Here is a number for the pile marked “nobody predicted this”: more than 750,000 US jobs linked to AI have been created since 2023, according to LinkedIn estimates cited by The Wall Street Journal and recapped by The Kobeissi Letter on 4 October.
The surprise is not the total. It is the composition — because the biggest single category in the AI jobs boom is a job almost nobody had on their radar five years ago.
🔍 THE BOTTOM LINE: Data annotators lead the AI employment surge with roughly 282,000 positions created between 2024 and 2026 — the people who label and quality-check the data models learn from. AI job postings pay about $177,000 against roughly $80,000 for non-AI roles, and demand for AI literacy across ordinary jobs is up 70% year on year.
Where the 750,000 jobs actually are
Breaking down the LinkedIn estimates as reported:
- 282,000 data annotators — the largest single category, positions created between 2024 and 2026
- 117,000 data centre jobs — construction and operations of the AI buildout’s physical footprint
- 105,000 AI engineers — still growing, but no longer the headline
- Around 29,000 Head of AI roles and 15,000 forward-deployed engineers round out the picture
The mix tells a story: the AI economy’s biggest employer so far is neither the labs building models nor the companies automating workflows — it is the enormous human effort of teaching those models what they are looking at. This site covered the new-collar shift behind these numbers: LinkedIn counts 1.3 million new AI-enabled roles globally, and most do not require a degree. The annotator figure is that thesis arriving in bulk.
The pay gap is doing the talking
The typical AI job posting lists about $177,000 in annual compensation, against roughly $80,000 for a typical non-AI position — a premium of more than double. LinkedIn’s August research additionally found US AI job postings have roughly doubled since 2023, and “AI engineer” has overtaken “machine-learning engineer” as the platform’s most common AI occupation. VP of AI postings increased about sixfold.
The premium is spreading beyond specialists. LinkedIn reports US jobs requiring AI-literacy skills up 70% year on year — employers increasingly want workers who can use AI wherever they sit, not only those who build it. That mirrors the pattern in LinkedIn’s safety-roles data, where an AI-adjacent specialisation grew 91% in a year, and with McKinsey’s estimate of 11 million US career shifts by 2035, the demand side of that transition is now visible in pay and postings.
The caveat: the broader market is slow
The same week’s data puts September US payroll growth at just 29,000 — so the AI boom is hiring into a labour market that overall is not going anywhere fast. Two things can be true: hiring-not-firing is the adjustment channel, and the new AI-linked roles are concentrated in a handful of categories rather than spread across the economy. A worker in retail or transport is not going to become a data annotator by proximity alone — McKinsey’s work suggests nearly half of transitioning workers face an unclear path.
For job-seekers the composition matters more than the headline. The categories hiring in volume — annotation, data centre operations — are also the most accessible: annotation work in particular is entry-friendly, which is precisely why it appears in such numbers, and why researchers have dinged AI labs over its pay and conditions offshore. The $177,000 figure applies to the postings where it applies; it is not the annotator median. Read the composition, not just the total.
What it means for New Zealand
New Zealand will not see 750,000 of anything, but the shape travels. The accessible rungs of the AI economy here are the same: digital-technologies curricula are being rewritten around AI skills, and polytechnic-level data and AI programmes have appeared across the country without fanfare. The annotator lesson applies directly — the AI jobs wave’s biggest entry point so far is a job that rewards care and consistency rather than a machine-learning PhD, and AI-skills demand is already visible in NZ job ads. For career-changers weighing a start: the entry door into the AI economy is wider than the engineering headlines suggest, and it does not start with a degree.
❓ FAQ
How many jobs has AI created in the US? LinkedIn estimates cited by The Wall Street Journal put AI-linked US job creation at more than 750,000 since 2023, led by roughly 282,000 data annotator positions, 117,000 data centre jobs and 105,000 AI engineers.
What do AI jobs pay compared to other jobs? The typical AI job posting lists about $177,000 in annual compensation, against roughly $80,000 for a typical non-AI position, per LinkedIn data. The annotator category itself sits lower — it is entry-level work by design.
What is a data annotator? A person who labels, checks and quality-scores the data AI models learn from — images, text, audio. It is the largest single new job category of the AI boom, and it generally does not require a degree.
Do AI jobs require a technical degree? Increasingly, no. LinkedIn’s new-collar research counts 1.3 million new AI-enabled roles globally, and demand for AI literacy spread across ordinary job types is up 70% year on year in the US.
🔍 THE BOTTOM LINE
The AI jobs story of 2026 is not a handful of million-dollar lab researchers — it is three-quarters of a million people, most of them doing work that did not exist as a job title in 2021. The boom is real, the pay premium is real, and the biggest door in is labeled data: read the composition before you decide what the number means for your own career.
📰 Sources
- Blockonomi — “AI Job Boom Adds 750,000 U.S. Roles as Data Centers Fuel Hiring” (4 October 2026)
- The Kobeissi Letter — LinkedIn AI employment estimates (4 October 2026)
- The Wall Street Journal — LinkedIn estimates on AI-linked job creation (2026)
- LinkedIn — August 2026 research on US AI job postings, occupation mix and AI-literacy demand
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
