Morgan Stanley: Four major catalysts of AI reshape the internet landscape next year, with Amazon, Meta, and Alphabet being the most favored among the giants

Morgan Stanley pointed out in its latest report that four key generative AI (GenAI) catalysts—model advancements, agentic experiences, capital expenditures, and custom chips—are reshaping the landscape of the internet industry.

According to news from the trading desk, Morgan Stanley analyst Brian Nowak stated in a report released on September 16 that the continuous breakthroughs in leading AI models and the proliferation of "agentic" AI experiences aimed at automating tasks are driving the industry into a new growth phase. These technological advancements will not only enhance user experience but also further drive the digitalization of consumer spending.

To support this wave of technological revolution, tech giants are investing at an unprecedented scale. The firm predicts that the total capital expenditures of six major tech companies will reach approximately $505 billion by 2026 and further increase to $586 billion in 2027, which will put pressure on the companies' free cash flow but also lay the foundation for future growth.

Based on an analysis of these trends, Morgan Stanley has clarified its preference order for large tech stocks over the next 12 months, in the following sequence: Amazon, Meta, and Google. The firm believes that these three companies have the capability to leverage AI catalysts to solidify their market positions and create new revenue streams.

Four Major Macro AI Catalysts

Morgan Stanley believes that the performance of the internet industry in the coming years will be primarily driven by four macro AI forces:

• Acceleration of Model Development: The report expects that top AI models will continue to improve, or even accelerate in their advancements. Ample capital, continuously improving chip computing power, and significant room for development in agentic capabilities will drive companies like OpenAI, Google, and Meta to release new generations of more powerful models.
• Proliferation of Agentic Experiences: Agentic AI products can provide more personalized, interactive, and comprehensive consumer experiences, further promoting the digitalization of consumer wallets. To achieve large-scale application, obstacles such as computing capacity, reasoning ability, and transaction process smoothness still need to be overcome.
• Surge in Capital Expenditures: The report predicts that by 2026, the total capital expenditures of the six giants (Amazon, Google, Meta, Microsoft, Oracle, CoreWeave) on data centers will reach $505 billion, a year-on-year increase of 24%. The firm noted that building a gigawatt (GW) level data center still requires approximately $40-50 billion in capital expenditures.
• Increased Importance of Custom Chips: The report believes that the likelihood of third-party companies testing and adopting custom ASIC chips like Google TPU and Amazon Trainium is increasing. Although NVIDIA's software ecosystem poses certain barriers, cost-effectiveness and capacity limitations will drive customers to seek ASIC solutions, especially in reasoning workloads. Morgan Stanley pointed out that if Google and Amazon can make progress in this area, it will represent significant upside potential beyond their current valuations

Capital Expenditures Surge, Squeezing Free Cash Flow

The massive capital expenditures are a heavy bet by tech giants on the future of AI, but they also directly impact their financial conditions. Morgan Stanley's model shows that from 2024 to 2027, the capital expenditures of the six major tech giants are expected to grow at a compound annual growth rate of 34%.

The report estimates that such a scale of investment will significantly affect the companies' free cash flow. By 2026, the infrastructure capital expenditures of Google, Meta, and Amazon are expected to account for approximately 57%, 73%, and 78% of their pre-tax free cash flow (FCF), respectively. This indicates that these companies are willing to sacrifice short-term profitability in exchange for long-term technological and market advantages to maintain their lead in the AI race.

Amazon: AWS Acceleration and Retail Margin Improvement

Amazon is Morgan Stanley's top pick among large tech stocks, with a target price of $300 and a rating of "Overweight." Its bullish logic is primarily based on two pillars: the re-acceleration of AWS business and the continued improvement of North American retail margins.

For AWS, the report analyzes the pace of its data center construction, expecting a significant increase in data center space in 2025 and 2026, which provides a physical basis for achieving over 20% revenue growth in 2026, higher than the bank's current base forecast of 19%.

In terms of retail business, the report points out that there is still significant room for improvement in Amazon's North American retail margins. As of the second quarter of 2025, the profit margin for this business was -1%, far below the approximately 1% level in 2018. Morgan Stanley's base model expects it to return to 2018 levels by 2028, indicating the possibility of further upward adjustments to its earnings per share forecasts for 2026/2027.

Meta: Core Business Improvement and "Call Options"

Morgan Stanley also maintains an "Overweight" rating on Meta, with a target price of $850. The firm believes that investors should focus on improvements in its core platform, the release of the next-generation Llama model, and several undervalued "call options."

The report states that Meta is leveraging GPUs to drive improvements in its core advertising business, with significant room to enhance user engagement and monetization capabilities. Meanwhile, the market anticipates the company will release a thoroughly tested and improved next-generation Llama model in early 2026 In addition, new businesses such as Meta AI search and Business Messaging are important long-term growth drivers. Morgan Stanley estimates that Meta AI search alone could create an annual revenue opportunity of approximately $22 billion by 2028. Survey data also shows that despite the late release of Meta AI, its user adoption rate has quickly caught up with ChatGPT and Google's Gemini.

Google: Search Innovation and Cloud Growth Prospects

Morgan Stanley maintains an "Overweight" rating on Google, with a target price of $210. The firm focuses on three core issues: AI-driven search growth, potential shifts in user commercial behavior, and the growth of Google Cloud (GCP).

The report points out that innovations in AI Overviews, AI Mode, and other areas are expected to drive accelerated growth in search revenue, with projected growth rates of 12% in the second half of 2025 and 9% in 2026. Morgan Stanley's survey data indicates that Google remains the preferred platform for consumers in commercial behaviors such as product research and price comparison, ahead of Amazon and other competitors.

In terms of cloud business, Google's Gemini model and TPU chips are seen as key drivers of GCP growth. The report mentions that companies like Meta have signed partnerships with GCP, which is expected to contribute approximately 300 basis points to Google's cloud business growth by 2026. The firm believes that the market has not fully priced in Google's advantages in custom chips, which constitutes a potential upside catalyst.