Global revenue from principal operations was $148 billion, up 11% year-over-year (excluding the impact of currency translation).
North America segment revenue was $90 billion, up 9% year-over-year.
International segment revenue was $31.7 billion, up 10 percent year-over-year (excluding the impact of currency exchange rates).
AWS segment revenue was $26.3 billion, up 18.8 percent year-over-year (excluding the impact of foreign exchange rates).
Global operating profit was $14.7 billion, nearly doubling year-over-year.
North America segment operating profit was $5.1 billion, up $1.9 billion year-over-year.
Operating margin was 5.6 percent, up 170 basis points year-over-year and down 20 basis points sequentially.
International segment operating profit amounted to $0.3 billion, up $1.2 billion year-over-year.
Operating margin was 0.9 percent, up 390 basis points.
AWS segment operating profit was $9.3 billion, up $4.0 billion year-over-year.
Overall unit sales increased 11 percent year over year.
Net income was $13.5 billion, up 100 percent year-over-year.
Diluted earnings per share were $1.26. Advertising revenue was $12.8 billion, up 20% year over year.
While advertising revenue missed expectations, overall financial performance remained strong, particularly in net sales and net income, which improved significantly.AWS, a key growth engine for Amazon, continued to perform well in the second quarter, driving the company's overall revenue growth.
Total revenue for the first half of 2024 was $291.3 billion, up 11% year-over-year, while total expenses were $261.3 billion; net income was $23.9 billion, up 141% year-over-year.
For the third quarter of 2024, Amazon expects net sales to be in the range of $154 billion to $158.5 billion (up 8% to 11% year-over-year), and projected operating profit to be in the range of $11.5 billion to $15.0 billion, which is below the market's expectations of $15.66 billion.
With an updated market capitalization of $1.9 trillion, the Wind consensus estimates 2024 net profit of $49.6 billion, up 63% year-over-year, and a PE of 36x corresponding to $1.76 trillion.
Q: What's the common belief that it's possible to overinvest rather than underinvest in AI? Thinking about utilizing AI from a longer term perspective, how do you see the underlying investment potential of AWS? What is the view on semiconductor components in terms of investment velocity? What is the view on moving to more semiconductor components in medium to long term portfolios? What is their return on investment?
A: The first thing to understand about investment in AWS and AI is that there are significant logistical challenges to running an AWS business. The company faces about 35 regions, and a region can be considered a cluster of multiple data centers and about 110 availability zones. Thousands of SKUs need to be logged in at the right number for each of the 200 AWS services in each availability zone. if the final capacity is too small, it can lead to service disruption. As a result, most companies offer more capacity than they need. However, if too much capacity is actually provided, the economics can be terrible and the return on operating revenue can be less than optimal.
As evidenced by AWS's disclosed operating revenues and profits, the company has learned how to manage this wisely. By building complex long-term models and algorithmicity, the company is able to provide the right amount of capacity. The company has done the same thing when it comes to artificial intelligence. The company has likewise developed a great deal of experience and skill in building capacity signals and models, as well as getting a lot of signals from customers about what they need. While a lot of money has been invested in the AI space and infrastructure, this has also created more demand for the company and will be a very big business.
Regarding semiconductor components, the strategy and approach comes from 18 years of operating AWS. The company has had a very deep relationship with Intel in the general purpose CPU space since it started operating AWS, and it still does. When the company found a product with high value and high return, the spend didn't actually go down. Even though it costs less per unit, it pushes the company to do more inventing and building for its customers and frees up more time. When more is spent, the company will again want better value for money than it is getting now.
The company's customers wanted to improve price/performance, and to that end, the company developed Trainium (a training chip) and Inferentia (an inference chip). The Company is developing a second version of both of these chips that will have a very compelling price/performance ratio. The company's customized chips are being produced at breakneck speed. Like Graviton, this will pay off very well and will be another differentiator for AWS relative to other companies.
