U.S. Tech Giants Benefit From Expanding AI Investment Cycle

2026-05-13

Big tech companies such as Amazon.com and Google are deepening a structure in which they invest in AI startups that are also customers, capturing both cloud revenue and valuation gains. While boosting AI demand, the flow of money is also heightening signs of overheating.

Unrealized gains lift earnings

In Amazon's January-March 2025 results, Anthropic's valuation gains accounted for about 80% of the $16 billion in non-operating income. Amazon has invested $8 billion in the company by 2025 and has a partnership with it.

Alphabet, Google's parent, also posted $37.7 billion in non-operating income, explaining that most of it came from unrealized gains on unlisted shares. It had invested $3 billion in Anthropic by 2025 and also held more than 6% of Elon Musk's SpaceX as of the end of 2025.

Net profit at both companies rose by about 80% year on year in the January-March quarter. On a pretax basis, Alphabet recorded $39.7 billion in operating profit versus $37.7 billion in non-operating income, while Amazon posted $23.9 billion in operating profit versus $16 billion in non-operating income. By simple calculation, about 70% of the increase in profit came from outside core operations.

AI startups supporting cloud demand

Major tech companies are increasingly taking on the character of venture capital firms that profit from investing in promising private companies. Anthropic, SpaceX, and OpenAI are all looking toward initial public offerings (IPOs).

Anthropic's valuation, which was in the $4 billion range in 2023, is expected to soon reach about $900 billion. Amazon and Alphabet plan to make up to a combined additional $115 billion in investments in OpenAI and Anthropic in 2026, which could further expand their unrealized gains.

For companies that provide cloud infrastructure like Amazon and Google, AI startups are also major customers. According to The Information, roughly half of the combined $1.5 trillion in backlog held as of the end of March by the three cloud giants Amazon, Alphabet, and Microsoft was for OpenAI and Anthropic.

The pattern is becoming entrenched in which unlisted AI companies support cloud demand as major customers, while valuation gains at their investors lift the earnings of the big players. Massive investments are increasingly circling back mostly as cloud usage fees, creating a risk that AI demand is being inflated beyond reality.

Rising investment and fundraising

Nvidia, which has an edge in semiconductors for AI, has also expanded investments in startup cloud companies and AI firms that buy GPUs. Such money loops have long been pointed out, with some arguing they fuel overheated demand and make the underlying situation harder to assess.

Tech giants are continuing to invest in AI data centers while also using debt. Capital expenditure in 2026 by the four US tech giants, including Meta along with the major cloud providers, is expected to reach as much as $725 billion, up 76% from the previous year, nearly 120 trillion yen. In the January-March quarter, the four companies' investing cash flow exceeded operating cash flow from core business.

Large bond issuance is also spreading as a funding method. Meta issued bonds worth tens of billions of dollars in October 2025 and April 2026, while Amazon and Alphabet did so in November 2025. Alphabet is also planning its first yen-denominated bond issuance.

The method Meta began in 2025 is especially complex. Through a joint venture with U.S. investment firm Blue Owl Capital and others, it is building a new data center worth about $27 billion, which Meta will use by paying rent. Such financial engineering, which secures computing capacity while keeping debt down, is also spreading.

During the 2000 dot-com bubble, investment in telecom equipment and fiber-optic networks swelled, and vendor financing, which involved lending to or investing in suppliers, helped drive excessive investment. US big tech is leading the recent stock rally, but if the circular structure strengthens further, AI investment will be harder to slow. With warnings growing about market overheating, the risks of a reversal in early-stage investment are also likely to rise.

Enjoyed this article? Share it with your network!