AI Investments of Hundreds of Billions of Dollars Finally Bear Fruit! Amazon and Meta's Revenue and Cloud Growth Begin to Materialize

In recent years, tech giants have invested hundreds of billions of dollars in expanding data centers, procuring AI chips, and deploying artificial intelligence infrastructure. This massive capital expenditure (capex) has become one of the most controversial investment topics in financial markets. However, Evercore analyst Mark Mahaney believes that recent earnings reports from some large internet companies are beginning to prove that these massive investments are not just 'burning cash,' but are gradually generating tangible returns—particularly Meta Platforms (META-US) and Amazon.com (AMZN-US), which stand out.

The scale of AI-related capital spending by major tech firms often reaches hundreds of billions, even exceeding a trillion dollars. Combined with concerns over inflated valuations from the AI boom, stocks like Meta, Amazon, Alphabet (GOOGL-US), Microsoft (MSFT-US), and Oracle (ORCL-US) have shown divergent performance recently. Since 2026, the S&P 500 index has risen 13%, yet Oracle, Meta, and Microsoft shares have declined, reflecting investor skepticism about the scale of AI investments and future return on investment.

Mahaney, however, argues that Q2 earnings indicate that some high-quality internet stocks have passed a critical turning point in AI investment returns. He suggests the market is entering a recovery phase for high-quality internet stocks, driven by strong 'return on AI' (ROAI) and successful product cycles evident in Q2 results.

Meta has drawn the most market attention. After reporting Q2 earnings in late July, Meta’s stock briefly plunged 8%, as investors remained uneasy about the company’s further increase in capital spending. However, the stock gradually rebounded, indicating that the market is beginning to reassess the long-term returns from Meta’s large-scale AI investments.

Meta has raised its 2026 capital expenditure forecast by 1.9%, to $137.5 billion. While this figure is nearly double that of 2025, the market had already anticipated Meta’s significant increase in AI-related investments this year, primarily for building data centers, procuring AI chips, and expanding AI infrastructure.

More importantly, according to FactSet forecasts, Meta’s annualized capex growth rate over the two years following 2026 is expected to slow to 22%. In other words, although Meta will continue to invest heavily, the pace of capex growth may gradually decline, which will help improve the company’s future profit margins.

At the same time, Meta’s Q2 results are already showing the business benefits of AI investments. The company’s Q2 revenue increased 28% year-over-year to $60.8 billion, driven primarily by higher ad prices and increased ad impressions.

Meta is using AI to analyze user behavior and deliver more relevant ads to different users, thereby increasing user engagement and improving advertisers’ return on investment. For Meta, this means AI is not just a cost center, but a tool that directly enhances the efficiency of its core advertising business.

WhatsApp is also emerging as another AI and monetization opportunity. Meta’s WhatsApp is still in the early stages of using AI to increase user commercial value, while paid messaging and subscription revenue have driven a 73% surge in 'other revenue'.

Meta CEO Mark Zuckerberg also revealed during the earnings call that the company is receiving numerous offers to sell computing capacity at a clear premium, and more coding and productivity tools are on the product roadmap.

This implies that Meta may not need to rely solely on its core advertising business to recoup AI investments. If the computing power generated by its AI infrastructure can be sold externally, or if AI tools can create new software and service revenues, Meta could further increase the utilization and return on its AI assets.

Mahaney believes the real question for investors is whether Meta can use AI to create more non-core products, services, and monetization opportunities. He considers this possibility highly plausible and more likely than current stock valuations suggest.

In terms of valuation, Meta’s forward P/E ratio is currently around 17x, a relatively low level since the full-scale AI era began in 2023. In contrast, the S&P 500’s forward P/E is slightly above 20x.

Meta has traditionally enjoyed a valuation premium over the broader market due to expectations of faster earnings growth. Therefore, if AI investments begin driving sustained revenue and profit growth, Meta’s relatively low current valuation could serve as a foundation for stock re-rating.

Amazon is also viewed by Mahaney as a representative company where AI capex may be starting to yield returns. He describes Amazon as being at a 'fundamental inflection point'.

Amazon has raised its 2026 capex forecast by 10%, to $220 billion—an astonishing scale. However, CEO Andy Jassy emphasized that a key reason for this increase is rising memory costs, not a sudden massive expansion in data center investments.

Since memory prices are highly volatile, this cost increase does not necessarily mean Amazon’s future capex will permanently remain at a higher level. Mahaney therefore believes the market should not simply interpret the $220 billion figure as Amazon suddenly making excessive investments.

More importantly, Amazon’s Q2 EPS exceeded market expectations, driven by higher-than-expected sales and profit margins. Even excluding the one-time gain from its investment in Anthropic, core performance remained strong.

The most direct beneficiary of Amazon’s AI investment is Amazon Web Services (AWS). This cloud computing business saw a clear acceleration in revenue growth in Q2, with a year-over-year increase of 37%, far exceeding the 17% growth rate from the same period last year.

AWS’s accelerated growth is crucial for Amazon, as it represents one of the primary business returns from the company’s massive investments in data centers, AI chips, and cloud infrastructure. If AWS can maintain high-speed growth, Amazon’s massive capex can be gradually recouped through higher cloud revenue and profits.

Beyond Meta and Amazon, Mahaney also highlights companies like Airbnb (ABNB-US) and Shopify (SHOP-US), noting they continue to maintain strong revenue growth, making their AI investment returns appear more attractive.

Shopify, for instance, reported a 32% year-over-year increase in Q2 GMV (Gross Merchandise Volume), reaching $115.6 billion. GMV represents the total value of goods transacted through Shopify’s platform, before Shopify’s revenue share is deducted.

Strong GMV growth has also helped Shopify improve its free cash flow margin, while the company’s capital expenditures have not significantly increased. This means Shopify can continue expanding transaction volume and cash flow without major capex expansion—contrasting sharply with AI giants that require massive data center investments.

The debate over whether AI capex is excessive will not end soon. Large tech companies are investing unprecedented amounts, and the market fears AI investments could lead to overbuilding or even repeat past tech bubble-style capital waste.

However, some companies’ earnings are now sending different signals. Meta is improving ad efficiency and monetizing WhatsApp through AI, while Amazon is accelerating growth via AWS—indicating that AI infrastructure investment is gradually transforming from a pure cost item into a pillar of revenue and profit growth.

Mahaney’s core view is that if investors wait until the market fully confirms that massive AI investments are generating returns before entering, they may already have missed the re-rating phase. In other words, current market concerns over AI capex may actually keep valuations of tech stocks that are already showing returns relatively cheap.

As Meta, Amazon, and other tech giants sequentially announce future capex plans, the market’s focus will shift from 'how much AI will cost' to whether these companies can consistently convert every dollar of AI investment into higher revenue, profit, and free cash flow. If this conversion efficiency continues to improve, AI capex may no longer be just a market risk source, but could become the core engine of the next wave of tech stock profit growth.