Summary of Key Points
Recently, the financial reports of three tech giants—Google, Amazon, and Microsoft—showed stark contrasts: Google’s stock price fell despite an increase in AI-related capital expenditures due to a lack of clarity around the return on investment; Amazon’s stock price soared despite also increasing its AI investments, thanks to the CEO’s clear explanation of how the money would be generated; Microsoft, on the other hand, used strong financial data to demonstrate the robust demand for AI computing power and its profitability, dispelling concerns about the high costs associated with AI.
The underlying shift in market perception is that Wall Street’s evaluation criteria for AI investments have shifted from focusing on the amount of money being spent to whether there are clear orders, revenue streams, and payment timelines for each dollar invested.
1. Why Did Google Face Negative Reactions While Amazon’s Stock Rose? The Difference Lies in the Explanation of Return Paths
Google’s issue was that it only mentioned the expenses without providing a detailed plan for generating returns:
- Its cloud business grew by 82%, but this included one-time hardware revenue, and the company did not disclose the actual growth rate after excluding this factor;
- It plans to increase capital expenditures to $195-205 billion in 2026 (the second time this year), and for the first time, it reported negative free cash flow of $5.86 billion. The company even had to borrow $100 billion and issue additional shares to fund these investments. Investors were concerned about where the return on such a large investment would come from.
Amazon’s strategy was to turn these expenses into well-planned investments:
- Although it also increased its expenditures to $220 billion, resulting in negative free cash flow, CEO Jassy provided specific figures: servers would pay off in 3 years and have a useful lifespan of 5-6 years; AI contracts are typically signed for at least 5 years; data centers can last for 30 years, allowing for the replacement of 5-6 generations of servers without the need for repeated infrastructure investments. This provided investors with a clear plan for generating revenue, making them understand that the money was being invested with the expectation of long-term profitability.
2. Microsoft Played a Crucial Role in Shifting Market Sentiment: Proving That AI Computing Power Can Be Profitable
Microsoft’s financial report dispelled fears that AI is an unprofitable endeavor:
- Azure’s cloud business grew by 43% (with expectations for another 45% in the next quarter), and it has unsold orders worth $678 billion (an 84% increase);
- New data center capacity was quickly snapped up by customers, indicating strong demand;
- 90% of its cloud revenue comes from companies outside the AI model development sector, showing a solid customer base;
- It even reported positive free cash flow of $19.6 billion, meaning it earned more than it spent.
These figures convinced investors that AI computing power is not just a costly endeavor but has significant commercial potential. Microsoft’s stock price rose by 15.5% in one day, and its market value increased by $450 billion, providing reassurance to the market.
3. Amazon CEO Jassy’s Clear Explanation of Revenue Streams
During a conference call, Jassy broke down the company’s capital expenditures into two parts, making them understandable to investors:
- Data Centers: Construction takes 1-2 years, but they last for 30 years. Once operational, they generate revenue and eliminate the need for ongoing costs such as land, electricity, and construction;
- Server Equipment: Customer needs are identified before purchasing, ensuring that investments pay off in 3 years with a usable lifespan of 5-6 years, and contracts are typically signed for at least 5 years.
This approach transformed the perception of AI investments from a costly waste to a model that generates continuous profits.
4. Wall Street’s New Focus: From Concerns About Expenses to Evaluation of Return Paths
Previously, the market focused on the total amount invested in AI, viewing high expenditures as a risk. Now, it is more interested in whether each dollar is tied to specific orders, revenue, and profit generation, along with clear payment timelines:
- Google claimed its investments would be profitable but failed to provide detailed evidence;
- Amazon provided concrete figures, giving investors confidence in the return on its investment;
- Microsoft’s data demonstrated strong demand and profitability, leading to positive market reactions.
This shift influenced other regions as well: SK Hynix’ stock rose by nearly 30%, and Samsung’s stock increased by 27%. Stocks in AI computing power that were previously sold off have seen a surge in demand.
5. Future Challenges: Can the “Renting Out Computing Power” Model Remain Profitable?
Cloud companies currently profit by providing AI computing power, similar to renting out assets. However, there are two long-term concerns:
- Pressure on Costs: The cost of AI processing is decreasing rapidly (cheaper chips and higher efficiency), which could lower the profitability of cloud services in the future;
- Cultural Differences in the Chinese Market: Companies like Alibaba Cloud and Volcano Engine are also investing heavily in AI infrastructure. However, the customer structure in China (e.g., business contracts) differs from that in the U.S. Whether the same revenue generation models apply in China remains uncertain.
These factors will be crucial to monitor as we move forward in the development of AI computing power.
In Summary: The high costs associated with AI are not necessarily a problem if there is a clear plan for generating returns. Investors now favor investments with defined return paths, rather than blind spending. This change reflects the new criteria adopted by Wall Street.