Amazon recorded Q4 2025 revenue of $213.4 billion, once again proving the strategic flexibility that comes with being a company of scale. Revenue grew double-digits, and both operating income and net income expanded. Simultaneously, the company continues aggressive investment in AI infrastructure even at the cost of free cash flow.
The core of this performance is one thing.
Amazon has entered not a 'profit maximization phase' but a phase of 'determining AI infrastructure leadership.'
Significance of 24% AWS Growth: Not a Simple Rebound but 'Structural Acceleration'
Q4 2025 AWS revenue was $35.6 billion, up 24% year-on-year. This is the highest growth rate among all Amazon business segments. More important is that this growth is not a short-term AI trend effect but the result of long-term contracts and infrastructure choices.
AWS secured many large customers including OpenAI, Visa, BlackRock, Salesforce, and the U.S. Air Force. These are not customer groups in a simple experimentation stage but those migrating mission-critical workloads to the cloud. That is, AWS growth is occurring not in the 'AI testing' stage but in the 'AI operations stage.'
In this sense, AWS's 24% growth is interpreted as a signal demonstrating that the 'cloud growth slowdown' phase the market once worried about is being re-accelerated by AI inference, agent, and large-scale operations demand.
Free Cash Flow Decrease Is a Choice, Not a Warning
Amazon's operating cash flow for the past 12 months was $139.5 billion, up 20%, but free cash flow decreased to $11.2 billion. On the surface this appears burdensome.
However, the cause is clear: capital expenditure (Capex) increased $50.7 billion year-on-year.
Amazon stated that most of this investment is AI infrastructure expansion. This is the decision to preempt 'physical capacity' capable of handling AI demand for the next 5 to 10 years, even at the cost of short-term cash flow. In other words, free cash flow decrease is closer to strategic transformation costs than performance deterioration.
Proprietary AI Semiconductor Strategy: Attempt to Dominate Cost Structure
Amazon disclosed proprietary AI semiconductor results in relatively detail in the earnings announcement. Trainium and Graviton surpassed $10 billion on an annualized revenue basis, and Trainium2 already has 1.4 million chips deployed.
This represents important changes:
- Core of AI competition moving from model performance to inference cost
- Controlling costs with proprietary silicon instead of GPU dependency
- Defending margins through vertical integration with AWS Bedrock
Amazon aims not simply to be 'a cloud that provides AI' but a platform capable of operating AI most cheaply and at scale. At this point Amazon's strategy clearly diverges from competitors.
Changing Role of Retail, Advertising, and Media: Enhancement as Cash Generation Engines
If AWS is the growth driver, retail, advertising, and media are the cash generation engines supporting investment.
- Prime same-day delivery volume increased 70%
- 'Amazon Now' ultra-fast delivery global expansion
- Prime Video sports broadcasting strengthened
- AI shopping assistant 'Rufus' contributed $12 billion in revenue
The common point of these segments is that AI creates efficiency rather than cost. Amazon is applying AI to retail automation, advertising recommendations, and media personalization to generate immediate revenue contribution. This is a case refuting the perception that 'AI burns money.'
The True Meaning of the 2026 $200 Billion Investment Plan
Andy Jassy announced plans for approximately $200 billion in capital expenditure in 2026. This number goes beyond simply being large — it reveals that Amazon views AI infrastructure competition as a 'physical endurance contest.'
This strategy rests on three premises:
- AI demand is not temporary
- Infrastructure becomes more expensive the later it's built
- Scale once established becomes barriers to entry
Amazon is a company that already proved this logic in retail. It is repeating the same approach in AI infrastructure.
Amazon Is Becoming the 'Logistics Company of the AI Era'
Amazon's 2025 performance can be summarized thus:
- AWS has entered the acceleration phase again
- Free cash flow decrease is a strategic choice
- Proprietary semiconductors and infrastructure create cost competitiveness
- Retail and advertising are the safety nets for AI investment
Amazon chose the path of dominating the infrastructure that supports all of these competitions rather than directly diving into AI model competition.
In the AI era, the strongest company is likely not the one with the smartest model, but the one capable of running the most AI most cheaply.
And currently the company closest to that condition is Amazon.
