# AI Created 1 Million US Jobs: What It Means for Swiss SMEs

> Author: Chris Jon Graf (AI Strategist & CEO)
> Updated: 2026-09-08
> URL: https://ai-outsourcing.ch/insights/ai-created-1-million-us-jobs-what-it-means-for-swiss-smes

## Summary

An Economist analysis finds AI has already created roughly 1 million new jobs in the US - far more than the roughly 200,000 AI-related layoffs. The driver is the build-out of AI infrastructure. For Swiss SMEs, the real bottleneck isn't technology - it's skilled labour.

The debate about AI and jobs is usually framed around the wrong question: does AI destroy jobs? A September 2026 analysis by The Economist gives a surprisingly clear answer for the US labour market: AI has already created around 1 million new jobs there, against only about 200,000 AI-related layoffs over the same period. The net effect is clearly positive - just not where most people expected it: not in software development, but in skilled trades and infrastructure.

## The numbers that flip the debate

**1,000,000** — new jobs created in the US since the AI boom, according to The Economist (September 2026)

**200,000** — AI-related layoffs over the same period - far fewer than the jobs created

These figures contradict the common assumption that AI leads to broad-based job losses. What is true: individual roles are changing radically or disappearing. What is false is the idea that this automatically produces a net loss of employment. The reason lies in a second, often overlooked effect of the AI revolution: the physical build-out of the infrastructure that makes it possible in the first place.

## Where the new demand comes from: the infrastructure boom

Since ChatGPT launched in 2022, infrastructure investment in the US has risen by around $500 billion per year, according to Goldman Sachs. A large share of that flows into data centre construction, which now runs at roughly $75 billion annually - up about 60 percent year on year. Every one of these data centres needs power, cooling, wiring and a building - and therefore people to build and maintain it.

**$500B / year** — additional US infrastructure investment since ChatGPT launched in 2022, according to Goldman Sachs

## Where the jobs are actually appearing

- Electricians and electrical planners for data centre power and wiring
- Construction workers and site supervisors for the physical build-out
- HVAC specialists for cooling and climate systems
- Data centre technicians for operations and maintenance
- Utilities specialists for grid expansion and capacity

The scale is significant. McKinsey estimates that building out AI infrastructure in the US alone will require roughly 130,000 additional electricians, 240,000 additional construction workers and 150,000 additional supervisors by 2030. Since 2023, Bureau of Labor Statistics data shows around 320,000 new jobs created in electrical work, construction, HVAC and utilities. These aren't 'AI jobs' in the traditional sense - but they exist only because AI does.

## The real bottleneck: people, not algorithms

> We need hundreds of thousands of electricians, plumbers and carpenters.
>
> — Jensen Huang, CEO, Nvidia

This comment from Nvidia CEO Jensen Huang captures the problem precisely: the bottleneck of the AI revolution isn't compute or capital, it's skilled trades labour. The Associated Builders and Contractors estimate that the US will be short around 349,000 construction workers in 2026, with that gap widening to 450,000 in 2027. Meta has already responded with a $115 million training programme designed to train 5,000 construction workers per year. The Center for Strategic and International Studies (CSIS) projects up to 140,000 additional skilled trades jobs by 2030 in its high scenario, driven by generative AI alone.

**349,000** — construction workers the US is projected to be short in 2026, rising further in 2027 (Associated Builders and Contractors)

## What this means for Swiss SMEs

The numbers differ in Switzerland, but the pattern doesn't. According to jobs.ch, around 44 percent of Swiss jobs are AI-exposed in 2026, and entry-level positions have already dropped by roughly 32 percent. That sounds alarming at first - but it mainly signals that work is shifting, not disappearing. Companies that read this shift correctly, rather than fearing it, recognise that the real bottleneck isn't technology - it's skilled people with the right capabilities, just as in the US.

This matches what Swiss companies are already observing. According to PwC Switzerland (2026), headcount at companies actively using AI is growing by around 52 percent, compared to 36 percent at companies without systematic AI adoption. EY Switzerland finds a similar pattern: 89 percent of surveyed companies use AI, yet only 7 percent are cutting jobs as a result - 18 percent are creating new ones. This pattern is known as people amplification: AI multiplies the impact of existing teams rather than replacing them.

> **People amplification, not headcount cuts**
>
> The companies with the strongest results aren't using AI to cut headcount - they're using it to get more impact from the same people, and reinvesting the freed-up capacity into new, often technical, roles.

## What Swiss decision-makers should do now

1. Map skills gaps systematically: which roles are AI-exposed, and where are technical and skilled-trades shortages already emerging?
2. Plan automation and upskilling together, instead of rolling out automation in isolation and leaving training to chance
3. Secure technical and skilled-trades capacity early - through apprenticeships, career-changer programmes and partnerships with vocational schools
4. Communicate a clear people-amplification strategy internally to actively reduce uncertainty among staff

For a closer look at how different economies are positioning themselves in the global AI race - and the strategic mistakes companies most commonly make - the episode [AI strategy for Swiss SMEs: the three biggest mistakes in the global race](https://www.ki-podcast.ch/ki-standort-schweiz-kmu-strategie-und-globales-rennen) from the Swiss AI podcast is worth a listen.

## Bottom line: the bottleneck is skills, not technology

The US numbers reveal a pattern that is already emerging in Switzerland too: AI destroys fewer jobs than it creates, but it shifts demand toward exactly the skills that are already scarce. For Swiss SMEs, that means one thing: companies that invest now in both automation and upskilling their people will gain an edge while others are still debating the wrong question.

## FAQ

### Does AI actually create more jobs than it destroys?

According to a September 2026 analysis by The Economist, AI has already created around 1 million new jobs in the US, compared to roughly 200,000 AI-related layoffs over the same period. The net effect is clearly positive, with many of the new roles located outside traditional tech occupations.

### Which occupations benefit most from the AI boom?

Demand is growing fastest in occupations needed for the physical build-out of AI infrastructure: electricians, construction workers, HVAC specialists, data centre technicians and utilities workers. McKinsey estimates the US alone will need around 130,000 additional electricians and 240,000 additional construction workers by 2030 for this build-out.

### Is the situation in Switzerland comparable to the US?

The scale differs, but the pattern is similar. According to PwC Switzerland, companies actively using AI grow headcount significantly faster (around 52 percent) than companies without systematic AI adoption (around 36 percent). EY Switzerland finds that only 7 percent of AI-using companies cut jobs as a result, while 18 percent create new ones.

### What does 'people amplification' mean?

People amplification describes using AI to multiply the impact of existing employees rather than cutting headcount. Companies following this approach show significantly stronger headcount growth in Swiss surveys than companies using AI primarily to cut costs.

### How should Swiss SMEs respond to the shortage of technical skilled labour?

The key is planning automation and targeted upskilling as one combined effort rather than two separate initiatives. Companies that invest early in technical and skilled-trades capabilities gain an advantage before competition for this talent intensifies further.

## Sources

- [The Economist: The jobs apocalypse is postponed. An AI jobs boom is here](https://www.hindustantimes.com/world-news/the-jobs-apocalypse-is-postponed-an-ai-jobs-boom-is-here-101788577683986.html)
- [The AI economy is rewriting the American Dream (CNBC, Mai 2026)](https://www.cnbc.com/2026/05/19/ai-hiring-slowdown-skilled-trade-workers.html)
- [AI companies recruiting electricians and carpenters (Seattle Times, Juli 2026)](https://www.seattletimes.com/business/ai-companies-are-recruiting-electricians-and-carpenters-by-the-thousands/)
- [Jensen Huang says electricians and plumbers will be needed by the hundreds of thousands (Fortune, Juni 2026)](https://fortune.com/article/nvidia-billionaire-ceo-jensen-huang-demand-for-gen-z-skilled-trade-workers-electricans-plumbers-carpenters-data-center-growth-six-figure-salaries/)
- [How Will AI Affect the US Labor Market? (Goldman Sachs, März 2026)](https://www.goldmansachs.com/insights/articles/how-will-ai-affect-the-us-labor-market)
- [GenAI's Human Infrastructure Challenge (CSIS, Sept 2025)](https://www.csis.org/analysis/genais-human-infrastructure-challenge-can-united-states-meet-skilled-trade-labor-demand)
- [LinkedIn Post: Alvin Foo (gewählter Source-Post)](https://www.linkedin.com/posts/alvinfsc_the-economist-finds-ai-has-already-added-activity-7502976545246846976-GwWp)