For 350 years physics has kept making the same discovery: two forces that looked different turn out to be one force wearing different clothes. Electricity and magnetism merged in 1865. Electromagnetism and the weak nuclear force merged in 1967 — and the merger was confirmed by experiment in 1983. The strong force joined the same mathematical framework in 1973. Gravity has never joined anything. Here is exactly how far the unification has got, what the near-misses look like, and what the next decisive measurement costs.
Gravity is roughly 1036 times weaker than electromagnetism — measured between two protons.
Electrostatic repulsion divided by gravitational attraction for a pair of protons, computed from CODATA 2022 constants: 1.2 × 1036. The other three forces sit within a handful of orders of magnitude of one another. Gravity sits thirty-six decades away — and it is the only one with no quantum description.
01The spectrum
Every force has a dimensionless coupling — a pure number saying how hard it pulls. Laid on a single logarithmic strip spanning thirty-nine decades, the strong, electromagnetic and weak forces huddle together at one end. Gravity is alone at the other.
Strong nuclear
≈ 1
8 gluons · range 10⁻¹⁵ m
Binds quarks into protons and neutrons, and those into nuclei. Confined: it never reaches beyond a nucleus.
Electromagnetic
7.3×10⁻³
Photon · range infinite
The fine-structure constant, α = 1/137.035999. Holds electrons to nuclei — which is to say, all of chemistry.
Weak nuclear
≈ 10−6
W± 80.369 · Z⁰ 91.188 GeV
Changes quark flavour. Feeble only because its carriers are heavy — it is electromagnetism's twin above 100 GeV.
Gravity
5.9×10⁻³⁹
Graviton · never detected
Gmp²/ħc, the gravitational coupling for two protons. Infinite range, always attractive, and still purely classical.
Couplings quoted at low energy; the strong and weak numbers are conventional order-of-magnitude values, since both "run" with energy. Boson masses and α from Particle Data Group 2024; α and Gmp²/ħc from CODATA 2022.
02The merger
Unification is a real historical process with dates and Nobel Prizes attached, not a slogan. Read this left to right: separate phenomena flow together into single theories. One line never joins.
Electroweak unification: Glashow, Salam and Weinberg, Nobel Prize in Physics 1979; W and Z bosons observed by UA1/UA2 at CERN in 1983 (Rubbia and van der Meer, Nobel 1984). Higgs boson observed by ATLAS and CMS, July 2012 (Englert and Higgs, Nobel 2013); mass 125.20 ± 0.11 GeV, PDG 2024.
03The near miss
Force strengths are not constants — they drift with the energy you probe them at. Extrapolate the Standard Model's three couplings upward and they very nearly converge around 1016 GeV. Nearly. They cross each other at three different energies instead of one point, and that near-miss has driven fifty years of theory.
Illustrative model: both line sets are one-loop renormalisation-group running from measured PDG 2024 values at mZ — the standard textbook calculation, not a measurement. Real two-loop and threshold corrections shift the crossings slightly; the qualitative result (Standard Model misses, MSSM nearly meets) is robust and well established.
04The gap
Quantum gravity becomes unavoidable at the Planck energy, 1.22 × 1019 GeV. The most powerful machine ever built reaches 1.36 × 104 GeV. Because the axis below is logarithmic, the bars flatter us enormously: the real ratio is about a million billion.
The shortfall
9 × 1014
Planck energy ÷ LHC energy
Collider physics is about fifteen orders of magnitude short of the energy where a theory of everything is forced to show itself.
Nature's best shot
2.4 × 1011
GeV · Telescope Array, 2023
The "Amaterasu" cosmic ray carried 2.4 × 10²⁰ eV in one particle — a thrown tennis ball's worth of energy in a single nucleus. Still 10⁸ short of Planck.
Brute force, priced
10,000
Light-years of linear collider
An illustrative estimate of the machine length needed to reach the Planck energy with conventional acceleration — longer than the Milky Way's disk is thick.
Illustrative: the 10,000-light-year collider is an order-of-magnitude estimate (A. Loeb), not an engineering study. Cosmic-ray energy from the Telescope Array Collaboration, Science, 2023. Planck energy from CODATA 2022.
05The clean tests
If the forces unify, protons should decay. The simplest grand unified theory, SU(5), predicted a proton lifetime near 1031 years — and detectors killed it in the 1980s. Since then, every push deeper has found nothing, which is itself the result: it rules out theories by the shelf-load.
Muon g−2 · final answer
127
Parts per billion, Fermilab 2025
aμ = 0.001165920705. The most precise magnetic-moment measurement ever made — for two decades the best hope of a crack in the Standard Model. It closed in June 2025 with no crack.
Higgs boson · found
125.20
GeV ± 0.11 · PDG 2024
The last missing Standard Model particle, predicted in 1964 and observed in 2012. Its discovery completed the theory — and told us nothing about gravity.
Neutrino mass · the one real crack
< 0.45
eV · KATRIN, Science 2025
Neutrinos oscillate, so they have mass — which the Standard Model does not allow. It is the only confirmed laboratory breach in the theory, and nobody knows what it points to.
Proton-decay limits at 90% confidence. p → e⁺π⁰ and p → μ⁺π⁰: Super-Kamiokande Collaboration, Phys. Rev. D 102, 112011 (2020), 0.37 Mton·yr exposure. Two-pion channels: Super-Kamiokande I–V, 0.401 Mton·yr (arXiv:2604.10975). Hyper-Kamiokande sensitivity from the Hyper-K Design Report (arXiv:1805.04163) and collaboration updates.
06The price
Nobody is going to build a galaxy-length collider. The realistic strategy is precision: measure the Higgs so exactly that a deviation appears, or watch enough protons for long enough that one of them dies.
Future Circular Collider
15 bn
Swiss francs · ~12 years
CERN's proposed 90.7 km ring under Geneva, recommended in 2025 as Europe's next flagship. About a third of the cost is digging the tunnel.
Hyper-Kamiokande
258
Kilotonnes of water · from 2028
Roughly eight times Super-K's fiducial volume, watching for a single proton to fall apart. Ten years of running buys one more decade of lifetime reach.
Standard Model dials
26
Free parameters, with neutrino mass
Nineteen in the minimal theory. Every one is measured, not derived. A genuinely unified theory should have far fewer — that is the aesthetic case for pressing on.
Quantum-gravity signals
0
Detected, ever
No graviton, no proton decay, no superpartner, no measured deviation from general relativity. A century of theory with an empty experimental column.
FCC figures from CERN's Future Circular Collider Feasibility Study (2025): 90.7 km circumference, ~15 billion CHF for the first stage. Hyper-K mass from the Hyper-K Design Report (2018).
Unification is not decoration. Every merger has paid out in technology within a lifetime: Maxwell's 1865 equations gave us radio, radar and the entire electromagnetic spectrum as an engineering resource. Electroweak theory gave us the collider physics that produced medical isotopes, hadron therapy, superconducting magnets and — in a side project at CERN in 1989 — the web.
The honest position today is that we have three of four. The Standard Model is the most precisely tested theory in science and it is visibly incomplete: it does not include gravity, it cannot explain why neutrinos have mass, and it needs 26 numbers that must be measured rather than derived.
What you can actually do. The results in this edition are all public. Super-Kamiokande, KATRIN and the muon g−2 collaboration publish their papers open-access; CERN streams its seminars; the Particle Data Group's Review of Particle Physics — the source of most numbers here — is free online and is the single best reference in physics.
And when a headline says a theory has been "proved," check the column that matters: what did the experiment measure, and to what precision? On unification, that column has been empty for fifty years. That is not failure. It is what a hard open problem looks like from the inside.
Like this? Get a new data story in your inbox every morning.