Open Research — 155 Technologies · 168 Years · CC BY 4.0

The future is predictable.
Here's how.

Wright's Law reveals a deep regularity: for every doubling of cumulative production, technology costs fall by a remarkably consistent percentage. This isn't hindsight — it's a forecasting tool. 155 technologies, 168 years of data, one simple equation.

0
Technologies
0
Years of Data
0
Median Learning Rate

A Simple Model That Predicts the Future

In 1936, Theodore Wright observed that every time cumulative aircraft production doubled, the cost per unit dropped by a fixed percentage. This "learning curve" turned out to be one of the most powerful prediction tools in economics.

A landmark study by the Santa Fe Institute (Nagy et al., 2013) tested six competing forecasting models against 62 technologies and found Wright's Law to be the best simple prediction model — outperforming Moore's Law, Goddard's, and others. The result is striking: a single-parameter power law, fit on past data, reliably forecasts future costs.

Gogerty (2025) explains why this works: learning curves emerge from the interaction of selfish symbols — ideas, techniques, and capital competing to reduce cost. The learning rate measures how efficiently a technology's ecosystem converts experience into savings. This dataset extends the SFI database from 62 to 155 technologies, adding AI, carbon removal, and 2025 price updates.

The learning rate is the percentage cost reduction per doubling of cumulative production. A 20% learning rate means costs drop 20% each time total output doubles — and this rate tends to persist, making it a forecasting input.

C(q) = C0 · q−b
C(q) Unit cost at cumulative quantity q
C0 Cost of the first unit produced
q Cumulative production volume
b Learning parameter (slope)
LR = 1 − 2−b (learning rate)

Why You Can Predict the Future

Wright's Law isn't just descriptive — it's predictive. Here's the science behind why a simple equation forecasts technology costs better than expert consensus.

The Best Simple Forecasting Model

The Santa Fe Institute's Performance Curve Database is the gold standard for empirical technology forecasting. In their landmark study, Nagy, Farmer, Bui & Trancik (2013) tested six competing models against 62 technologies and found Wright's Law produced the most accurate out-of-sample forecasts.

The result is remarkable: a single-parameter power law, fit on historical data, reliably predicts future costs — often decades ahead. Solar PV costs were forecast in 2010 to reach $0.50/W by 2030. They hit $0.10/W by 2024.

🏛 Explore the Santa Fe Institute PCDB →

Why Simple Models Win

Why does a single equation outperform complex models and expert panels? Gogerty (2025) argues that learning curves are an emergent property of selfish symbols — ideas, techniques, and innovations competing within technological ecosystems to reduce cost and increase efficiency.

This evolutionary view, drawn from The Symbology, explains why learning rates are so stable: they measure the rate of evolution in a technology's knowledge ecosystem. Technologies with rich, competitive idea-spaces (like solar PV and GPUs) learn fast. Technologies trapped in regulatory complexity (like nuclear) learn in reverse.

"The learning rate is not a property of the technology. It is a property of the ecosystem of ideas, capital, and competition surrounding it." — Gogerty, 2025

155 Technologies — Extending the SFI Database

This dataset extends the Santa Fe Institute's original 62-technology database to 155 technologies, adding AI/machine learning (GPU compute, LLM inference, algorithmic efficiency), carbon removal (DAC, BECCS, biochar, enhanced weathering), and updated 2024–2025 prices across energy and hardware.

The full methodology, regression results, and confidence intervals are available in the research paper: Experience Curves Extended: Wright's Law Across 155 Technologies.

📄 Read the Paper on SSRN →

Use This to Predict Costs

Every learning rate in this database is a forecast input. If solar PV has a 20.5% learning rate, and you can estimate future cumulative production, you can project future cost. This is how IRENA, BloombergNEF, and energy modelers build their scenarios.

Carbon removal technologies are early in their curves — tracking where solar was in the 1980s. If DAC follows a 15–20% learning rate (as BECCS and biochar suggest), costs could fall from $600/tonne to under $100/tonne within two decades of scaling.

Prediction is not certainty. Learning rates can shift if policy, regulation, or material constraints change. See methodology notes below.

Methodology & Data Notes

Inflation adjustment: 93 of 155 technologies use constant-dollar (inflation-adjusted) prices, with base years noted in each technology's cost label (e.g., "2005 USD/W", "1966 USD/lb"). 11 recent technologies (carbon removal, AI) are in nominal dollars over short time spans (3–8 years), where inflation distortion is minor. Physical-unit metrics (Wh/kg, GFLOPS, $/genome) are not affected by inflation.

Data currency: 49 technologies (32%) have data through 2024–2025. The remaining 106 are historical datasets — primarily from the Santa Fe Institute PCDB and Boston Consulting Group archives (1968–1972 endpoints). Learning rates from historical data remain valid: the power-law relationship does not expire.

Sources: All data points are traceable to primary sources with URLs. See Sources & References below. Full dataset: Download.

83 experience curves — each one a forecast

Each bubble is a technology. The x-axis shows its learning rate — the cost decline per doubling of cumulative production. Click any bubble to explore its data.

Experience Curve (log-log)
Cost Decline Over Time
Data Note

The median technology gets 21% cheaper per doubling

Distribution of learning rates across 83 technologies with Wright's Law fits. Each rate is a forecast input.

New Research

The Global Learning Rate

If individual technologies learn, does civilization itself learn?
And if so — fast enough?

20%
Apparent
Conventional composite
12%
True
Inclusive of externalities
52%
Required
For 1.5°C Paris target
Download the full paper (PDF) →

The Global Learning Curve

Energy intensity vs. cumulative GDP on a log-log scale, 1970–2023. The slope gives an 18.3% learning rate (R² = 0.96) — remarkably close to the median across 155 individual technologies.

Source: Maddison Project Database, BP/Energy Institute Statistical Review. OLS regression on log-transformed data.

The Three Numbers

What the world thinks it's learning, what it's actually learning, and what it must learn.

0%
Conventional
The apparent learning rate using market GDP — ignoring climate damage, biodiversity loss, and natural capital depletion.
−7.6 pp phantom gap
0%
Inclusive
The true rate, adjusted for unpriced externalities. 38% of apparent learning is cost-shifting to nature.
−39.6 pp Paris gap
0%
Required for 1.5°C
The carbon learning rate needed to limit warming to 1.5°C while sustaining 3% GDP growth. More than 4x the inclusive rate.

The Emissions Gap

Under the current learning rate, annual CO₂ emissions do not decline — they rise continuously. The Jevons paradox at planetary scale: GDP growth overwhelms intensity improvement.

Projection: 3% real GDP growth, 22.2% carbon learning rate held constant. Paris pathways per IPCC AR6.

Technology Learning Rates vs. the Global Economy

Individual technologies learn fast. The global economy, weighed down by materials, institutions, and externalities, learns much more slowly. Solar PV at 49% is the closest precedent for the 52% required for 1.5°C.

Sources: SFI Performance Curve Database, Farmer & Lafond (2016), IRENA (2023), Ziegler & Trancik (2021).

Carbon Removal: Wright's Law Promises Radical Cost Decline

Direct Air Capture, BECCS, and Enhanced Weathering follow their own experience curves. Current costs are high — but so were solar panels in 2005.

Projected via Wright's Law: C(x) = C0 × (x/x0)−b where b = −log2(1 − LR). Cumulative removal targets per IPCC AR6 & Global CCS Institute.

CDR Makes the Impossible Merely Very Hard

With 10 Gt/yr of carbon removal by 2050, the required learning rate drops from 52% (unprecedented) to ~35% (demonstrated by LED lighting and wind).

Required learning rate recalculated assuming CDR offsets gross emissions. Reference rates from SFI Performance Curve Database.

1.5°C Carbon Budget — Live
250.0
Gt CO₂ remaining

At 37.4 Gt/year, the 1.5°C budget (50% probability) is exhausted by approximately 2030–2031.
The budget may already be functionally exhausted once carbon cycle feedbacks and non-CO₂ forcing are included.

“The learning curve is a tool. What we learn on it is up to us.”

Nuclear: Wright’s Law’s Biggest Outlier

Every other technology in this database gets cheaper with scale. Nuclear got more expensive. Costs rose 5× while cumulative deployment grew 400× — the most dramatic failure of learning-curve economics in industrial history.

−23% learning rate — costs rise 23% per capacity doubling

US Nuclear: Actual Cost vs. Wright’s Law Prediction

Had nuclear followed even a modest 15% learning rate, costs would be under $500/kWe today. The 8× gap is the defining chart in energy economics.

Source: EIA, Lovering et al. (2016), IAEA PRIS. 2023 real dollars. CarbonSig Research, Feb 2026.

Real cost increase
1960s → 2024
Gap vs. Wright’s Law
at 15% learning rate
900+
NRC regulatory guides
(only tighten, never loosen)
10+ yrs
US build time
(vs. 5 yrs in South Korea)
117%
Average cost overrun
on nuclear megaprojects
2
US reactors built since 1978
(133 authorized before)

Learning Rates: Nuclear Is the Outlier

Solar PV has maintained a stable 20% learning rate for four decades. Nuclear’s negative rate means building more reactors has been associated with higher costs.

Learning rates from Grubler (2010), IRENA (2024), NREL ATB (2024), and this database (curve_161).

Can SMRs Reverse the Curve?

Small Modular Reactors are the first credible structural attempt to satisfy Wright’s Law preconditions for nuclear: factory fabrication, standardized designs, and shorter build times. But the evidence so far is sobering.

FOAK estimates from NREL ATB 2024, NuScale CFPP filings. NOAK projections assume 9.5% (optimistic) and 5% (base) learning rates.

The Promise

Factory fabrication moves 35% of capital cost into controlled environments (vs 5% for large reactors). Standardized designs enable worker continuity. Target build time: 3–5 years vs 10+ today.

The Reality

NuScale’s CFPP reached $20,139/kW before cancellation — comparable to Vogtle. Zero operational Western SMRs. No factory exists. The 9.5% learning rate assumption has never been demonstrated for nuclear.

The Math

At 9.5% learning rate, SMRs reach $3,600/kW target after ~50 serial builds. At 5%, it takes 200+. At the historical −23%, they never get there. The first 10–20 units reveal which trajectory prevails.

The Hidden Edge

Speed may matter more than overnight cost. At 8% WACC, a 10-year build doubles the overnight cost via financing. A 3-year SMR build at $10,000/kW beats a 15-year large reactor at $6,600/kW on total installed cost.

📄 Read the Full Analysis (20 Experts, 20 Charts)

Landmark Technologies

From record-breaking learners to anti-learners — each learning rate is a window into the future of that technology.

☀️
Solar Photovoltaics
20.5%
$30 → $0.10/W in 50 years. The textbook learning curve—one of the most studied technologies in Wright's Law history.
GPU Compute
89.2%
The fastest learning rate ever measured. GPU cost per GFLOPS dropped 89.2% every time cumulative shipments doubled.
☢️
Nuclear Electricity
−88.1%
An anti-learner: costs went UP with scale. Regulation, safety retrofits, and increasing complexity drove costs higher.
🌎
Carbon Removal
5–16%
Learning rates of 5–16% across BECCS, ocean alkalinity, and biochar—tracking solar's early trajectory.

Chemicals and energy learn fastest, but the spread is enormous

Distribution of Wright's Law learning rates by sector. Each point is a technology. The box shows the interquartile range; the line marks the median.

Data Explorer

83 technologies with Wright's Law fits. Search, sort, and click to explore.

Name Category Learning Rate Year Range Points Trend

Download the Database

The complete Wright's Law dataset — 155 technologies, 168 years of cost data, with 2025 updates and full source attribution. All data freely available under Creative Commons CC BY 4.0.

📊

Master Database

All 155 technologies with learning rates, R², confidence intervals, latest 2025 prices, cumulative production, and source URLs.

📈

New 2025 Data Points

Latest cost and production data points from 2024-2025, ready to extend existing experience curves.

🔗

Source Attribution

Every data source cited — 100+ URLs, publishers, years, and coverage. Full provenance chain.

📝

Individual Curve Data

Time series for each technology with annual cost, production, and cumulative totals. 155 JSON files.

📄

Wright's Law Extended

Gogerty (2025): "Experience Curves Extended: Wright's Law Across 155 Technologies." Full methodology.

🌍

Global Learning Rate

Gogerty (2025): "The Global Learning Rate: The World Economy as a Learning System." Energy, carbon, and material learning rates.

☢️

Nuclear Deep Dive

Gogerty (2026): "Why Nuclear Energy Costs Keep Rising — And Whether SMRs Can Reverse the Curve." 20 expert analyses, 20 charts.

</>

Embed This

Copy the iframe code below to embed the interactive scatter plot on your own site.

<iframe src="https://carbonfinancelab.com/wrights-law/" width="100%" height="800" frameborder="0" style="border:1px solid #e2e8f0;border-radius:12px;" title="Wright's Law Explorer"></iframe>
Cite Wright's Law: Gogerty, N. (2025). “Experience Curves Extended: Wright's Law Across 155 Technologies Including Carbon Removal and Artificial Intelligence.” SSRN Working Paper. SSRN 6198738. Data and interactive explorer: carbonfinancelab.com/wrights-law/

Cite Global Learning Rate: Gogerty, N. (2025). “The Global Learning Rate: The World Economy as a Learning System.” Working Paper. Download PDF.

Data lineage: Extends the Santa Fe Institute Performance Curve Database (Nagy, Farmer, Bui & Trancik, 2013). See also: “Why Simple Prediction Models Work”.

Sources & References

Every data point is traceable. Below are the 100+ sources underpinning this dataset, organized by domain. All data was verified as of February 2025.

Energy Technologies
  1. IRENA (2025). Renewable Power Generation Costs in 2024. International Renewable Energy Agency. irena.org
  2. IRENA (2025). Renewable Capacity Statistics 2025. irena.org
  3. BloombergNEF (2024). 3Q 2024 Global PV Market Outlook. Solar PV module price: $0.096/Wp. bnef.com
  4. BloombergNEF (2025). Lithium-Ion Battery Pack Prices Fall to $108/kWh. bnef.com
  5. BloombergNEF (2024). Battery Pack Prices Fall to $115/kWh (2024). bnef.com
  6. Lazard (2025). Lazard's Levelized Cost of Energy+ (LCOE+), June 2025. lazard.com
  7. IEA-PVPS (2025). Snapshot of Global PV Markets 2025. Cumulative PV: 2.25 TW. iea-pvps.org
  8. NREL (2025). Winter 2025 Solar Industry Update. nrel.gov
  9. GWEC (2025). Global Wind Report 2025. Wind capacity: 1,136 GW. gwec.net
  10. GWEC (2025). Offshore Wind Capacity: 83 GW. gwec.net
  11. EIA (2025). Annual Energy Outlook 2025 — LCOE Report. eia.gov
  12. World Nuclear Association (2025). Economics of Nuclear Power. world-nuclear.org
  13. IEA (2025). Global Hydrogen Review 2025. Electrolyzer capacity: 4.9 GW. iea.org
  14. Our World in Data. Solar PV Module Prices (historical). ourworldindata.org
  15. ThinkGeoEnergy (2025). IRENA: 16% Decrease in Geothermal LCOE in 2024. thinkgeoenergy.com
AI / Machine Learning
  1. Epoch AI (2025). How Much Does It Cost to Train Frontier AI Models? epoch.ai
  2. Epoch AI (2025). LLM Inference Price Trends. 150x decline in 4 years. epoch.ai
  3. Epoch AI (2025). B200 GPU Cost Breakdown ($6,400). epoch.ai
  4. Epoch AI (2025). NVIDIA Chip Production — Stock Doubling Every 10 Months. epoch.ai
  5. Epoch AI (2025). AI Trends Dashboard. Algorithmic efficiency, training compute, benchmarks. epoch.ai
  6. Epoch AI (2025). SWE-bench Verified Leaderboard. epoch.ai
  7. a16z (2025). LLMflation: LLM Inference Cost Going Down Fast. a16z.com
  8. NVIDIA (2025). Q3 FY2026 Financial Results — $57B Revenue. nvidia.com
  9. Anthropic. Claude API Pricing. anthropic.com
  10. Google. Gemini API Pricing. google.dev
  11. DeepSeek. API Pricing. deepseek.com
  12. IntuitionLabs (2025). NVIDIA Data Center GPU Specs & Comparison. intuitionlabs.ai
  13. DemandSage (2026). ChatGPT Statistics. 800M weekly active users. demandsage.com
Information & Hardware
  1. TrendForce (2025). DRAM Spot Pricing. DDR5: $0.50–$0.81/GB. trendforce.com
  2. TrendForce (2025). NAND Flash Prices Jump Double Digits in Q1. trendforce.com
  3. Backblaze (2025). Hard Drive Cost per Gigabyte. backblaze.com
  4. Tom's Hardware (2025). Hard Drive Prices Surged 46% — AI Demand. tomshardware.com
  5. NHGRI (2025). DNA Sequencing Costs: Data. genome.gov
  6. EE Times (2025). TSMC Price Hikes End the Era of Cheap Transistors. eetimes.com
  7. Tom's Hardware (2025). Cost Per Transistor Stopped Dropping at 28nm. tomshardware.com
  8. Statista (2025). LED Lamp Prices (US). ~$0.45/kilolumen. statista.com
  9. TeleGeography (2025). IP Transit Price Erosion — Regional Differences Remain. telegeography.com
  10. SEMI (2025). Silicon Wafer Shipments Increase 3% in Q3 2025. semi.org
Chemicals & Materials
  1. FRED (2025). PPI: Plastic Resins & Materials (WPU066). Index: 253.5 (Dec 2025). fred.stlouisfed.org
  2. Plastics Europe (2025). Plastics — the Fast Facts 2025. plasticseurope.org
  3. Statista (2025). Global Plastic Production 1950–2024. 413.8 Mt. statista.com
  4. Our World in Data. Cumulative Global Plastics. ~8.3 Bt all-time. ourworldindata.org
  5. ChemAnalyst (2025). Polypropylene, PVC, Polystyrene, Ammonia, TiO2, Carbon Black — Pricing Data. chemanalyst.com
  6. World Steel Association (2025). December 2024 Crude Steel Production & 2024 Totals. 1,883 Mt. worldsteel.org
  7. USGS (2024). Cement — Mineral Commodity Summaries 2024. usgs.gov
  8. Statista (2025). Global Cement Production Volume. 4.2 Bt. statista.com
  9. Trading Economics (2025). Aluminum (LME), Steel Commodity Prices. tradingeconomics.com
  10. Ethanol RFA (2025). Annual Ethanol Production — US: 16.22 Bgal (2024 record). ethanolrfa.org
Carbon Removal & Biotechnology
  1. CDR.fyi (2025). Durable Carbon Removal Market Dashboard. 42.5 Mt contracted, ~1 Mt delivered. cdr.fyi
  2. CDR.fyi (2025). Q3 2025 Market Update & DAC Market Snapshot. cdr.fyi
  3. CDR.fyi (2025). CDR Pricing Survey, January 2025. cdr.fyi
  4. State of CDR (2025). stateofcdr.org. >$9.8B invested. stateofcdr.org
  5. Climeworks (2025). Mammoth Plant. 36,000 t/yr design capacity. climeworks.com
  6. 1PointFive (2025). Stratos DAC Facility, Ector County TX. 500,000 t/yr. 1pointfive.com
  7. Frontier Climate (2025). Lithos Carbon — $57.1M Enhanced Weathering Deal. frontierclimate.com
  8. Frontier Climate (2025). Planetary Technologies — $31M OAE Deal. frontierclimate.com
  9. CarbFix (2025). First EU Onshore CO2 Storage Permit. carboncredits.com
  10. Carbon Credits (2025). Microsoft Doubles CDR Deals to 45 Mt in 2025. carboncredits.com
  11. NHGRI (2025). DNA Sequencing Costs: Data. genome.gov
  12. Twist Bioscience (2025). Gene Synthesis Pricing. $0.07/bp standard. twistbioscience.com
Foundational References & Data Lineage
  1. Santa Fe Institute — Performance Curve Database (PCDB). The foundational open dataset for technology experience curves. 135 performance curves across dozens of technologies. This project extends the PCDB dataset. pcdb.santafe.edu
  2. Nagy, B., Farmer, J.D., Bui, Q.M., & Trancik, J.E. (2013). “Statistical Basis for Predicting Technological Progress.” PLoS ONE, 8(2), e52669. The landmark SFI study showing Wright's Law outperforms competing forecasting models. doi:10.1371/journal.pone.0052669
  3. Gogerty, N. (2025). “Experience Curves Extended: Wright's Law Across 155 Technologies Including Carbon Removal and Artificial Intelligence.” SSRN Working Paper. Extends the SFI database from 62 to 155 technologies with 2024–2025 data. SSRN 6198738
  4. Gogerty, N. (2025). “Why Simple Prediction Models Work: Selfish Symbols and Experience Curves.” SSRN Working Paper. Explains the evolutionary dynamics behind stable learning rates. SSRN 5824023
  5. Gogerty, N. The Symbology: Selfish Symbols and Evolution in the Age of AGI and Blockchain. The broader theoretical framework connecting evolutionary information dynamics to economic phenomena. Author page on SSRN
  6. Gogerty, N. (2014). The Nature of Value: How to Invest in the Adaptive Economy. Columbia University Press. The foundational work on adaptive valuation and how economies evolve through competitive information dynamics. Amazon
  7. Wright, T.P. (1936). “Factors Affecting the Cost of Airplanes.” Journal of the Aeronautical Sciences, 3(4), 122–128. The original discovery of the learning curve.
  8. Arrow, K.J. (1962). “The Economic Implications of Learning by Doing.” Review of Economic Studies, 29(3), 155–173.
  9. Sahal, D. (1979). “A Theory of Progress Functions.” AIIE Transactions, 11(1), 23–29.
  10. Boston Consulting Group (1968). Perspectives on Experience. BCG. Historical cost data for 26+ chemicals and manufactured goods.

Feedback & Error Report