Observed Signal · May 29, 2026 · Research Publication · Source: SemiAnalysis · Impact: 4/5 · Sentiment: Neutral
AI Dark Output: Invisible Economic Value from AI
SemiAnalysis defines “Dark Output” as AI-generated economic value that national accounts currently fail to capture and warns this measurement gap could be large and accelerating. The report distinguishes substitution dark output (tasks AI replaces) from new dark output (tasks made affordable by AI) and states its Dark Output Monitor identifies roughly $1.5 trillion in labor-tied tasks where current AI has credible displacement potential. It documents how receipts, price collapse, boundary shifts, and sector misrouting can make real AI work invisible in GDP, prices, and labor statistics, and cites macroeconomic precedents and measurement reforms (BEA, BLS). The piece calls for new measurement efforts so investors, policymakers, and firms can correctly assess AI’s real economic impact.
Measurement gaps in GDP and labor statistics caused by AI could materially distort macro signals used by investors, policymakers, and firms; the report quantifies exposed labor (~$1.5T) and proposes monitoring approaches, making it highly relevant to tech and economic decision‑making.
Track Anthropic Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- SemiAnalysis coins the term “Dark Output” for AI-enabled economic value not captured in current national accounts.
- SemiAnalysis’ Dark Output Monitor identifies roughly $1.5 trillion in tasks where current AI could substantially augment or automate labor (exposed labor, not missing output).
- A 2013 BEA methodology revision to capitalize R&D and intellectual property increased measured 1990s output by about $3.6 trillion, a historical analogue cited for measurement change.
- Anthropic’s Economic Index (March 2026) reportedly shows 37% of tokens are used in computers and mathematics, while investment-in-software GDP contribution has not broken from pre-AI trends.
- The article was published on 2026-05-29.
Connected Companies & Entities
1 Entity mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Exponential View Releases AI Economy Report
Exponential View released a bottom-up, de-duplicated research report, The State of the AI Economy, estimating the generative AI economy generated $110 billion in revenue over the past 12 months with an annualized run rate exceeding $175 billion. The report models demand-side AI spending across consumer and enterprise stacks, avoids double-counting by reporting end-customer dollars, separates AI-directed CapEx from ordinary CapEx for hyperscalers, and examines token-price elasticity and quality-adjusted output tokens as a better measure of economic value. The v1 report excludes China, internal AI uplift (e.g., ad revenue improvements), and professional services revenue.
AI and Work: Risks, Numbers, and New Jobs
This Italian-language analysis reviews major international reports and European policy to describe how AI will transform work unevenly rather than trigger mass unemployment. It cites WEF, McKinsey, OECD, IMF and EU sources estimating both large job displacement and significant job creation by 2030, identifies the sectors most exposed (administration, customer service, manufacturing, retail/logistics, transport), and catalogs emerging roles (prompt engineers, MLOps, AI governance/compliance, data curators, red-team/safety researchers). The piece focuses on Italy and Europe, highlighting regional gaps, PNRR funding opportunities, and policy recommendations for reskilling, governance, and inclusive labor transitions. Publication date: 2026-08-16.
Measuring AI Value: Useful Intelligence per Dollar
OpenAI outlines a framework for assessing AI economics centered on a proposed metric, "Useful Intelligence per Dollar," which asks whether AI completes valuable work, how much successful tasks cost, how dependable results are, and whether value improves at scale. The piece argues businesses should measure end-to-end cost per successful task (including retries, human review, and employee time) rather than cost per token, and track dependability categories (ready to use, needs correction, needs escalation). OpenAI also describes infrastructure and compute as central to improving model capability and efficiency. The post announces GPT‑5.6 (three tiers: Sol, Terra, Luna), claims GPT‑5.6 Sol set a new state of the art on certain long-horizon engineering benchmarks while using fewer output tokens, and positions ChatGPT Work and ChatGPT Enterprise as enterprise offerings built on OpenAI's security, privacy, and compliance foundations.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.