Mini-Series: AI Job Market (Issue A)

On January 28, 2026, Amazon announced it would cut roughly 16,000 corporate roles. That same week, Meta cut 1,500 from its Reality Labs division. Pinterest cut about 15% of its workforce. Autodesk cut 1,000. UPS announced plans for 30,000 cuts. Dow announced a reduction of 4,500.

This all happened in the same month, by profitable, stable companies. Their narrative is around efficiency, automation, and AI alignment. But what else is happening underneath the surface to drive these job cuts?

In this issue, we dig beneath the headlines and press releases. Grab a beverage of choice and settle in, as we have plenty to share.

-Brandon S.

If the corporate spending pattern continues, your region, your industry, and your household will feel this in different ways.

L1: Natural Resources & Energy

Data centers need copious amounts of power. AWS alone added 3.8 gigawatts of capacity in 12 months ending Q3 2025. AI data centers require 3-4x the power of traditional servers, given the same physical size. This is a hard physical boundary to reconcile.

That boundary determines where infrastructure can actually be built, and regions with abundant, cheap power win. By the same token, regions with constrained, expensive power lose. The Southwest (cheap solar and hydropower), the upper Midwest (hydropower from dams), and the Pacific Northwest (natural cooling + hydropower) are the geographies where “hyperscalers” can afford to build, and are doing so now.

IREN, an energy infrastructure firm, secured a $9.7 billion deal with Microsoft to build power infrastructure for data centers. This prompts utilities to expand generation capacity including new natural gas plants, renewable buildouts, transmission expansion. The work is underway, and with it comes opportunity, but it's concentrated geographically.

L2: Manufacturing & Construction

The data center boom is real construction. U.S. data center employment grew from 306,000 in 2016 to 501,000 in 2023. Over 400 data centers are currently under development. The construction industry faces a 439,000-worker shortage, with data center jobs paying 30% more than typical construction. Electricians, MEP engineers, HVAC specialists, and project managers are all roles surging in demand in Arizona, Iowa, Virginia, Texas, and Oregon.

Union electrician membership in the D.C., Maryland, and Virginia area has doubled since 2018, reaching 14,700 members. The work is visible. The jobs are real. But they're temporary—construction typically runs 18–36 months, then stabilizes or potentially returns to pre-boom levels.

Meanwhile, in traditional manufacturing, automation is compressing headcount. Autodesk cut 1,000 workers while "betting on AI." Corporate support roles—finance ops, project management, engineering management—are squeezed hardest.

Semiconductor manufacturing is also booming, but most of it happens outside the U.S. TSMC and SK Hynix are expanding fabrication capacity for AI chips. NVIDIA maintains 80% of the AI GPU market, with demand growing even as costs have increased dramatically. High-bandwidth memory (a special kind of memory used in AI computing) demand surged to a projected $100+ billion by 2030; SK Hynix sold out its 2026 HBM4 production with prices rising 172% year-over-year. But these jobs are concentrated in Taiwan and South Korea, not America.

L3: Retail, Services & Distribution

Traditional tech hub job markets are weakening. The San Francisco Bay Area lost 20,000 net jobs in 2025 with the information sector alone losing 4,500 positions. Job listings in San Francisco fell 37% from February 2020 to October 2025. Seattle's unemployment rose to 5.1%, above the national average of 4.5%.

Meanwhile, data center regions are gaining. Arizona, Iowa, Virginia, Oregon, and parts of Texas are where construction jobs have clustered and wages rise.

This matters because the people hurt by corporate layoffs can't easily move to where the new jobs are. A laid off tech company manager who made $200,000 in San Francisco can't easily uproot (especially a whole family) and become an electrician in Phoenix. So people end up stuck in expensive cities with shrinking job markets.

This also has a domino effect: when corporate workers lose income, service workers feel it immediately. Restaurants near Amazon (and similar) offices see lunch spending drop. Childcare providers see their income vanish while their costs remain. Real estate agents close fewer deals. Retail sees discretionary spending collapse and sales drop off. These service workers didn't get laid off, but their demand was supported by those who did.

Professional services face another pressure: substitution. AI is now drafting legal documents, analyzing financial data, producing junior research. Entry-level roles in many professions are at risk. In the U.S., 23.2 million jobs have 50%+ of tasks automatable. Most won't disappear; they'll transform. But this creates challenges for entry-level roles and future growth and advancement opportunities.

AI-related job postings reached 4.2% of all postings in December 2025. But here's the catch: 90% of those postings come from just 1% of companies: hyperscalers and large tech firms. For ordinary workers in ordinary cities, AI hiring remains nearly invisible.

L4: Management & Politics

States are paying attention. California and Washington are worried about losing talent. Arizona and Iowa are actively courting companies with incentives. If this trend continues, and if visible second-order effects (service sector weakness, lower tax revenue, unemployment diverges by region) materialize, policy responses will follow. They’ll likely be things we’ve seen before: Retraining programs. Relocation incentives. Tariffs or reshoring measures. These aren't in place yet, but the groundwork is being laid.

Policy is currently reactive, not proactive. The scale (50,000+ announced Q1 2026 layoffs so far) is notable but not crisis-level—we've seen far worse in 2020, 2008 and 2001. So federal policy is not yet in action. But watch state-level responses and tax revenue tracking in tech hubs, as those will be early signals of whether this trend is accelerating or stabilizing.

Several markers will tell us whether this corporate spending trend is accelerating, stabilizing, or reversing:

Unemployment divergence: If tech hub unemployment stays above 5% while data center regions stay below 4%, the geographic split is real and widening. This should be visible by March 2026 BLS reports.

Job postings by metro area: Monitor Indeed data by region. Declining postings in SF/Seattle coupled with rising postings in Phoenix/Des Moines would confirm the pattern. This should be visible by May 2026.

State tax funds: California and Washington tax revenue compared to prior years will show whether income losses are material. Declining revenue would show that service sector weakness is real and spreading. This should be visible by June 2026.

Construction employment: BLS data on electricians, MEP (plumbers, HVAC), and construction trades by region. Rising employment in Southwest and Midwest coupled with flat or declining data center region construction would signal the boom is slowing. This should be visible by Q2 2026.

Wage trends by occupation: Are electrician wages in Phoenix staying elevated? Are corporate manager wages in SF declining? Wage divergence would confirm the sectoral/geographic split. This should be visible by mid-2026 in job postings with wages publicized.

Hyperscaler capex guidance: Watch Amazon, Microsoft, Google, and Meta quarterly earnings for updated capex spending plans. If capex slows, hiring might resume. If capex accelerates, layoffs likely continue.

Your Coalscoop-informed edge: Time will tell if this pattern holds true, but these January 2026 signals are clear. This isn't a recession narrative; it's a structural shift narrative. Capital is flowing to AI infrastructure investments at the expense of traditional human hiring. We’ve shown how some regions and sectors benefit while others will shrink. This creates a new kind of economic risk: corporate profit doesn't guarantee job security anymore. It guarantees ruthless capital allocation, in the hopes of boosting corporate profits even further.

Understanding this system—where companies are spending, how layers of the economy interact, and which regions and sectors intersect with this direction—doesn't directly reduce employment risk in certain sectors, but it tells you how to best position yourself to navigate it.