The investment cycle in artificial intelligence infrastructure is gaining momentum, and new data suggests we are only at the beginning of a massive supercycle. According to my calculations, based on an analysis of the latest reports, the total capital expenditures (CapEx) of the four largest cloud providers — Amazon (AWS), Microsoft (Azure), Alphabet (Google Cloud), and Meta — will exceed the $700 billion mark by 2026. This will be an absolute record and will dispel the doubts of skeptics who expected an imminent peak in investment.

The key driver of this growth is not just an increase in budgets, but a fundamental shift in strategy. Instead of cutting costs, tech giants are, on the contrary, revising their forecasts upward. They see more and more opportunities for monetizing AI, which pushes them to aggressively expand data centers, develop their own chips, and increase computing power.

Distribution of forces: who spends how much

The leader of the race remains Amazon (AWS) with projected costs of around $200 billion. The main funds will go towards expanding the data center network and developing its own Trainium and Inferentia chips. Microsoft (Azure) has raised the bar to approximately $190 billion, deepening its partnership with OpenAI and deploying infrastructure for Copilot. Alphabet (Google Cloud) is targeting $180-190 billion, focusing on the TPU network and developing Gemini models. Meta, despite all regulatory difficulties, has significantly increased its target to $125-145 billion, directing funds towards training open-source Llama models and improving recommendation algorithms.

Notably, Tesla has also joined this race, raising its target for 2026 to more than $25 billion. Its investments will go towards deploying AI servers and expanding the Dojo supercomputer. Meanwhile, Apple and NVIDIA adhere to different strategies: the former focuses on integrating AI into its own chips and devices, while the latter focuses on developing next-generation chips and investing in AI providers such as CoreWeave.

Value chain and the supercycle

Behind the sharp rise in semiconductor sales lies the entire value chain in AI computing. The top tier is advanced chip manufacturing (TSMC, CoWoS, HBM from SK Hynix). The middle tier is the cloud providers themselves, who connect equipment suppliers with AI developers. The bottom tier consists of large language models (OpenAI, Anthropic) and a growing number of AI applications. Both training and inference require enormous amounts of computing, which sustains stable demand across the entire chain.

Global semiconductor sales show a strong positive correlation with the capital expenditures of tech giants, and both indicators are hitting new records. The key indicator to track for a potential reversal is the AI monetization coverage ratio, which correlates potential AI revenue with CapEx. According to my estimates, the ratios for Meta (1.8), Microsoft (1.5), Google (1.1), and Amazon (0.7) currently confirm that these companies' investments are justified.

My expert opinion: These figures are not just accounting reports. They signal that we are witnessing not a speculative bubble, but a structural shift. The supercycle in semiconductors continues, and its foundation is real demand for AI computing power. Investors should closely monitor the dynamics of the monetization ratio: as long as it is above 1.0, the race makes sense, and any correction in the sector will only be temporary.