative, and the advertising revenue CAGR between 2020 and 2022 will reach 38%.
3.2. Reviewing Finance to See How Amazon Practices Long-term Doctrine
3.2.1 Continuously optimizing business models to drive increased profits and cash flow
Self operated e-commerce: occupying upstream funds to achieve business expansion and logistics construction. From the perspective of net profit margin, the company has always been at a loss during the pure self operation stage, but has rapidly expanded to become the world's largest e-commerce platform with less financing, due to the accumulation of cash flow through negative cash conversion cycles:
From 1996 to 2001, Amazon's cash conversion cycle remained below -32 days, meaning that the company was able to make payments to suppliers at least 32 days after receiving user payments. During this period, the funds could be used for capital expenses, and cash flow would increase with the expansion of the business scale. Therefore, the company was occupying more and more upstream funds during expansion. In addition, the increasing market position has also extended the time for payment to suppliers, ultimately maintaining low prices and completing the construction of logistics systems.
Platform e-commerce: Build a flywheel to continuously attract traffic and monetize traffic to achieve profitability. In a letter to shareholders in 2008, Bezos mentioned that he believes the key factors that determine user experience are low price, category richness, and logistics delivery. The company has also created a flywheel model for continuous flow based on these factors:
1) Launch Marketplace to attract third-party merchants to settle in and compete with self operated businesses to lower prices;
2) Low prices have attracted users, driving an increase in the richness of merchants and categories;
3) Release Prime to lower the shipping threshold for users (paying a membership fee of $79 per year to enjoy unlimited free shipping), and launch FBA to provide third-party merchants with logistics services that have more cost and performance advantages compared to self shipping. This can improve user experience while gaining more platform traffic tilt, thereby attracting more merchants to adopt FBA and improving delivery efficiency;
4) The optimization of prices, category richness, and delivery efficiency has driven a continuous increase in supply and demand side traffic.
In the following text, we will use data to demonstrate this logic one by one.
In terms of monetization mode, the company charges membership fees on the user end, opening fees, transaction commissions, and FBA service fees on the merchant end, and achieved its first annual profit (EBITDA conversion) in 2002. In 2022, the proportion of third-party merchant revenue increased to 23%, becoming the second largest source of revenue.
Advertising: Utilizing the advantages of retail ecological resources such as channels and data to grow into new profit points. In terms of channels, the company utilized the cash flow accumulated by 1P+3P e-commerce and its original intention to improve user experience. In 2011, it started streaming media business through acquisition and included free services in Prime membership rights to attract more users. Currently, it has multiple channels for advertising, such as e-commerce platforms, streaming media, and third-party websites.
In terms of conversion rate, the company combined years of accumulated user consumption data with interest recommendation algorithms to achieve a higher transaction conversion rate compared to Google and Facebook, prompting its share in the online advertising market in the United States to rapidly increase from 2% in 2016 to 15% in 2023. Advertising has become the core profit point of retail business.
3.2.2 High capital expenditure Capex creates logistics scale barriers
Building an efficient logistics system by overlaying intelligent logistics technology with a global logistics network.
The advantage of dense distribution network is significant. Before 2010, the company's capital expenditure was mainly for logistics construction, and during the pandemic, it doubled the storage space in the United States. Therefore, in 2023, the company operated over 400 fulfillment centers, 150 sorting centers, and 1000 distribution stations worldwide, with an average single fulfillment center covering an area of approximately 800000 square feet and capable of employing over 1500 employees. Supply Chain Dive data shows that in 2021, 77% of Americans were within 60 minutes of driving from Amazon delivery stations.
Leading in intelligent logistics technology. On the other hand, the company has also continuously invested in intelligent logistics technology to accelerate the effective operation of the network, including the acquisition of Kiva System in 2012 to apply AGVs (unmanned transport vehicles). By 2020, there were 45000 AGVs; Launched drone service Prime Air in 2013; In 2017 and 2019, they respectively acquired unmanned delivery vehicle Dispatch and warehousing robot company Canvas Technology.
