v2060Notes · updated 29 Sep 2026
Headroom across 25 fields
How many times better the limit is than typical practice today (dot) and around 2010 (ring), on a log scale. The line between them is sixteen years of progress.
Dashed amber line at 3×: below it, remaining gains are incremental. Hollow blue markers show a requirement that isn't met yet, not a gap to a limit. Grey rings mark roughly 2010; where a ring sits on the dot, the field has barely moved. Hover or tap a row for the numbers.
Also on the radar, and where not to aim
Big but different
- Launch to orbit: about 3,200–4,200× above the energy floor. The gap is hardware and operations cost, not physics.
- Fusion: NIF reached a target gain of 4.13 in 2025. A plant needs about 50–100 at the target, before counting how efficiently the laser turns grid power into light.
- Quantum computing: breaking RSA-2048 needs under a million physical qubits at 0.1 % error. Today's gate-based devices have hundreds to about a thousand; atom arrays now trap over 6,000.
- Crops: photosynthesis runs at 0.5–1 % against a 4.6–6 % limit. Sub-Saharan yields sit at 20–30 % of their water-limited potential.
Already near the ceiling
- Electric motors: 95–97.5 % efficient.
- Wind turbines: Cp about 0.48 against the Betz limit of 0.593.
- Gas combined cycle: 64 % against about 84 % Carnot.
- Electrolysers: 1.4× their minimum.
- Fibre and radio coding: within 1–2 dB of Shannon per hertz. Capacity gains now come from more bandwidth and more channels (fibre: 10–20× above).
Method and caveats
The home page ranks sixteen problems three ways. Limit headroom ranks only problems with a proven or well-established limit, and discounts limits that assume ideal, defect-free materials (battery weight, strong materials); weather’s two-week limit comes from chaos models rather than a proof, and quantum speedups assume factoring is classically hard; problems measured against humans (robots, learning from little data) or other countries (building) or with no known ceiling are grouped separately. Current bottleneck is judgement: how much progress this decade the problem blocks, across how many sectors. Value if solved is a rough annual value by 2060 times the chance of solving it, good to about 3×. The five picks sit in the top half on both bottleneck and value. Biology and health are out of scope because this site is about engineering limits.
Headroom is the fundamental limit divided by the typical commercial value, or the other way round where lower is better. Where sources disagree, the bar shows the range. Limits are physical (Landauer, Carnot, Shockley–Queisser, Betz, Shannon, minimum work of separation) except where marked as a biological benchmark, a practical ceiling, a lab demonstration or an engineering requirement.
A few figures are estimates rather than sourced values: the physics floors for aircraft and road vehicles and typical EV consumption. Treat headroom as an order of magnitude, not a precise number.
- IEA, Key Questions on Energy and AI (2026)
- Ho et al., limits of CMOS energy efficiency
- Green500, June 2026
- Carlsmith, brain compute estimate
- Sholl & Lively, Nature 2016 (separations)
- Keith et al., Joule 2018 (direct air capture)
- Elimelech & Phillip, Science 2011 (desalination)
- IEA Ammonia Technology Roadmap
- US DOE cement bandwidth study
- US DOE iron & steel bandwidth study
- Hall–Héroult process
- ICCT, electric aviation
- Amprius 500 Wh/kg cells
- Specific strength of materials
- Falcon 9 pricing
- Legged robot cost of transport, Frontiers 2018
- YASA axial-flux motor record
- UN, Ageing
- Perovskite–silicon tandem records
- LLNL, NIF 2025
- Gidney 2025, RSA-2048 resources
- NICT/UCL fibre record, 2026
- Kalff et al., atomic memory
- Zhu, Long & Ort 2008, photosynthesis
- Global Yield Gap Atlas, Africa
- US DOE electrolysis assessment
- Keadby 2, 64.18 % combined-cycle record
- Betz's law
- Epoch AI / Ho et al., CMOS energy limits
- Thermodynamics of air capture with compression
- Amine sorbent heats of adsorption (RSC 2023)
- Liu et al., Nature Energy 2019 (Li-metal cells)
- Cui group, practical Li–S cells
- Ideal tensile strength of iron (DFT)
- Collins et al., Science 2005 (walking efficiency)