The semiconductor industry's exponential progress is frequently cited as a benchmark for technological advancement, but this framing obscures a crucial insight: Moore's Law represents just one point in a remarkably wide distribution of improvement rates. Some technologies have improved by factors exceeding a trillion, while others—despite being critical to modern life—have advanced by less than tenfold over comparable periods. This analysis ranks 50 technologies by their total improvement magnitude, revealing which domains experienced the most dramatic sustained progress and which approached physical or economic limits far earlier.
At the pinnacle of technological improvement stand communication and precision measurement technologies, with undersea cable capacity achieving the most extreme advancement in human history—approximately 10¹² improvement (one trillion-fold) from the 1858 telegraph's fraction of a bit per second to today's 350 Tb/s systems. This represents information transmission improving by roughly a factor of 10 every 13 years for over 160 years.
Atomic clock accuracy and frequency measurement precision share second place with approximately 10¹⁰ improvement—from fractional uncertainties of 10⁻⁹ in 1955 to 10⁻¹⁹ today. These clocks now lose less than one second over the age of the universe, enabling GPS, financial trading synchronization, and tests of fundamental physics. The improvement continues rapidly; JILA's 2021 demonstration achieved 7.6×10⁻²¹ differential precision.
Semiconductor purity achieved 10⁹ improvement in impurity reduction, from 99% pure metallurgical-grade silicon in the 1950s to 99.9999999% (nine-nines) electronic-grade silicon today—meaning fewer than one non-silicon atom per billion. This extraordinary purification enabled the entire microelectronics revolution.
| Rank | Technology | Improvement Magnitude | Time Period | Status |
|---|---|---|---|---|
| 1 | Undersea cable capacity | ~10¹² (1 trillion×) | 166 years | Slowing |
| 2 | Atomic clock accuracy | ~10¹⁰ | 70 years | Rapidly improving |
| 3 | Frequency/time measurement | ~10¹⁰ | 75 years | Rapidly improving |
| 4 | Semiconductor purity (impurity reduction) | ~10⁹ | 70 years | Plateaued |
| 5 | Genetic modification precision (breeding→CRISPR) | ~10⁶-10⁹ | 12 years (CRISPR) | Rapidly improving |
| 6 | HDD storage density | ~5×10¹¹ (500 billion×) | 68 years | Slowing |
| 7 | Particle accelerator energy | ~10⁸ | 85 years | Continuing |
| 8 | Fiber optic bandwidth | ~10⁷ | 47 years | Continuing |
| 9 | Network backbone bandwidth | ~10⁷ | 55 years | Continuing |
| 10 | Gravitational wave detection | ~10⁷ | 55 years | Rapidly improving |
| 11 | CPU energy efficiency | ~10⁶-10¹² | 79 years | Slowing |
| 12 | DNA sequencing cost | ~10⁶ | 23 years | Continuing |
| 13 | Telescope angular resolution | ~10⁶ | 413 years | Continuing |
| 14 | Camera sensor resolution | ~10⁷ | 49 years | Maturing |
| 15 | Temperature measurement precision | >10⁶ | 200 years | Continuing |
| 16 | Satellite communication bandwidth | ~10⁵ | 62 years | Rapidly improving |
| 17 | DRAM cost per GB | ~10⁵ | 44 years | Continuing |
| 18 | Semiconductor transistor density | ~2.5×10⁷ | 53 years | Slowing |
| 19 | Telephone switching capacity | ~10⁴-10⁵ | 110 years | Technology shifted |
| 20 | Drug screening throughput (HTS) | ~10⁴ | 35 years | Continuing |
| 21 | GPS positioning precision | ~10⁴ | 32 years | Continuing |
| 22 | MEMS sensor precision | ~10⁴ | 30 years | Continuing |
| 23 | Laser frequency stability | ~2×10⁴ | 40 years | Continuing |
| 24 | Mass spectrometer resolution | ~10⁴ | 77 years | Continuing |
| 25 | Optical fiber attenuation | ~7,000× | 55 years | Plateaued at limit |
| 26 | Solar PV cost | ~1,000× | 47 years | Rapidly improving |
| 27 | Laser diode cost | ~1,000× | 40 years | Continuing |
| 28 | Wireless spectrum efficiency | ~600× | 45 years | Continuing |
| 29 | SSD/Flash storage cost | ~100-150× | 24 years | Continuing |
| 30 | Gene synthesis cost | ~100-150× | 24 years | Slowing |
| 31 | CNC machining precision | ~100× | 50 years | Continuing |
| 32 | Industrial robot precision | ~100× | 40 years | Continuing |
| 33 | Synthetic fiber tensile strength | ~75× | 85 years | Incremental |
| 34 | Li-ion battery cost | ~65× | 33 years | Rapidly improving |
| 35 | Permanent magnet strength (BHmax) | ~50× | 100+ years | Slowing |
| 36 | PCR diagnostic sensitivity | ~10⁴ | 40 years | Continuing |
| 37 | PCR speed | ~50× | 40 years | Continuing |
| 38 | Rocket payload cost to LEO | ~40× | 43 years | Rapidly improving |
| 39 | Container loading cost | ~37× | 68 years | Mature |
| 40 | Container ship capacity | ~40× | 68 years | Slowing |
| 41 | Superconductor critical temperature | ~32× | 113 years | Stalled* |
| 42 | Drone payload capacity | ~30× | 10 years | Rapidly improving |
| 43 | MRI resolution | ~25× | 40 years | Slowing |
| 44 | Electron microscopy resolution | ~25× | 50 years | Continuing |
| 45 | Carbon fiber cost | ~20× | 55 years | Continuing |
| 46 | Video compression efficiency | ~20× | 35 years | Continuing |
| 47 | EV battery cost | ~10× | 16 years | Rapidly improving |
| 48 | 3D printing resolution | ~10× | 30 years | Rapidly improving |
| 49 | LED lighting efficiency | ~8× | 28 years | Continuing |
| 50 | Display pixel density | ~7× | 25 years | Plateauing |
*Superconductor Tc improved ~32× at ambient pressure but stalled since 1993; room-temperature claims require impractical pressures (~270 GPa).
Several important technologies showed more modest improvement magnitudes despite their economic significance:
- Crop yields (corn): ~6× over 100 years (26→177 bushels/acre)
- Aircraft cruise speed: ~5× (1940s-1970), then plateaued for 50+ years
- Automotive fuel efficiency: ~2× over 50 years (13→27 MPG)
- Rail freight efficiency: ~2× over 44 years
- Wind turbine capacity factor: ~2.5× over 40 years
- Li-ion energy density: ~3.5× over 33 years (improving slower than cost)
- Nuclear capacity factor: ~1.9× (50%→93%)
- Natural gas combined cycle efficiency: ~1.8× (35%→64%), near theoretical limit
- High-speed rail operating speed: ~1.5× over 60 years
- Electric motor efficiency: ~12 percentage points (84%→96%)
- Cement production energy: ~38% reduction over 40 years
- Aluminum smelting energy: ~30% reduction over 120 years
The technologies showing the most dramatic improvements share several characteristics. Information and measurement technologies dominate the top rankings because they benefit from the compounding effects of digitization—once a signal becomes digital, it can exploit semiconductor scaling for processing, storage, and transmission. The top 10 technologies are predominantly about moving, storing, or measuring information.
Physical transformation technologies—moving atoms rather than bits—show more modest gains. Aluminum smelting, cement production, and steel manufacturing have improved their energy efficiency by only 30-40% over many decades because they face irreducible thermodynamic constraints. You cannot smelt aluminum below its minimum energy requirement no matter how clever the process.
Three patterns distinguish rapid improvers:
-
Combinatorial scaling: Fiber optics improved through wavelength division multiplexing—each new wavelength channel multiplies capacity. Adding 80 channels at 400 Gb/s each yields 32 Tb/s from technology that started at 45 Mb/s.
-
Miniaturization cascades: Semiconductor scaling improved transistor density, which improved sensors, which improved measurement precision, which improved manufacturing control, which enabled further miniaturization.
-
Platform shifts: DNA sequencing improved ~10⁶× not through incremental Sanger method improvements but through the paradigm shift to massively parallel next-generation sequencing around 2008.
Transistor density (rank 18 at ~2.5×10⁷ improvement) is remarkable but not exceptional within this distribution. Several technologies improved by similar or greater magnitudes: HDD density (~5×10¹¹), fiber bandwidth (~10⁷), atomic clocks (~10¹⁰). What made Moore's Law culturally prominent was its predictable doubling rate (~2 years), enabling long-term planning in ways that more erratic improvements could not.
The median improvement in this list is approximately 100-1,000× (10²-10³), meaning half of these notable technologies improved by less than three orders of magnitude over their development history. Technologies improving by 10⁶ or more are genuine outliers—only about 15 of the 50 achieved this.
Accelerating technologies (room for continued rapid improvement):
- Atomic clocks and precision measurement
- Solar PV and battery costs
- DNA sequencing and gene editing
- Rocket launch costs (SpaceX Starship targeting another 10-100×)
- Satellite bandwidth (LEO constellations)
- Gravitational wave detection
Decelerating or plateaued technologies:
- Transistor density (2D scaling ending; 3D offers continued but slower gains)
- Optical fiber attenuation (within 20% of Rayleigh scattering limit)
- Aircraft speed (no commercial improvement since 1973)
- Nuclear construction costs (actually increasing)
- Semiconductor purity (current levels exceed requirements)
- Permanent magnets (approaching crystallographic limits)
This analysis suggests several forecasting principles. First, don't generalize from Moore's Law—it represents the upper quartile of improvement rates, not a universal law of technological progress. Expecting new technologies to match semiconductor scaling will usually lead to disappointment.
Second, information technologies genuinely are different. The top 15 technologies by improvement magnitude are overwhelmingly about information processing, storage, transmission, or measurement. Physical transformation technologies—despite their importance—improve far more slowly.
Third, plateaus are common and predictable. Many technologies approach fundamental limits: optical fiber attenuation (Rayleigh scattering), electric motors (already >95% efficient), NGCC plants (~64% versus ~70% theoretical maximum). Recognizing these limits early prevents unrealistic expectations.
Fourth, paradigm shifts matter more than incremental improvement for achieving 10⁶+ gains. DNA sequencing, genetic modification, and gravitational wave detection all achieved their extraordinary improvements through fundamental reconceptualization rather than optimizing existing approaches.
The distribution of technological improvement rates—spanning from 10¹² to less than 10×—reflects the remarkable diversity of the challenges humanity has tackled. That undersea cables improved by a trillion-fold while aircraft speed stagnated for half a century tells us something profound about which problems admit exponential solutions and which hit walls far earlier. Understanding this distribution is essential for realistic technology planning in the decades ahead.