The skilled trades have spent the last three years being held up as the obvious answer to "what's safe from AI." That framing is mostly right, and the data backs it. But it's not uniformly right — and one trade in particular gets cited as durable that the data, and what's actually happening on U.S. highways in 2026, suggest is quietly more exposed than people realize.
This piece uses the same AI exposure dataset CertOutlook applies to healthcare credentials: peer-reviewed task-level substitutability scores mapped onto BLS Standard Occupational Classification codes. Lower numbers mean less exposure, more durable through 2030. We pull the trades from our profession set and rank them.
The trades exposure ranking
Wage data is from the BLS May 2024 OEWS. Growth projections are from the Occupational Outlook Handbook 2024–2034 projections.
| Trade | AI exposure | BLS growth (2024–2034) | Median wage (May 2024) |
|---|---|---|---|
| Welder, Cutter, Solderer | −1.20 (very low) | +2% (slower than average) | $51,000 |
| CDL Truck Driver (heavy + tractor-trailer) | −1.15 (very low) | +4% (as fast as average) | $57,440 |
| Plumber / Pipefitter / Steamfitter | −1.12 (very low) | +4% | $62,970 |
| HVAC Technician | −0.85 (low) | +8% (much faster than average) | $59,810 |
| Electrician | −0.78 (low) | +9% (much faster than average) | $62,350 |
| Barber | −0.73 (low) | +5% | $35,420 |
| Cosmetologist | −0.68 (low) | +5% | $35,250 |
What's most durable through 2030 — and why
The pattern is consistent with the healthcare exposure data: physical work performed in a non-standardized environment with real-time judgment is extraordinarily resistant to AI. Every trade in this list scores in the lowest AI-exposure tier. The differences within the list are small relative to the gap between any of these trades and a knowledge-work occupation.
Plumbers and HVAC technicians sit in the most defensible position. The work isn't just physical and judgment-heavy — it's done inside specific, irregular building environments where each job has unique constraints. There's no path to robotic plumbing that competes with a human plumber in your basement at consumer prices in this decade. Plumbers will be in demand in 2030, and they'll command a wage premium because the supply pipeline is producing graduates more slowly than retirements are removing tradespeople. The gap is not subtle: industry analyses show roughly five retirements for every two new skilled-trades entrants, and 2025 saw nearly 600,000 trade job postings against only ~150,000 new apprenticeship entries.
Welders and electricians are similar. Some welding (especially production welding in standardized factory environments) is more automatable, and some electrical work in industrial automation contexts has competing technology. But the median welder and electrician work in custom contexts where AI exposure is minimal. Specialized welding paths — particularly ASME-qualified pressure-vessel welding — command substantial wage premiums over general welding work.
Barbers and cosmetologists are quietly some of the most durable trades on the list. Personal services performed on humans, with continuous aesthetic judgment and conversation, are nearly impossible to automate at consumer prices. The wage band is lower than other trades, but the durability is exceptional. This is the trades equivalent of massage therapy on the healthcare side.
The CDL caveat — read this carefully (the picture has changed)
CDL truck driving scores in the lowest exposure tier in our dataset (−1.15). At face value, that suggests the trade is durable through 2030. Read the score in context, because the on-the-ground picture in 2026 is moving faster than the score implies.
The score reflects the aggregate task mix of a heavy/tractor-trailer driver: pre-trip and post-trip inspections, cargo securing, loading-dock interactions, route improvisation when GPS is wrong, and long-haul highway driving. The non-driving tasks are genuinely hard to automate. The long-haul highway portion is the part actually being displaced — and it's already happening on specific U.S. corridors:
- Aurora Innovation is running fully driverless commercial freight on Dallas–Houston and Fort Worth–El Paso, with a 1,000-mile driverless lane between Phoenix and Fort Worth, and plans for 200+ driverless trucks in operation by the end of 2026 — including night and adverse-weather routes.
- Kodiak Robotics has surpassed 10,700 revenue-generating driverless hours by early 2026, operating Atlas Energy Solutions' driverless fleet in the Permian Basin and routes between Dallas and El Paso.
- Gatik reported $600M in contracted driverless revenue at the start of 2026, with 60,000+ fully driverless orders completed since mid-2025 — daily Fortune-50 retailer deliveries across Texas, Arkansas, and Arizona, no safety driver.
- Plus.ai launched a commercial autonomous pilot in Texas with Ryder and International Motors.
- Waabi has been running supervised commercial deliveries on Uber Freight loads in Texas since 2023 and is launching fully driverless service in 2026.
- California adopted regulations in April 2026 permitting heavy-duty driverless deployment, lifting its prior ban — meaningful because California is the largest U.S. freight market.
The setbacks people remember (TuSimple shutting down U.S. operations in late 2023, Embark being acquired by Applied Intuition in 2023, Waymo Via stepping back from trucking) read like the technology is dying. They aren't — the surviving players consolidated and accelerated. The "five years away" framing has been replaced by specific corridor deployments and named contracted-revenue numbers.
The honest read for someone considering a CDL credential in 2026: the credential still pays back fast (often in under 12 months), the labor market through 2027–2028 is robust, and regional/last-mile/specialty work is much less exposed than long-haul. But long-haul tractor-trailer work on Sun Belt corridors is being displaced now, not in 2030. If you're 25 today and planning a 30-year long-haul career, you're planning around a job description that won't exist in its current form for that duration.
The actionable framing: pursue CDL with eyes open, plan to migrate toward regional/last-mile/specialty (HAZMAT, oversize, fuel, refrigerated, intermodal drayage) work over the medium term, and avoid building a long-term plan around long-haul on the Texas/Arizona/Southwest corridors as a destination.
Where the upward path lives in each trade
Low exposure makes a credential durable, but doesn't make it a high-ceiling career. The trades vary substantially on upward optionality.
- Plumber and electrician have the strongest upward paths. Apprenticeships typically run 4–5 years to journeyman; another 1–4 years of experience (varies by state) opens master licensure. Master tradespeople in major metros routinely clear $100,000+, and the path to ownership of a small contracting business is well-trodden.
- HVAC tech has a moderate upward path. EPA 608 certification has four levels — Type I (small appliances), Type II (high-pressure), Type III (low-pressure), and Universal — plus specialty work in commercial refrigeration and industrial process cooling. The growth rate (+8%) is the strongest in the trades.
- Welder upward path is concentrated in specialty welding (underwater, pipeline, aerospace, ASME-qualified pressure-vessel) where wages can substantially exceed the median. The general welder labor market is more competitive than the specialty market.
- Barber and cosmetologist upward paths are largely entrepreneurial — owning your chair, building a client book, opening a shop. Wage growth as a salaried employee is limited; ownership transitions are where the income lives.
- CDL driver upward path is via specialty (HAZMAT, oversize, refrigerated, fuel, intermodal drayage) and via owner-operator status, both of which carry their own risk profiles — and both are meaningfully less exposed to autonomous displacement than general long-haul.
The 2030 trades thesis
If you want fastest credential payback + acceptable durability: HVAC Technician (12–18 month programs) or Welder (6–18 months). Both pay back quickly with reasonable durability and clear specialization paths.
If durability matters more than wage: Barber or Cosmetologist. Lowest exposure, modest wages, ownership path is real.
If you're considering CDL: pursue it for the fast payback and 2027-2028 labor market, plan the medium-term migration to regional, last-mile, or specialty work, and don't build a 2035 plan around long-haul on Sun Belt corridors. The displacement isn't theoretical anymore.
Compared to healthcare
The trades are, in aggregate, more AI-resistant than healthcare credentials. The lowest-exposure trade in our dataset (welder at −1.20) sits near massage therapy (−1.41) on the healthcare side; the highest-exposure trade is more durable than the most-exposed healthcare credentials.
That's a real signal, but it doesn't mean the trades are unconditionally a better credentialing bet — wage trajectory, work conditions, and physical demands across a 30-year career all factor in. For a side-by-side, see trades vs healthcare: which fits you. For the parallel healthcare AI exposure analysis, see healthcare certifications most and least exposed to AI through 2030.
For state-specific wage data on any of these trades, see the CertOutlook state hubs — wages vary substantially across markets, and the highest-paying state for any given trade is rarely the most populous.