AI adoption in the United States

US companies took 79% of global AI funding in 2025. McKinsey’s global survey finds use still shallow and returns going to a small group.

The numbers

  • 79%

    US share of global AI funding in 2025

    Crunchbase

  • 38%

    of all AI funding went to OpenAI, Anthropic and xAI

    $86.3 billion, CB Insights. A different data provider and total than the 79%

  • 80%

    of 2025 generative AI spending forecast to go to hardware

    Gartner forecast, March 2025

Show all figures
  • $159 billion

    AI funding that went to US companies in 2025

    Crunchbase

Reported AI use and high performers

  • of respondents use AI regularly in at least one function88%
  • are high performers who credit AI with 5% or more of EBIT6%

Both are shares of the respondents of McKinsey’s global survey, not of US companies. The second is given as about 6%.

I planned a country-by-country series on AI adoption, with one article per country and an overview for the EU. This is the first one, the United States.

The global surveys below provide context; they do not estimate adoption among US businesses.

What I found

  • The money is in the United States. Stanford’s 2025 AI Index puts US private AI investment at $109.1 billion in 2024. Crunchbase counted $159 billion of AI funding going to US companies in 2025, or 79% of the world total. CB Insights counts $86.3 billion raised by OpenAI, Anthropic and xAI that year, about 38% of all AI funding.
  • Use is wide but shallow. In McKinsey’s 2025 global survey, 88% of respondents report regular AI use in at least one function, and only about one in three have begun to scale it across the enterprise. In BCG’s surveys of software developers, more than 85% are at stages two and three of adoption and fewer than 10% have reached semiautonomous collaboration.
  • A small group gets the returns. About 6% of McKinsey’s respondents, its “high performers”, credit AI with at least 5% of earnings before interest and tax and report significant value. They are nearly three times as likely to redesign how the work is done.
  • The rules come from the states. The study found no comprehensive federal AI law, and the July 2025 AI Action Plan aims to speed up innovation and remove barriers. Colorado passed a broad act on high-risk systems in 2024 (SB24-205). Illinois requires notice and consent for AI video interviews. After the study, in May 2026, Colorado repealed its act before it took effect and replaced it with a narrower law (SB 26-189).
  • The copyright cases concern use of copyrighted material in training and, in some cases, model outputs. The New York Times sued Microsoft and OpenAI in the Southern District of New York (No. 1:23-cv-11195). Music publishers sued Anthropic over song lyrics (No. 3:23-cv-01092). Record companies sued Suno and Udio; RIAA announced the cases in June 2024.
  • The study argues that companies write internal AI policy from the lawsuit headlines, while the risk inside a company is what an agent may change in its systems and what record is left. It did not measure this.
  • Gartner forecast in March 2025 that about 80% of worldwide generative AI spending in 2025 would go to hardware, mostly devices with AI built in.

How it was made

I asked Perplexity Deep Research on 26 April 2026 and it showed 71 sources. The AI gathered the sources and drafted the study. I asked the question. The McKinsey, BCG and Gartner numbers are global, not US-only. Figures were current on that date, except the Colorado change, which was added later.

Sources

The question
How far has AI adoption gone in the United States, and how do the copyright lawsuits compare with the risks companies actually face?
How it was made
Perplexity Deep Research, 71 sources.