Forecasting Record AI Spending into 2025

By Big Data Federation, Inc.

Generative AI’s rapid adoption has ignited a massive spending boom, particularly among the largest technology companies, which have committed unprecedented capital expenditure (CapEx) on the hardware essential for AI model development and operation. In this article, we examine the CapEx trends of major hyperscalers1 worldwide and forecast spending through 2025, leveraging extensive macroeconomic data and ecosystem interconnections.

In the first half of 2024 (1H24), hyperscalers CapEx climbed fifty percent year-over-year compared to 1H23. Will this unprecedented surge continue into next year? Our machine-generated fundamental forecasts indicate that, while the pace may decelerate, hyperscalers CapEx will still reach new heights, with substantial expansion projected into 2025.

For the full year 2024, our models forecast hyperscalers CapEx growth between 58% and 64% compared to 2023, fueled by the Generative AI infrastructure rollout. Cloud computing, remains the bedrock of generative AI, supporting both its computational demands and extensive data storage needs. Data center construction is at a record, with advanced interest resulting in spaces leased before being built.

Looking ahead to 2025, we project a robust year-over-year CapEx growth rate of 40% to 50%. Recent earnings calls underscore that hyperscalers, wary of underinvestment, are racing to expand capacity ahead of demand, choosing to risk overinvestment rather than miss future opportunities.

Hyperscalers’ CapEx typically progresses through build-up, absorption, and expansion phases. Presently, we are still deep in the build-up stage, where massive projects have sparked a multi-year surge in hyperscalers CapEx, potentially prolonging this phase before moving into absorption.

1 AAPL, AMZN, BABA, BIDU, CRM, EBAY, GOOGL, IBM, META, MSFT, NOW, ORCL, PYPL, SAP, and 700-HK (Tencent Holdings Limited)

Published by Big Data Federation · All insights · Subscribe to Insights

Financial AI ®