Investors are in the 'First Inning'—The Perspective to Master the AI Rally
In this article, based on a discussion by Allison Nathan, George Lee, and Eric Sheridan from Goldman Sachs Exchanges' 'AI Exchanges: How tech giants are navigating the AI landscape,' we explain how leading U.S. tech giants are responding to the AI era and what investors should focus on. We will make the technical content as accessible as possible, weaving in quotes and specific examples.
1. Tech Giants' Commitment—Capital Expenditure and Supply Constraints
1-1. Maintaining Capital Expenditure for the Remainder of 2025
At the beginning of the discussion, George Lee pointed out, 'These major companies have indicated their intention to maintain capital expenditures for AI support at least through the end of 2025.' Specifically, Meta is maintaining a capital intensity of approximately 40% of revenue, and Alphabet (Google) and Amazon are adhering to similar policies.
'We are first batter, second strike of the first inning.'
As this statement shows, while they hold high expectations for the future of this new 'game' called AI, they also acknowledge that it is still in the early stages.
1-2. Increasing Demand Under Supply Constraints
Meanwhile, as many customers seek AI workloads, 'supply constraints' in hardware such as GPUs are becoming more serious. In response, emerging companies like CoreWeave, which provides rendering and AI training infrastructure centered on NVIDIA GPUs, have appeared, playing a role in strengthening the supply chains of tech giants.
2. AI as the Third Computing Shift
2-1. Historical 'Shifts' and Essential Similarities
Eric Sheridan looked back at the history of computing and explained that there have been three stages of shifts:
Web 1.0 (Desktop Computing)
Web 2.0 (Mobile Computing)
Web 3.0 (A New Wave Centered on AI)
He explained that we are currently moving rapidly from the 'latter half' of the latter to the 'initial phase'.
'Before the arrival of ChatGPT, most people didn't know what ChatGPT was.'
He emphasized the 'buzz speed' of the AI interface, which gained over 800 million monthly active users in just two and a half years, based on this fact.
2-2. Progress from Infrastructure Layer to Platform Layer to Application Layer
He also stated, 'Every computing shift has a three-layer structure: infrastructure, platform, and application,' and defined them as follows:
Infrastructure Layer: Training and building Large Language Models (LLMs)
Platform Layer: Providing AI APIs and development environments
Application Layer: Actual business and consumer-facing services (e.g., Gemini, ChatGPT, etc.)
He points out that this follows a certain trajectory. We are currently in a transition period to the platform layer, and it is a phase where the 'true contenders' of the application layer will appear one after another.
3. Key Points for Investors to Watch
3-1. The Trade-off Between Capital Intensity and Growth Rate
While the capital expenditures of tech giants are reaching enormous levels,
Meta: Capital intensity of approximately 40% of revenue
Alphabet: 2025 CapEx projected at $75 billion
Amazon: Planned CapEx in the $100–$110 billion range
Such levels are being maintained, and it was suggested that 'if these levels continue, growth rates could potentially settle from 40–60% down to the mid-teens.'
3-2. Manifestation of Tariff Impacts
With recent U.S.-China trade friction and rising component procurement costs, cases where 'tariffs are driving up CapEx input costs' have been reported. Meta is taking measures to address macro factors, such as raising CapEx in anticipation of tariffs while downwardly revising Opex (operating expenses).
3-3. The Future of Advertising Automation and Search
Regarding the automation of advertising operations through AI adoption,
A virtuous cycle of 'automatically repeating ad creation, delivery, and performance measurement to achieve improved ROI' has been established, contributing to the strengthening of Google and Meta's core businesses. On the other hand, while 'search cannibalization' is a concern, the view is that no significant impact on search advertising revenue has been observed at present.
4. Future Outlook and Risk Factors
4-1. Winners and Losers in the Application Layer
The most important thing is 'which applications will be adopted after passing through the platform layer.' Just as Uber and Airbnb transformed the existing taxi and lodging industries during the Web 2.0 era, there is a possibility that emerging companies could threaten existing players in the AI era as well.
4-2. Investor Patience
While Eric expresses concern that investors are 'prone to being swayed by quarterly earnings,'
'The biggest key is whether you can make investment decisions with an eye on five years from now.'
This suggests that while future uncertainty is high, capital deployment with a long-term perspective is what will determine the winners and losers of the AI rally.
Tech giants are sparing no capital expenditure to expand AI infrastructure and are rapidly advancing from the platform layer to the application layer. Investors are required to discern macro factors (tariffs and economic fluctuations), each company's capital intensity and growth prospects, and ultimately, which applications will dominate the market. The 'early game' of AI is still ongoing—therefore, I would like to keep a close watch on future developments from a long-term perspective.
Related Articles
