How AI and ‘experience creep’ are making it harder for new graduates to find jobs

Unemployment for recent college graduates (ages 22-27) rose to ~5.7% in Q4 2025 versus 4.2% for all workers, and job postings show ‘experience creep’: listings open to candidates with 2–4 years fell from 46% (mid-2022) to 40% (mid-2025) while those seeking 5+ years rose from 37% to 42%. Tech hiring is notably weak—software developer postings are down 29% and data/analytics roles down 38% from pre-pandemic—and firms appear reallocating spend toward AI capital (e.g., Oracle layoffs amid data-center builds), potentially reducing entry-level opportunities. Recommend monitoring AI capex versus labor spend and tech hiring metrics for talent-supply risks to long-term sector growth.

Analysis

Corporates are reallocating spend toward large, lumpy AI-capex projects that produce a near-term substitution of capital for entry-level labor. That reallocation creates a two-phase P&L effect: (1) an immediate hit to free cash flow and incremental depreciation that compresses near-term margins, and (2) a delayed productivity dividend if the projects scale — a dynamic that will produce volatile guidance surprises over the next 2–4 quarters.

The talent pipeline is the hidden macro risk. Fewer on-ramps for junior roles raises the marginal cost of hiring experienced engineers and forces firms to rely more on contractors and M&A for talent; I model a plausible 100–200bp increase in median tech wage growth over 2–4 years and an attendant rise in buyout activity for niche engineering shops. That structural scarcity also increases platform lock-in for incumbents who can pay to secure scarce senior talent.

Sector winners and losers will bifurcate on balance-sheet strength and capex exposure rather than pure software/tech labels. Hardware, cloud infrastructure and facility suppliers capture the bulk of incremental capex dollars and should see revenue visibility improve over a 12–24 month build cycle, while mid-cap SaaS and services vendors with tight free-cash-flow will be most vulnerable to margin compression and hiring freezes.

Oracle is emblematic of the tradeoff: firms taking on heavy infrastructure investment become particularly sensitive to guidance vs. investor expectations. Watch quarterly capex cadence, non-GAAP margin trajectory and deferred revenue conversion as 1–2 quarter leads on whether capital investment is being monetized or simply penalizing EPS.

Key Decisions for Investors

• Pair trade (6–12 months): Long NVDA (or NVDA 12-month calls) / Short ORCL (or ORCL 3-month puts). Rationale: Express exposure to AI hardware and infrastructure demand versus an enterprise capex-heavy name exposed to margin/cashflow pressure. Risk/reward: asymmetric — NVDA upside if AI demand persists (target +30–80%); ORCL downside limited to ~15–25% near term if guidance disappoints. Use 30% notional on short leg as hedge.
• Long Equinix (EQIX) or AMT (12 months): Buy the stock or 9–12 month calls to capture secular data-center leasing growth as hyperscalers accelerate deployments. Risk: cyclical tenant delays or supply glut inbuild; reward: stable high-margin leasing revenue and yield expansion if utilization rises (target +20–35%).
• Long staffing/training plays (6–12 months): Buy Robert Half (RHI) or Udemy (UDMY) to capture increased spend on upskilling/contract labor as firms pivot away from junior hiring. Risk/reward: cyclical sensitivity to hiring freezes but double benefit from higher billing rates and increased training revenues (target +15–30%).