The artificial intelligence arms race is entering a fundamentally new level. According to my analysis of recent reports and forecasts from leading technology corporations, the combined capital expenditures of the four largest cloud service providers (CSPs) — Amazon, Microsoft, Google, and Meta — will exceed $700 billion by 2026. This unprecedented surge in investment clearly indicates that we are witnessing not just a cycle, but the beginning of a full-fledged supercycle in the semiconductor industry.
The key driver of this growth is not scattered startups, but rather the cloud giants, who see virtually limitless opportunities for monetizing AI. Instead of an expected peak in investment, we see a consistent upward revision of forecasts. Amazon (AWS) is preparing to spend around $200 billion on expanding its data center network and developing its own Trainium and Inferentia chips. Microsoft (Azure) has raised the bar to $190 billion, deepening its partnership with OpenAI. Alphabet (Google Cloud) is targeting $180–190 billion, focusing on TPUs and the Gemini model. Meta, in turn, has significantly revised its budget to $125–145 billion, allocating funds for training Llama.
Importantly, Tesla has also joined this race, increasing its 2026 targets to over $25 billion for deploying AI servers and the Dojo supercomputer. Meanwhile, Apple and NVIDIA are pursuing different strategies: the former is betting on integrating AI into its own devices, while the latter focuses on developing next-generation chips and investing in AI providers.
How this impacts the supply chain
The growth in semiconductor sales directly correlates with the capital expenditures of tech giants. The top tier of the chain is advanced chip manufacturing (TSMC, CoWoS) and high-bandwidth memory HBM (SK Hynix). The middle tier consists of the cloud providers themselves, connecting manufacturers with AI developers. The bottom tier comprises large language models (OpenAI, Anthropic), which ultimately create the demand for computing power.
The AI monetization coverage ratios for Meta (1.8), Microsoft (1.5), Google (1.1), and Amazon (0.7) currently confirm that the investments are justified. Demand for GPUs from NVIDIA and AMD, as well as for custom ASIC chips, remains stable.
My expert opinion: These figures are not just forecasts, but plans already embedded in budgets. The semiconductor supercycle will continue, and the key indicator to watch for a reversal is not a decline in capital expenditures, but a drop in the monetization ratio. As long as it remains above 1.0, the market stays bullish.