The investment cycle in artificial intelligence infrastructure is gaining momentum. According to my calculations, based on market data analysis, the combined capital expenditures of the four largest cloud service providers (CSPs) will exceed $700 billion by 2026. This is not just another record—it is a signal of the start of a new supercycle in the semiconductor industry that will reshape the entire value chain.
The latest earnings season only confirmed this trend. Companies from the "Magnificent Seven" once again showed strong results, but the main focus of investors was on the upward revision of capital expenditure forecasts. Instead of signs of a peak, new data indicates that hyperscalers see more and more opportunities for investment in data centers, proprietary chips, and computing infrastructure.
How Tech Giants Are Ramping Up Investments
The main drivers of this cycle remain the four dominant cloud providers. Amazon (AWS) will be the largest investor with capital expenditures of around $200 billion, directed at expanding data centers and developing its own Trainium and Inferentia chips. Microsoft (Azure) raised its forecast to approximately $190 billion, deepening its partnership with OpenAI and expanding infrastructure for Copilot. Alphabet (Google Cloud) increased its target to $180–190 billion, focusing spending on expanding its TPU network and developing Gemini models.
Meta significantly raised its capital expenditure target to $125–145 billion. The funds will mainly go toward training open-source Llama models and improving recommendation algorithms. In addition to traditional cloud providers, Tesla has also joined the race, raising its 2026 target to over $25 billion. Investments will be directed at deploying AI servers and expanding the Dojo supercomputer.
Meanwhile, Apple and NVIDIA are pursuing different strategies. Apple is betting on research and integrating AI into its own chips and devices. NVIDIA, on the other hand, is focused on developing next-generation chips and investing in AI providers, such as CoreWeave.
How This Affects the Supply Chain
Behind the sharp rise in semiconductor sales lies the entire value chain in AI computing. At its core are AI servers, with chips serving as their "brains." The top level includes advanced chip manufacturing, such as at TSMC, advanced CoWoS packaging, and high-bandwidth memory (HBM) from companies like SK Hynix. The middle level consists of the cloud providers themselves, linking equipment suppliers with AI developers.
The bottom level is large language models, which serve as the ultimate source of demand. This includes developers like OpenAI and Anthropic, as well as a growing number of AI applications. Both training and inference require enormous amounts of computing power, sustaining steady demand across the entire chain.
Noticing the acceleration in demand, providers are expanding capacity—which in turn spurs orders across the semiconductor industry. Sustained demand from cloud providers has already resulted in stable orders for GPUs from NVIDIA and AMD, as well as custom ASIC chips.
Global semiconductor sales show a strong positive correlation with the capital expenditures of major tech companies, and both indicators are hitting new records. At the same time, analysts point to a key indicator for tracking a potential reversal—the AI monetization coverage ratio, which correlates potential AI revenue with capital expenditures. 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, meaning the semiconductor supercycle continues.
My expert opinion: The market is underestimating the scale of the upcoming supercycle. $700 billion is just the tip of the iceberg. If AI monetization continues to grow at this pace, we will see even more aggressive capital expenditures, which will directly pressure chip manufacturers and, consequently, the cryptocurrency market, where mining and decentralized computing are directly dependent on semiconductor availability.