Four of the largest technology companies in the world, Alphabet, Amazon, Meta, and Microsoft, are set to collectively spend approximately $650 billion on capital expenditures in 2026, with the bulk of that money going into data centers, AI chips, servers, and the power infrastructure needed to run them.
The figure, first reported by Bloomberg and later confirmed by Bridgewater Associates in a letter to clients, represents a 60% increase from the roughly $410 billion the four companies spent in 2025, and marks the largest single-year capital spending commitment by any group of companies in at least a decade.
What Each Big Tech Is Spending
Amazon leads the group with the most aggressive plan to spend $200 billion in capital expenditures for 2026, announced during its most recent earnings call. That is nearly a 50% jump year-over-year, and the majority is earmarked for Amazon Web Services (AWS) to handle growing demand for cloud-based AI workloads.
Alphabet, Google’s parent company, comes in second with a projected spend of between $175 billion and $185 billion. The company is directing those funds toward its Gemini AI models, its Vertex AI enterprise platform, and the expansion of Google Cloud infrastructure.
Microsoft has an annual run rate of $145 billion in capital expenditures for its 2026 fiscal year, which began in July 2025. The spending covers the buildout of data centers that support Azure, its cloud platform, as well as the Copilot AI tools embedded across its Microsoft 365 suite. The company reported a 66% increase in second-quarter capital spending, exceeding analyst estimates.
Meta rounds out the group with a projected spend of $115 billion to $135 billion. The company is focusing its investment on infrastructure for its Llama open-source AI models and AI-powered advertising systems. At $135 billion, that would represent nearly an 87% increase from its 2025 capex.
What the Money Is Actually Buying
The spending is not going into software development or research labs in any significant proportion. What these companies are building, at scale, is physical infrastructure. That means land, construction, specialized AI chips, high-capacity servers, networking cables, cooling systems, backup generators, and the electrical grid connections to power it all.
Meta has already broken ground on a 1-gigawatt data center campus in Lebanon, Indiana, describing it as one of the company’s largest infrastructure investments, valued at over $10 billion. Alphabet is building out facilities across multiple regions to meet enterprise demand. And Amazon has separately announced a $12 billion investment in northwest Louisiana for data center infrastructure.
Bridgewater’s co-chief investment officer Greg Jensen noted in the letter to clients that “compute demand continues to significantly outpace supply, driving hyperscalers to invest even more rapidly to try to someday get ahead of the demand.” This supply-demand gap is one reason all four companies are accelerating spending at the same time rather than waiting.
Broader Economic Effects
Bridgewater also estimates that AI-related capital expenditure from these companies added about 50 basis points to the U.S. GDP growth in 2025. That contribution is projected to reach around 100 basis points in 2026, which is comparable to what business investment contributed during the 2000s Dotcom boom.
The spending is also creating upward pressure on electricity prices in some regions and contributing to higher prices for technology and communications equipment. Jensen cautioned that the AI boom has entered what he described as a “more dangerous phase,” where rising dependence on outside capital and the sheer scale of physical investment create meaningful downside risk if demand forecasts don’t materialize or if financial markets pull back sharply.
The chief investment officer also made mention of the ripple effects of AI on other sectors like software. “It is no longer possible for AI leaders to satisfy their investors’ expectations without creating existential risks to other sectors like software,” Jensen noted.
Whether $650 billion in infrastructure spending ultimately justifies itself will depend on how quickly and broadly AI tools generate revenue for businesses and consumers. What is clear right now is that Alphabet, Amazon, Meta, and Microsoft have each decided that the risk of underbuilding is greater than the risk of overcommitting.
