|
Issue 17
|
|
The weekly brief AI didn’t shrink the job market. It rerouted it into hard hats.This issue looks at where the AI economy’s real hiring is happening right now — not in prompt engineering, but on job sites — and what that means for how you read every “AI jobs” headline from here on. September 14, 2026 · 6 min read
|
Sourced from The Economist, PwC, WRITER, and U.S. Census Bureau figures. | No hype Every number here is sourced, dated, and attributed. | Why now September data just confirmed where AI hiring actually lands. |
|
If you read nothing else The Economist estimates AI has created about one million new U.S. jobs since 2023 — nearly five times the roughly 200,000 layoffs it’s been blamed for. But almost none of that growth is in tech: it’s electricians, HVAC technicians, and grid engineers building the data centers AI runs on, with data-center construction spending now topping $75 billion a year. If you’re waiting for a job literally titled “AI,” you’re reading the boom from the wrong seat. |
|
|
Nova’s signal The most in-demand “AI skill” in America right now doesn’t show up on a resume — it’s a journeyman electrician’s license. Nobody’s LinkedIn headline says that yet. Nova’s call. Stop scanning job boards for the word “AI.” Start scanning them for “data center,” “grid,” and “commissioning.” Last issue we tracked one company’s deployment push; this week we zoom out to the whole labor market. Story 02 covers why more layoffs are blaming AI even as executives privately say it changed nothing, 03 breaks down the enterprise disappointment numbers behind those layoffs, 04 has the specific skills actually paying a premium, and 05 covers the widening split between companies grooming an “AI elite” and everyone else.
|
|
This week 5 developments worth your attention.
|
|
01 / Labor Market AI didn’t shrink the job market — it rerouted it into concrete and copperThe Economist estimates AI has added roughly one million jobs to the U.S. economy since 2023, comfortably outpacing the roughly 200,000 layoffs attributed to it over the same period. Nearly all of that net growth, though, is showing up in construction and the skilled trades that build and power data centers — electricians, HVAC technicians, and grid engineers — not in software roles. Data-center construction is running north of $75 billion a year, up nearly 60% from a year ago per Census Bureau figures, and employment across the five industries most tied to that build-out has grown roughly 320,000 more than general construction trends alone would predict. The Associated Builders and Contractors trade group says the industry will need nearly half a million new workers in 2027, up from 349,000 this year. Why it matters. If your mental model of the “AI jobs boom” is prompt engineers and ML researchers, you’re missing where the actual net job creation is happening — and where the leverage is for anyone willing to get a trade certification instead of a bootcamp certificate. The signal. Data-center electricians with liquid-cooling and medium-voltage experience are reportedly clearing up to $300,000 with overtime in hot markets like Northern Virginia; the median base for the role is closer to $94,500. Market Research Media, on The Economist’s jobs estimate →
|
|
02 / Layoffs 56% of 2026 layoffs now blame AI — but the verification doesn’t hold upMore than half of layoff events tracked in 2026 explicitly cite AI, automation, or machine learning as a reason — up sharply from roughly 5% of all cuts in 2025. The same companies leading the cuts, Alphabet, Microsoft, Meta, and Amazon, are on track to spend close to $700 billion combined on AI infrastructure this year. But a National Bureau of Economic Research working paper found 90% of executives say AI has had zero measurable employment impact at their own company, and separate reporting says nearly six in ten companies admit they frame layoffs as “AI-driven” when the real reason is financial. OpenAI’s Sam Altman has called some of this “AI washing.” Why it matters. A layoff headline that says “AI” is doing work for the company’s narrative that may have nothing to do with what actually happened on the org chart — treat every AI-blamed layoff as a claim to verify, not a data point to trust. The signal. Challenger, Gray & Christmas, the firm that actually tracks layoff announcements, puts AI’s cited share closer to 17–26% — well below the 56% figure making headlines, depending on methodology. The Interview Guys, on the AI-layoff verification gap →
|
|
03 / Enterprise AI Half of enterprises now call their own AI rollout a disappointmentWRITER’s second annual enterprise AI survey, of 2,400 C-suite leaders and employees across the US, UK, Ireland, Benelux, France, and Germany, found 48% now call gen-AI adoption a disappointment, up from 34% a year earlier. Only 29% of leaders report real ROI from generative AI, and just 23% from AI agents. Three-quarters of executives admit their company’s AI strategy is “more for show” than actual internal guidance, and 39% say they have no formal plan to turn AI tools into revenue. Why it matters. The gap between AI spending and AI results is where budgets — and jobs — move next; a widening disappointment number usually shows up a few quarters later as either a re-org or a vendor swap. The signal. 79% of executives admit to struggling with lagging ROI, strategy gaps, or internal power struggles around their AI rollout — this isn’t a minority problem, it’s most companies. WRITER, 2026 Enterprise AI Adoption survey →
|
|
04 / Skills & Pay Jobs needing AI skills are growing 8x faster than the market — and paying a 62% premiumPwC’s 2026 Global AI Jobs Barometer, built on more than a billion job postings across six continents, found roles requiring specific AI skills grew 69% year over year, against 9% for the overall jobs market — nearly eight times faster. The average wage premium for those skills has climbed to 62%, and total AI-related job postings are now almost double where they stood in 2024. Growth is concentrated in tech, media, and telecom (11%) and professional services (6%), and PwC notes employers are weighting judgment, creativity, and leadership more heavily as pure AI-tool fluency becomes table stakes. Why it matters. The wage premium isn’t paying for knowing how to use a chatbot anymore — that’s assumed. It’s paying for people who can apply AI output with judgment inside a specific domain. The signal. If a job posting asks for a specific AI skill plus three-plus years of domain experience, that combination is exactly the intersection PwC’s data says is compounding fastest. PwC, 2026 Global AI Jobs Barometer →
|
|
05 / Workplace 92% of the C-suite is grooming an “AI elite.” 60% plan to cut everyone else.The same WRITER survey found 92% of C-suite executives admit they’re actively cultivating a group of “AI elite” employees, while 60% say they plan layoffs targeting workers who don’t adopt AI tools. Employees WRITER calls “AI super-users” save nine hours a week — 4.5x more than laggards — and are 3x more likely to have gotten both a promotion and a raise in the past year. Separately, 77% of executives say employees who don’t become AI-proficient won’t be considered for promotion going forward. Why it matters. This is no longer a hypothetical “adapt or get left behind” talking point — it’s now a stated management practice at a majority of large employers, with numbers attached. The signal. 29% of employees in a related survey say they’re already quietly sabotaging their company’s AI rollout — the backlash to being sorted into a lower tier is starting to show up as passive resistance, not just attrition. WRITER / BusinessWire, 2026 AI Adoption survey →
|
|
Signal & chatter A few more things worth knowing.Robotic technicians up 107%. Between 2022 and 2026, demand for robotic technicians grew 107%, HVAC/cooling engineers 67%, and industrial automation technicians 51% — all feeding the same data-center build-out as story 01. Explore → | Data-center PMs are getting cold-called weekly. The sector is on track for 650,000 permanent data-center positions in 2026, with an estimated 340,000 potentially unfilled; senior project managers on hyperscale programs can clear $290,000 in total comp. Explore → |
|
|
From the community This week’s color comes from trade press and network coverage rather than a single Reddit thread — the electrician boom hasn’t generated the kind of canonical forum thread that, say, AI-training gig work has in past issues. Worth reporting as what it is: on-the-record quotes, not crowd sentiment. TV host Mike Rowe says he personally met three electricians under 30 earning $240,000 to $280,000 a year on AI data-center builds — a figure several outlets have now repeated. Treat it as the ceiling, not the floor: the median data-center electrician salary is closer to $94,500, and the six-figure outliers usually come loaded with overtime, medium-voltage certification, and a hot local market. DC Geeks, on what data-center electrician pay actually looks like →
|
|
FOWL prediction 17 By the second half of 2027, at least one major trade union or apprenticeship program will start explicitly marketing itself as “the AI jobs pipeline.”Nearly half a million new construction workers are needed in 2027, and almost nobody under 25 associates “electrician” with “the AI economy.” The instant one apprenticeship program or union local realizes it’s sitting on the actual AI jobs boom and rebrands around it, competitors follow — the same way coding bootcamps rebranded around “tech jobs” a decade earlier. FOWL AI · September 14, 2026 · We’ll score this in a future issue.
|
|
Put it to work 3 things to do this week.| 01 | If you have no interest in construction but want to ride this boom anyway — look at the software and workflow layer around it: building-automation-system (BAS) commissioning software, hyperscale project-management tools, and grid-interconnection consulting are all whitespace right now (story 01). | | 02 | If a layoff headline at your company or a peer’s cites “AI” — don’t take it at face value. Check Challenger, Gray & Christmas’s actual attributed-cause data and weigh whether the company’s other numbers support the AI framing before it shapes your own career decisions (story 02). | | 03 | If your company is quietly sorting people into an “AI elite” and everyone else — get explicit with your manager, in writing, about what “AI-proficient” means for your specific role before the next review cycle. Vague expectations turn into vague performance reviews (story 05). |
|
Career radar Data Center Commissioning EngineerThe specialist who tests and verifies that a newly built data center’s power, cooling, fire-suppression, and controls systems actually work together before it goes live — the last gate before a hyperscaler accepts a finished building. As AI data centers add liquid cooling and medium-voltage power faster than the trade has built standard training for, commissioning engineers are one of the roles hyperscalers say they can’t hire fast enough. A place to start. Most commissioning engineers come up through electrical or mechanical trades or building-automation backgrounds, not a dedicated commissioning degree. A journeyman electrician’s license or a Certified Commissioning Professional (CCP) credential is a realistic on-ramp; total comp on hyperscale programs runs $165,000–$290,000+. Explore AI careers → |
|
|
The opportunity board Know where to look.If you’d rather monetize expertise directly than chase a trades apprenticeship, the AI training and evaluation market is still hiring across a wide pay band: generalist annotation work pays $15–25/hr, coding and CS RLHF work $50–65/hr, and specialized expert review (medical, legal, finance) can reach $250–1,000+/hr on platforms like Mercor and Surge AI. See the full FOWL AI platform breakdown → On the trades side, hiring is happening through union apprenticeship programs, ABC (Associated Builders and Contractors) member contractors, and hyperscaler-adjacent GCs directly — not through typical tech job boards. Browse this week’s AI jobs board →
|
|
For two years, the loudest AI jobs story has been about who builds the models and who prompts them well. This week, the actual net-job-creation data says the biggest winner so far is neither — it’s the trades building the physical plant AI runs on. That’s not a story about a new AI skill. It’s a story about an old skill — a trade certification, not a bootcamp certificate — suddenly being priced like a new one. Until next Monday,
TYB
Data scientist & founder, FOWL AI
|
|
One question for you Would you take a six-figure data-center trade job over a lower-paid “AI” title?Hit reply. I read every response, and it helps shape what we cover next.
|
|
Atlanta · Vibe Code Saturdays Read about the future.
Then build it.A free, hands-on Saturday to build a website, newsletter, portfolio, or useful tool with AI. Beginners welcome. Bring your laptop. First date and venue coming soon. Apply and we’ll follow up with the details.
|
Know someone who’d find this useful?Share this briefing → Forwarded this email? Get the free weekly briefing.
|
|
|
|