AI Completely Failing to Boost Productivity, Says Top Analyst

JP Gownder of Forrester and multiple academic studies highlight a growing disconnect between heavy AI investment and measurable economic returns, citing historical productivity declines (2.7% annual growth from 1947–1973 vs. 1.5% from 2007–2019) and an MIT finding that 95% of companies integrating generative AI saw no meaningful revenue growth. Empirical tests show AI coding tools can slow programmers and autonomous AI agents completed under 3% of remote work tasks in one study, while Forrester predicts automation could structurally replace about 6% of jobs (~10.4 million roles) by 2030—suggesting downside risk to AI revenue expectations and potential labor-market dislocation rather than near-term productivity-driven upside.

Analysis

Market structure: The immediate winners are low-cost IT services and outsourcing (Accenture ACN, Infosys INFY, TCS) that can monetize re-staffing and remediation work; hyperscale cloud providers (AMZN, MSFT, GOOGL) retain leverage as capital-light consumers of AI compute. Direct losers are high-valuation, AI-native software vendors with unproven ROI (public proxy C3.ai AI, and some PLTR-sized data/analytics plays) and smaller AI startups that depend on continued aggressive capex. Expect buyer consolidation and pricing pressure on SaaS vendors as procurement shifts from speculative pilots to measured ROI thresholds, implying 10–20% downside to stretched revenue multiple names if guidance softens over 3–12 months.

Risks: Tail risks include (1) regulatory restrictions on generative models or data use (10–20% probability over 24 months), (2) a sudden hyperscaler capex pause that produces a chip/inventory glut (15% probability, 6–12 months), and (3) a major public failure that forces write-downs and client pullback. Short-term (days–weeks) risk is sentiment-driven equity repricing; medium (3–12 months) is guidance-driven earnings revisions; long-term (years) is structural labor displacement (Forrester ~6% jobs by 2030) that re-prices wages and demand. Hidden dependency: measurable productivity gains require organizational change and retraining—without that, ROI remains elusive.

Trade implications: Tactical: establish modest longs in IT services (ACN 2–3%, INFY 2%) and 1–2% short in AI-native public names (C3.ai ticker AI) sized to fund puts. Options: buy 3–6 month put spreads on AI (AI) to cap carry; buy a 6–9 month protection package on NVDA (small, 1% notional put spread) if hyperscaler guidance cuts occur. Sector rotation: trim high-growth SaaS exposure by 5–10% range over next 30–60 days, redeploy to IT services and select cloud infra names on >10% pullbacks. Enter within 30 days ahead of next earnings; set stops at 8–12% per position.

Key Decisions for Investors

• Establish a 2–3% long position in Accenture (ACN) and a 2% long in Infosys (INFY) within 30 days; target 12–24 month horizon, take profits if positions gain 15–25% or cut to breakeven on a 10% drawdown. Rationale: beneficiaries of rehiring/outsourcing and remediation work as AI pilots fail to produce ROI.
• Initiate a 1–2% short (or equivalent bought put spread) on C3.ai (AI) sized to portfolio risk within 14 days; use a 3–6 month 15% OTM put / 30% OTM call/put structure to limit carry. Cover or cut on any confirmed multi-quarter positive revenue acceleration or acquisition bid.
• Buy a small protective put spread (≈1% notional) on Nvidia (NVDA) with 6–9 month expiry (buy ~10% OTM puts, sell ~25% OTM puts) as insurance against a hyperscaler capex pause; deploy if NVDA rallies then drops >10% on guidance cuts.
• Reduce high-growth SaaS exposure by 5–10% over the next 60 days, reallocating proceeds into IT services (ACN/INFY) and selective cloud infra names on >10% pullbacks. Trigger: implement reallocation when a SaaS name trades >10x forward revenue and/or cuts FY guidance.