Perplexity CEO says AI layoffs aren’t so bad because people hate their jobs anyways: ‘That sort of glorious future is what we should look forward to’

ServiceNow CEO Bill McDermott warned unemployment could exceed 30% in coming years; Block cut ~40% of staff (~4,000 employees) and the Alliance for Secure AI counts >101,000 AI-linked U.S. job losses since Feb 2025. Perplexity CEO Aravind Srinivas and VC Bill Gurley counter that AI displacement can spur entrepreneurship and market adaptation; Srinivas highlighted smaller, AI-optimized startups as routes to one-person/$1B businesses. Bank of America data shows business applications with clear hiring plans fell 4.4% YoY in January while “high propensity” businesses rose >15%, and a cited startup (TurboAI) claims ~$1M/month revenue with 13 employees, illustrating the shift toward capital-efficient, AI-enabled firms.

Analysis

AI-driven labor compression is not just a productivity story — it reshapes unit economics for starting businesses. Expect a sustained shift from headcount-driven cost curves to capital- and software-driven leverage: startups that historically needed 50–200 employees can now operate with single-digit headcounts, compressing demand for seat-licensed enterprise tools and payroll-dependent services over a 12–36 month horizon. This reduces TAM growth for vendors tied to per-employee pricing while increasing addressable markets for orchestration, inference infrastructure, and transaction rails that monetize volume rather than payroll.

ServiceNow-style incumbents face a two-fold threat: secular erosion of seat growth plus shorter sales cycles and lower ACV on small, AI-enabled customers who buy lightweight automation rather than enterprise platform rollouts. The second-order impacts extend to commercial real estate, MSPs, and HR/payroll SaaS — industries that feed into traditional renewal pools and cross-sell motion for large workflow vendors. Conversely, players that own inference stack, developer platforms, and micro-billing/payment flows capture recurring economics from many small customers instead of a few big ones.

Key risks that could reverse the trend are (1) a macro funding freeze that halts formation of microbusinesses within 6–12 months, (2) regulatory or export controls fragmenting compute supply chains, and (3) an accuracy/productivity cliff where claimed AI gains fail to materialize in repeatable revenue increments. Near-term we should watch capex guides from chip OEMs and enterprise seat trends in earnings; medium-term triggers are sustained shifts in business-application pricing models and SMB deposit/transaction growth metrics.

Key Decisions for Investors

• Pair trade: Long NVDA (12-month) / Short NOW (6–12 month). Express via equal-notional NVDA stock or Jan-2027 LEAP call (buy) funded by selling NOW Jan-2026 covered calls or initiating a NOW 6–9 month put spread. R/R: asymmetric — NVDA retains multi-bagger upside if AI capex continues; NOW downside concentrated if seat-based monetization decays by 20–40%. Monitor NVDA guidance and NOW seat/ACV commentary for early entry/adjustment.
• Long BAC (9–18 month tactical). Buy BAC stock or a Jan-2026 call spread to position for higher account/transaction velocity from proliferating microbusinesses and deposit stickiness. R/R: modest upside (20–40%) with contained downside relative to cyclicals; tail risk is credit-cycle stress or steep yield curve inversion compressing NIMs — size position accordingly.
• Short legacy seat-based SaaS exposure (6–12 month). Target vendors with >60% revenue tied to per-employee licensing via put spreads (use NOW as primary candidate). R/R: short-term earnings downside if enterprises accelerate automation pilots into production; catalyst window around next two quarterly earnings releases for seat guidance revisions.