See this visualization first on the Voronoi app.

Use This Visualization
Charted: The Rise of AI Hyperscaler Spending
This was originally posted on our Voronoi app. Download the app for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
The five big hyperscalers poured an estimated $197 billion into AI infrastructure in 2024, with spending set to rise further.
For Microsoft, Alphabet, Meta, Amazon, and Oracle, spending is soaring. In order to power and train AI models, these companies are investing heavily in data centers, graphics processing units (GPUs), and cloud-hosted infrastructure. In 2025, Microsoft is projected to spend $80 billion on data centers alone.
This graphic shows the growth in AI hyperscaler capital spending, based on data from J.P. Morgan Asset Management.
AI Hyperscalers’ Spending Boom
Below, we show the big five hyperscalers’ capital expenditures over time, driven by increasing demand for AI that requires significant storage capacity and computing power to run:
Year | AI Hyperscaler Capital Spending (B) |
2026P | $249 |
2025P | $234 |
2024P | $197 |
2023 | $126 |
2022 | $123 |
2021 | $92 |
2020 | $73 |
2019 | $67 |
2018 | $65 |
2017 | $40 |
2016 | $31 |
2015 | $25 |
2014 | $23 |
2013 | $17 |
2012 | $10 |
2011 | $8 |
2010 | $7 |
2009 | $4 |
2008 | $7 |
2007 | $5 |
2006 | $4 |
2005 | $2 |
2004 | $2 |
With the highest projected capital spending, Amazon spent an estimated $75 billion in 2024, up from $48.4 billion in the previous year.
Amazon Web Services (AWS), its cloud-services unit, is ramping up investment in AI infrastructure amid global demand for advanced applications. As Amazon’s most profitable division, CEO Andy Jassy claims it offers nearly double the AI features of other major cloud providers put together.
Meanwhile, Meta Platforms announced it plans to spend up to $40 billion in 2024, equal to about a quarter of projected annual revenues. In 2024, the company sought to acquire 350,000 GPUs from Nvidia to train its large language models. Together, these cost an estimated $10.5 billion.
When it comes to Google Cloud, massive investments in AI generated billions in revenue in 2024, with over two million software developers using the platform.
As AI hyperscalers increasingly require more data center capacity to train AI models, these providers will struggle to keep up with demand. This comes as U.S. data center property inventory has expanded by 43% annually in 2023 and 2024, on average, amid demand from big tech firms. For perspective, U.S. apartment inventory increased by a mere 2% over the period.
Learn More on the Voronoi App 
To learn more about this topic from a data center perspective, check out this graphic on big tech spending on data centers in 2024.