VeltrixVeltrix.
← All articles
02 / 62March 10, 2026

The Jobs Reckoning

92 million roles displaced, 170 million created by 2030. The net number is positive. The transition is not.

Which jobs AI is replacing — and when.

92M
jobs projected displaced by 2030 WEF
14%
of workers already displaced or significantly affected NU
60%
of jobs will see significant task-level changes by 2030 NU
Customer ServiceCritical risk
Customer service rep SSRN
80%
2.24M of 2.8M US jobs at risk
IBM AskHR: 11.5M calls, <5% human oversight
AdministrativeCritical risk
Data entry clerk SSRN
95%
AI processes 1,000+ docs/hr
vs 2–5% human error rate
Legal supportHigh risk
Paralegal / legal researcher DS
80%
Paralegals by 2026
Legal researchers: 65% by 2027
FinanceHigh risk
Banking / financial processing GS
54%
200,000 Wall Street roles at risk
by 2027–2030 (Bloomberg)
RetailMedium risk
Cashier / checkout DS
65%
Walmart: 8,000 roles
Sam's Club: 12,000 roles
ManagementLower risk
Mid-level management DS
50%
By 2026, 20% of orgs will use AI
to flatten hierarchy (Gartner)
Timeline of impact: who gets hit and when
Now → 2026Entry-level white-collar
NOW
2025 → 2027Admin & data roles
ACTIVE
2027 → 2030Mid-level knowledge work
2027+
2028 → 2030Transport / logistics
2028+
2030+Complex cognitive / expert
2030+
Goldman Sachs nuance GS: If AI use cases were expanded across the full economy today, only 2.5% of US employment would face immediate displacement. The bigger near-term effect is hiring slowdowns and task shifts rather than mass layoffs — but the rate is accelerating.
So what does this mean?

92 million jobs displaced by 2030 — and it's already started. Entry-level white-collar roles are being hit right now. If your job involves repetitive data processing, customer queries, or routine admin, the timeline isn't "someday." It's this year.

But "displacement" doesn't always mean "fired." For most people, it means your role is changing. The tasks AI can do will be handed to AI. The question is whether you're the person directing it — or the person being replaced by it.

Jobs AI is building that didn't exist 5 years ago.

Net global job impact by 2030 WEF SSRN
92M
Jobs displaced
vs
170M
New jobs created
+78M
Net gain in global jobs by 2030. But: they won't be in the same places, industries, or skill levels.
RoleGrowth trajectoryMedian salary
AI / ML Engineer VERFastest growing AI title
+143% YoY demand · +41.8% quarterly
$206K
avg 2025 salary
Prompt Engineer PLRole didn't exist 3 years ago
+135.8% demand surge · 32.8% CAGR to 2030
$123K
avg · up to $335K
Data Scientist (AI-focused) VERTraditional role, AI-supercharged
+10% YoY · 58,263 open roles (data/AI combined)
$170K
median 2025
Chief AI Officer FRNew C-suite role emerging fast
90% of companies creating new AI roles (CIOs/CTOs)
$352K
avg CAIO · up to $644K
AI Ethics Officer SSRNRegulation-driven growth
Part of ~350K new AI-specific roles projected
$130K+
est. range
AI-Augmented MD / Clinician NUHealthcare: AI augments, not replaces
Nurse practitioners +52% projected 2023–2033
Growing
demand + premium
The skills gap warning: 77% of new AI jobs require a master's degree or higher SSRN. The new jobs won't automatically absorb the workers whose old ones disappear. Geographic concentration is also an issue — new AI roles cluster in tech hubs, not the towns where displacement is highest.
So what does this mean?

The net numbers look positive: 170M new jobs vs 92M lost. But these aren't a like-for-like swap. The new roles pay more, demand different skills, and concentrate in different places.

You don't need to become a machine learning engineer to benefit. But you do need to become AI-literate. The people who learn to work alongside AI — in whatever field they're in — will have access to the new roles. The ones who don't will be competing for a shrinking pool of unchanged positions.

"AI won't replace humans — but humans with AI will replace humans without AI." The data increasingly backs this. The wage premium for AI skills doubled in a single year. Productivity in AI-exposed industries is growing four times faster. The gap is accelerating.

Wage premium: same job, different AI skill level PwC
No AI skills
Base
Same job, no AI skill requirement. Reference point. This was the norm in 2022.
Some AI skills
+25%
The 2024 baseline. Even basic AI literacy commanded a premium over peers.
2025 reality
Strong AI skills
+56%
Premium doubled in one year. PwC analysed close to 1 billion job ads across 6 continents.

Average worker with AI skills earns $18,000+ more per year than an equivalent peer without them (Lightcast, 1.3B job postings) IG. Premium exists in every industry and region analysed — not just tech.

Financial servicesMost AI-exposed
Pre-AI
Now
27%
Software publishingHigh AI exposure
Pre-AI
Now
~27%
Mining / hospitalityLeast AI-exposed
Pre-AI
Now
9%

Pre-AI baseline: 7% productivity growth (2018–2022). Post-AI: 27% in exposed industries. 3x higher revenue per employee in most-exposed vs. least-exposed sectors.

"AI literacy has become as important as computer skills were in the 1990s. The difference is this transition is happening in months, not years."
— Synthesis of McKinsey, PwC & WEF research, 2025
66%
faster skill change in AI-exposed jobs vs. last year's 25% rate PwC
39%
of current skill sets will be outdated or transformed by 2030 IDX
75%
of employers plan to reskill or upskill workforces for AI collaboration PwC
83%
of companies say AI-skilled employees are more likely to retain their jobs H5
49%
of US companies using ChatGPT have already replaced some workers with it NU
$19.9T
AI contribution to global economy projected by 2030 GS
So what does this mean?

Workers with AI skills earn 56% more than their peers doing the same job. That premium doubled in 12 months. This isn't a future trend — it's the current reality across every industry PwC analysed.

You don't need to switch careers. You need to add AI to the career you already have. The data is unambiguous: the single biggest thing you can do for your earning power and job security right now is become competent with AI tools.

5 things you can do this week
to get ahead of the curve.
1.

Subscribe to stay informed. The landscape is shifting weekly. Sign up to Veltrix Collective and get one practical AI use case every Tuesday — no jargon, just tools and techniques you can apply immediately. Knowing what's changing is half the battle.

2.

Start using whatever AI your company already has. If your workplace has Copilot, use it. If there's a custom internal tool, learn it. If neither — open ChatGPT, Claude, or Gemini and start asking it to help with real work tasks: drafting emails, summarising documents, analysing data, brainstorming ideas. The goal is daily reps, not perfection.

3.

Learn prompt engineering and system prompts. This is the skill that separates "I tried AI and it was meh" from "AI doubled my output." Learn how to write clear instructions, provide context, and shape AI responses. If you have access, start building your own custom agents or GPTs tailored to your specific workflows.

4.

Get hands-on with AI tools at home. Download Claude Code and use it for small personal projects — automating file organisation, writing scripts, building simple apps. Try n8n or Make for workflow automation. The skills you build on personal projects transfer directly to work.

5.

Pick one part of your job and AI-augment it. Don't try to overhaul everything at once. Choose the most repetitive, time-consuming task in your week and figure out how AI can handle 80% of it. Then move to the next one. Small wins compound fast — and they're the evidence you need when the next performance review or promotion comes around.

The data is clear: a 56% wage premium, 4x productivity growth, 92M jobs shifting. The people who start now will be the ones who thrive. The rest will be playing catch-up. Which side do you want to be on?

Source references

WEF
World Economic Forum — Future of Jobs Report 202592M displaced / 170M created / +78M net by 2030. 41% of employers plan workforce reductions.weforum.org →
PwC
PwC 2025 Global AI Jobs Barometer — ~1 billion job ads, 6 continents56% wage premium for AI skills (up from 25%). 4x productivity growth. Skills changing 66% faster.pwc.com →
GS
Goldman Sachs Research — AI Workforce Analysis (Aug 2025)6–7% US job displacement baseline. 2.5% at immediate risk today. $19.9T GDP contribution by 2030.goldmansachs.com →
SSRN
SSRN — AI Job Displacement Analysis 2025–2030 (Nartey)Role-level automation rates. 350K new AI roles. 77% require master's. 80% customer service risk.ssrn.com →
NU
National University — 59 AI Job Statistics (May 2025)14% already displaced. 60% task-level changes. 49% of US ChatGPT users replaced workers.nu.edu →
DS
DemandSage — 77 AI Job Replacement Statistics 2026Sector-level risk data. Retail, legal, finance displacement projections.demandsage.com →
VER
Veritone — Q1 2025 AI Labor Market AnalysisMedian AI role salary $156,998. AI/ML Engineer +41.8% YoY. 35,445 AI positions Q1 2025.veritone.com →
PL
PromptLayer — AI Prompt Engineering Jobs 2025$123K avg salary. $90K–$335K range. 32.8% CAGR to 2030.promptlayer.com →
FR
FinalRoundAI — 15 Highest Paying AI Jobs 2025CAIO avg $352K. AI Engineer avg $206K (+$50K YoY).finalroundai.com →
IG
The Interview Guys — Analysis of 15 major studies on AI wage premiums (2025)Cross-verified 2B+ job postings. 19–56% premium range. Premium doubled in 12 months.theinterviewguys.com →
IDX
Index.dev — AI Job Growth Statistics 202639% skill sets outdated by 2030. Revenue growth 27% in AI-exposed vs 8.5% least exposed.index.dev →
H5
High5Test — AI Replacing Jobs Statistics 202583% of companies link AI skills to job retention.high5test.com →
Veltrix Collective
The window to adapt
is now open

Weekly signal on AI's impact on work, careers and organisations. What's actually happening, backed by data. No hype. Join readers staying ahead.

Weekly, every Tuesday · No spam · Privacy policy · Unsubscribe anytime
Data synthesis March 2026. All figures are approximate, drawn from cited sources above. Job displacement projections vary significantly across methodologies — treat ranges as directional. The wage premium data (PwC) is based on job ad analysis, not matched individual data, and should be interpreted accordingly. "Net positive" job creation figures are global aggregates and do not reflect local or sectoral concentration of losses.
Written by Luke Madden, founder of Veltrix Collective. Data synthesis and analysis by Vel.