The artificial intelligence arms race is entering a new phase. Based on my calculations from analyzing MacroMicro data, the combined capital expenditures of the four largest cloud service providers (CSPs)—Amazon, Microsoft, Alphabet, and Meta—will exceed $700 billion by 2026. This is not just another record, but a clear signal of the beginning of a new supercycle in the semiconductor industry.
The latest earnings season only confirmed this trend. Instead of signs of a slowdown, which many skeptics had predicted, we see hyperscalers actively increasing their budgets. Investors are now most concerned not with current profits, but with the potential for future AI revenue. And companies are responding to this demand with aggressive expansion of data centers, development of their own chips, and modernization of computing infrastructure.
Who is spending how much?
The leader in investment volume will be Amazon (AWS) with projected spending of around $200 billion. The funds will go toward 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 and scaling infrastructure for Copilot. Alphabet (Google Cloud) is also keeping pace, targeting $180-190 billion, with a focus on TPUs and the Gemini model. Meta, in turn, has significantly revised its plans upward—to $125-145 billion—allocating resources to training open-source Llama models and improving recommendation algorithms.
Tesla has also joined this race, increasing its 2026 target to over $25 billion. These funds will go toward deploying AI servers and expanding the Dojo supercomputer. Notably, Apple and NVIDIA are pursuing different strategies: the former is betting on integrating AI into its own devices, while the latter is focusing on developing next-generation chips and investing in AI providers such as CoreWeave.
Impact on the supply chain
The sharp rise in semiconductor sales is not a coincidence, but a direct consequence of this investment cycle. At the core of the entire value chain are AI servers, whose "brains" are chips. The top tier includes advanced manufacturing at TSMC, complex CoWoS packaging, and high-bandwidth memory HBM from SK Hynix. The middle tier consists of the cloud providers themselves, linking equipment suppliers with AI developers. The bottom tier includes large language models from OpenAI and Anthropic, as well as a growing number of AI applications that require enormous computing power for both training and inference.
My analysis shows that global semiconductor sales exhibit a strong positive correlation with the capital expenditures of tech giants. Both indicators are hitting new records. The key indicator for tracking a reversal of this supercycle is the AI monetization coverage ratio, which correlates potential AI revenue with capital expenditures.
Currently, the ratios for Meta (1.8), Microsoft (1.5), Google (1.1), and Amazon (0.7) confirm that these companies' investments are justified. This means the semiconductor supercycle continues, and we are only at the beginning of this journey.
My expert opinion: The market is clearly underestimating the long-term implications of this investment cycle. Current spending is not just a "bubble," but a fundamental laying of infrastructure for the next decade. Investors should closely watch the companies enabling this infrastructure, from chip manufacturers to energy suppliers for data centers.