The measurable pattern in 2026 is concentration, not AI taking over. Stanford's AI Index 2026 reports that industry built more than 90 percent of the notable frontier models in 2025, that the United States hosts 5,427 data centers (more than 10 times any other country), and that almost every leading AI chip is made by one company, TSMC. Whether AI itself could one day act against human wishes is a separate and still open question; the data here concerns who holds the power today.
Key takeaways
- "Dominate" can mean AI controlling people, a few firms or countries controlling AI, or AI spreading through daily life, and each has different evidence.
- Control of frontier AI is concentrated in industry, in the United States and China, and in a very small number of chip and data center suppliers.
- Use is growing but far from universal: 19.2 percent of Canadian businesses used AI in the 12 months to spring 2026, up from 6.1 percent two years earlier.
- Capability is uneven. A model can win a gold medal at a math olympiad and still misread analog clocks about half the time.
- Public concern is rising: 52 percent of US adults in a June 2026 survey were more concerned than excited about AI.
Three meanings of "dominate"
The question mixes at least three claims. The first is that AI systems could gain control over people. That claim concerns future capabilities, and it has its own literature that this article does not repeat. The second is that a handful of companies and governments could use AI to hold outsized power. That one can be checked against ownership, chips, and money today. The third is that AI could dominate daily life and work by sheer spread. That can be checked against adoption surveys. Mixing the three leads to answers that fit none of them, so the sections below take the second and third, where numbers exist.
Where frontier AI is concentrated
The AI Index 2026, published by the Stanford Institute for Human-Centered AI in April 2026, gives several markers. Industry produced over 90 percent of notable frontier models in 2025. The United States hosts 5,427 data centers, and TSMC fabricates almost every leading AI chip, so physical supply runs through a few places and firms. U.S. private AI investment reached USD 285.9 billion in 2025, more than 23 times China's USD 12.4 billion. Yet the model gap has narrowed: as of March 2026, the top U.S. model led the top Chinese model by only 2.7 percent, and China held 85 public or public-private AI supercomputers, the most of any country.
These figures describe concentration among a few actors. They do not show that those actors act together, or that they control what the models do after release.
Dominance indicators at a glance
The table gathers the main figures from the two sources used here. The reading in the last column is an interpretation, not a finding of either source.
| Indicator | Latest figure | What it can and cannot show |
|---|---|---|
| Share of notable frontier models built by industry, 2025 | Over 90 percent | Shows who builds the models, not who benefits from them |
| Data centers in the United States | 5,427, more than 10 times any other country | Shows where computing capacity sits |
| Lead of top U.S. model over top Chinese model, March 2026 | 2.7 percent | Shows a close race, so one country does not hold a durable lead |
| Generative AI adoption, population share, within three years | 53 percent globally; 28.3 percent in the United States (24th) | Shows spread, but the measure is use, not dependence |
| Canadian businesses using AI, 12 months to spring 2026 | 19.2 percent; 40.0 percent say AI is not relevant to them | Shows most firms are not yet users |
How far AI has spread in Canada and the US
The Statistics Canada survey, collected from April 1 to May 6, 2026, found that 19.2 percent of businesses had used AI to produce goods or deliver services in the preceding 12 months. The share was 12.2 percent a year earlier and 6.1 percent two years earlier. Information and cultural industries led at 42.3 percent, followed by finance and insurance at 40.4 percent. Statistics Canada says the effect of AI on employment remains unclear. In the United States, the AI Index puts generative AI use at 28.3 percent of the population, ranking 24th among the countries measured, well behind the global figure of 53 percent.
Fast growth from a low base is not domination. It does mean that decisions made now about procurement, data, and oversight will shape how many people depend on a small set of providers.
Why capability does not equal control
The AI Index describes capabilities as uneven. One model earned a gold medal at the International Mathematical Olympiad, yet the top model reads analog clocks correctly only about 50 percent of the time. Such gaps matter when judging claims of near-total takeover, because control over the physical world, institutions, and legal systems takes more than strong test scores. They also mean that tools which look reliable in one area can fail in another, so human review stays necessary.
What the public thinks
A Pew Research Center survey of 3,488 US adults, fielded June 22 to 28, 2026, found 52 percent more concerned than excited about AI in daily life, up from 37 percent in 2021. Nine percent were more excited, and 37 percent equally both. Among adults under 30, 55 percent were more concerned than excited. The AI Index adds that US respondents showed the lowest trust in their own government to regulate AI of any country surveyed, at 31 percent, and that 73 percent of experts expect a positive effect on how people do their jobs, against 23 percent of the public.
This article presents general information drawn from surveys and reports. Forecasts about AI remain uncertain, and no source cited here predicts a takeover.
FAQ
Is any single company or country in control of AI?
Not completely. Industry builds most frontier models and the United States leads in investment and data centers, but the 2.7 percent gap between top U.S. and Chinese models in March 2026 shows a close contest. Chip manufacturing is more concentrated than model building.
Will AI take over most jobs soon?
The sources used here do not show that. Statistics Canada reports that 19.2 percent of businesses used AI and calls the employment effect unclear. Job-specific exposure is a different measure from job loss.
Can AI act against human wishes?
This article does not assess that. The question concerns future systems and is covered by international safety reports and expert surveys that this article does not summarize.
Why do experts and the public see AI differently?
The AI Index records a 50-point gap on job effects, with 73 percent of experts and 23 percent of the public expecting a positive effect. The reasons were not measured in the figures cited here.
Related reading
- Why no single robot is called best
- Documented benefits of AI
- How deep learning works
- Major AI systems in use
More in AI Policy and Ethics.
Last verified: October 2026
Sources
- The 2026 AI Index Report, Stanford Institute for Human-Centered AI, April 2026
- Inside the AI Index: 12 Takeaways from the 2026 Report, Stanford HAI, April 13, 2026
- Analysis on artificial intelligence use by businesses in Canada, second quarter of 2026, Statistics Canada, June 11, 2026
- Young adults in the US are increasingly wary of AI, concerned it will take jobs, Pew Research Center, August 18, 2026