As AI models grow more powerful, so do their demands on the infrastructure that supports them. Behind the headlines about language models and robot assistants is a less flashy — but equally transformative — trend: an unprecedented surge in data center investment.
Big Tech is leading the charge. In 2025, cloud giants like Amazon (AWS), Google (Google Cloud), and Microsoft (Azure) are scaling up capital expenditures at an aggressive pace to support the next generation of AI workloads. That spending is creating ripple effects across chipmakers, energy infrastructure firms, and the companies that keep hyperscale data centers running.
For investors, this buildout represents a multi-year tailwind — and a real opportunity to position early in the companies powering AI behind the scenes.
The Hyperscalers Are Going All In
Recent earnings calls and public filings reveal just how massive the data center expansion has become:
- Amazon Web Services (AWS) expects to spend approximately $105.2 billion on property and equipment in 2025. According to CFO Brian Olsavsky, the majority of that spend is earmarked for tech infrastructure, specifically to support the “growing need” for AI-powered workloads through AWS.
- Alphabet (Google Cloud) is planning to increase its capital expenditures to $75 billion in 2025 — up from $52.5 billion in 2024. CFO Anat Ashkenazi told investors that the spending will go “primarily for servers, followed by data centers and networking.”
- Microsoft plans to invest $80 billion in fiscal 2025. That’s part of a broader push to double its data center capacity to support AI services through Azure, with Vice Chair Brad Smith citing the need to meet rapidly expanding demand for AI-enabled cloud infrastructure.
These aren’t one-time splurges — they signal the start of a sustained infrastructure race. The cloud arms race is being fueled by surging AI compute needs, especially with more resource-intensive workloads like large language models (LLMs), video generation, agentic AI, and enterprise AI platforms.
Who Stands to Benefit?
Hyperscaler spending doesn’t just benefit the cloud divisions themselves — it flows into an entire ecosystem of companies that provide chips, servers, cooling, power management, and physical infrastructure.
Semiconductors & AI Hardware
- Nvidia (NVDA) – The dominant supplier of GPUs and AI accelerators, Nvidia continues to secure massive orders from all major cloud providers. Its new Blackwell chips and upcoming Rubin architecture are central to next-gen AI computing.
- AMD (AMD) – Competing in the AI chip space with its Instinct MI300 accelerators, AMD has also gained traction with major cloud platforms.
- Broadcom (AVGO) and Marvell Technology (MRVL) – Both supply custom chips and networking silicon for hyperscale data centers.
Power, Cooling & Infrastructure
- Vertiv Holdings (VRT) – Specializes in critical infrastructure for data centers, including cooling systems, power management, and battery backups. Shares have surged over 300% in the past 12 months (as of March 2025) on AI demand tailwinds.
- Eaton Corp (ETN) – Provides advanced electrical systems and power distribution critical to hyperscale installations.
- Schneider Electric (EPA: SU) – A global leader in energy management systems for data centers and industrial operations.
Cloud & Infrastructure Providers
While the hyperscalers themselves are investing heavily, they’re also building out proprietary hardware and leasing to enterprise clients — making their stocks potential long-term compounders in the AI theme:
- Amazon (AMZN)
- Alphabet (GOOGL)
- Microsoft (MSFT)
How to Gain Exposure: A Watchlist for the AI Infrastructure Boom
For investors looking to tap into the broader data center opportunity, here’s a starter watchlist broken down by category:
| Category | Company | Ticker |
| AI Chips | Nvidia | NVDA |
| AMD | AMD | |
| Networking Silicon | Broadcom | AVGO |
| Marvell Technology | MRVL | |
| Cooling & Power | Vertiv Holdings | VRT |
| Eaton Corp | ETN | |
| Schneider Electric | SU.PA | |
| Cloud Providers | Amazon | AMZN |
| Microsoft | MSFT | |
| Alphabet | GOOGL |
Want broader exposure? Consider ETFs focused on infrastructure and semiconductors:
- Global X Data Center REITs & Digital Infrastructure ETF (VPN)
- iShares Semiconductor ETF (SOXX)
- First Trust Cloud Computing ETF (SKYY)
Bottom Line
The AI boom isn’t just about models and software — it’s about massive physical infrastructure that’s being built to power them. The hyperscalers are placing their bets with billions in capex, and the companies enabling that transformation are already seeing demand soar.
For investors, following the money means following the data center gold rush — and positioning accordingly.