NextFin News - In a decisive move that has reshaped the technology sector's financial landscape, Amazon and Google have taken a commanding lead in the global race for artificial intelligence infrastructure. According to GeekWire, Amazon CEO Andy Jassy announced on February 5, 2026, that the company plans to spend a record $200 billion in capital expenditures (capex) this year, primarily targeting AI, custom chips, and robotics. This follows a Wednesday announcement from Alphabet, Google’s parent company, which projected its own 2026 capex at between $175 billion and $185 billion—nearly double its 2025 spending levels. These disclosures, made during the fourth-quarter earnings cycle in Seattle and Mountain View, represent the largest single-year infrastructure commitments in the history of the internet era.
The scale of this spending has triggered immediate volatility in the capital markets. Amazon shares fell 10% in after-hours trading following the report, as investors grappled with the reality that the company’s free cash flow dropped to $11.2 billion in 2025, down from $38.2 billion the previous year. Similarly, Microsoft and Meta have signaled aggressive spending, with Microsoft reporting $37.5 billion in capex for the most recent quarter alone. The primary driver for this unprecedented capital flight is the urgent need to build out data centers and procure the specialized hardware required to run generative AI models at scale. Jassy defended the strategy during the earnings call, stating that Amazon is "monetizing capacity as fast as we can install it," citing a 24% growth rate for Amazon Web Services (AWS), which reached $35.6 billion in quarterly revenue.
The underlying cause of this spending surge is a fundamental shift in the cloud computing business model. For the past decade, cloud providers focused on general-purpose compute; today, the industry is pivoting toward "AI-native" architecture. A critical component of this transition is the move toward custom silicon. For the first time, Amazon disclosed that its in-house chips, Trainium and Graviton, have reached a combined annual run rate of over $10 billion. By designing their own processors, Amazon and Google are attempting to break their dependency on external chipmakers and reduce the long-term cost of running massive AI workloads. This vertical integration is no longer a luxury but a prerequisite for maintaining margins in an era where AI compute requirements are doubling every few months.
From an analytical perspective, the "Capex Race" of 2026 is less about current-quarter profits and more about securing a seat at the table for the next decade of enterprise computing. The impact of this spending is twofold. First, it creates a massive barrier to entry. With the "entry fee" for a competitive AI cloud now measured in the hundreds of billions of dollars, smaller players are effectively being locked out of the foundational model market. Second, it shifts the risk profile of these tech giants. By plowing nearly all operational cash back into physical infrastructure, companies like Amazon are betting that the demand for AI services will remain inelastic. If the anticipated AI productivity boom fails to materialize for enterprise clients, these companies will be left with billions of dollars in depreciating hardware and underutilized data centers.
However, the data suggests that the demand is currently outstripping supply. AWS’s 24% growth—its fastest in three years—indicates that corporate America is aggressively migrating workloads to the cloud to take advantage of AI capabilities. Google’s decision to double its capex similarly reflects a need to defend its search dominance through integrated AI features and to expand its Google Cloud Platform (GCP) footprint. The trend toward "Sovereign AI," where nations require data to be processed within their borders, is also forcing these providers to build localized infrastructure at a pace never seen before.
Looking forward, the market should expect a period of "margin compression" followed by a potential "monetization harvest." As U.S. President Trump’s administration continues to emphasize American leadership in emerging technologies, the regulatory environment is likely to remain favorable for large-scale domestic infrastructure projects. By 2027, the focus will likely shift from how much these companies are spending to how efficiently they are utilizing their custom silicon to lower the "cost per inference." The winners of the 2026 capex race will be those who can successfully transition from being builders of AI infrastructure to being the indispensable utility providers of the AI economy.
Explore more exclusive insights at nextfin.ai.
