From: allin
Google’s significant investment in Artificial Intelligence (AI) and its strategic direction have been a focal point of recent discussions, particularly concerning the company’s financial performance and future outlook. Despite strong earnings, a projected surge in capital expenditure has garnered attention, signaling Google’s aggressive push in the AI space [01:10:09].
Financial Performance and AI Investment
Google’s recent financial reports show robust growth. Revenue was up 12% year-over-year, reaching 12 billion and YouTube ad revenue surging by 14% to 26.5 billion [01:09:24]. For the full year, Google reported 100 billion in net profit [01:09:33].
A key concern for investors, leading to a 7% stock drop, was Google’s announced capital expenditure (capex) [01:09:09]. The company plans to invest $75 billion in capex in 2025, marking a 42% jump over 2024 and 29% more than analysts anticipated [01:10:00]. This substantial investment is primarily allocated to data centers, servers, and the overall AI buildout [01:10:09].
AI Models and Productization
Google’s AI models are considered among the best across a wide range of capabilities [01:10:40]. The challenge for Google is to effectively translate these advanced models into better products [01:10:53]. While competitors like OpenAI are noted for being faster and having better margins in deep research, these are aspects that can be improved through post-training and productization efforts [01:11:03].
Google’s significant revenue stream from ad targeting is expected to directly benefit from AI-driven optimizations [01:11:27]. The company is encouraged to provide more detailed disclosures on how the $75 billion capex is segregated, especially how much goes towards AI-enabled models for ad optimization versus more speculative pre-training and post-training initiatives [01:11:40].
Strategic Outlook and Return on Invested Capital (ROIC)
The focus on Return on Invested Capital (ROIC) is a key framework for Google’s AI investment [01:12:50]. It is argued that if Google could demonstrate that a significant portion of its investment is for immediate ad optimization, which has a higher ROIC, the market would react more positively [01:12:53]. The company’s ROIC is already increasing due to the implementation of AI models into its operations [01:13:01].
Google’s historical approach to computing infrastructure investment, characterized by frugality and thoughtful management, suggests that the $75 billion capex is a positive signal [01:14:14]. Their past strategies included using cheap, disposable racks in the 98-2005 era, focusing on energy efficiency from 2005-2010, and extending depreciation schedules from 2-3 years to 6 years by 2021 due to AI optimization on maintenance [01:14:19]. This history implies a strong line of sight on how they will achieve full utilization and a great return on their investment [01:14:43].
With an assumed 20% ROIC on 15 billion in profit annually, plus the amortization, which is approximately 27 billion in operating profit to meet their ROIC performance, which is seen as achievable, representing less than 20% of their current annual operating profit [01:16:07].
The confidence in this investment also stems from Google’s in-house capabilities, such as their TensorFlow chips, which contribute to cost-effectiveness and efficiency [01:13:50]. This investment is viewed as a “strong and positive indicator” that Google has confidence in its strategy to transition from traditional search to chat-based interfaces [01:16:43], reinforcing the idea that they know where to invest rather than underinvesting [01:17:03].