Every star, planet, comet, dust grain, neutron star and living thing is built from one ingredient — ordinary atomic matter. Line that up against the full contents of the cosmos and it disappears. The rest of the universe is two substances nobody has ever held, seen, or identified, measured only by the way they bend and stretch everything else.
4.9%
of the universe's energy budget is ordinary matter — protons, neutrons and electrons. The other 95.1% is dark matter and dark energy. Planck 2018 (A&A, 2020)
01 The inventory
Three ingredients. We can only see one.
The Planck satellite mapped the cosmic microwave background — the light released 380,000 years after the Big Bang — and read the universe's recipe out of its ripples. Two of the three ingredients have no known particle.
Source: Planck Collaboration 2018 (Astronomy & Astrophysics 641, A6, 2020). Ωm = 0.315 ± 0.007; baryon share Ωb ≈ 0.049 derived from Ωbh² = 0.0224 with H₀ = 67.4 km/s/Mpc. Sub-split of the 4.9% follows the baryon census in §04; the ~0.3% stellar share is derived, not directly measured.
02 The ladder
Everything, across five decades of scale.
A linear chart flattens the small stuff into nothing. On a logarithmic strip you can see the whole inventory at once — from the substance that dominates the cosmos down to the relic light that fills every cubic centimetre of it.
Photon share Ωγ ≈ 5.4 × 10⁻⁵ and the minimum neutrino share (Σmν ≥ 0.06 eV from oscillation experiments; Planck 2018 bounds Σmν < 0.12 eV) are derived from Planck 2018 parameters and standard relic physics. Planck's matter density already includes the small massive-neutrino term; it is broken out here for scale. Stellar share is derived (§01).
03 Inside the 4.9%
Nearly all of it is still the gas the Big Bang made.
In the first three minutes, nuclear reactions fixed the universe's elemental recipe: hydrogen and helium, and essentially nothing else. Thirteen point eight billion years of stellar fusion have since converted only about one percent of that gas into everything heavier — carbon, oxygen, iron, us.
Primordial helium mass fraction Yp = 0.2463 ± 0.0003 (Planck 2018 / Big Bang nucleosynthesis). Present-day solar abundances X = 0.7381, Y = 0.2485, Z = 0.0134 (Asplund, Grevesse, Sauval & Scott 2009, Annual Review of Astronomy & Astrophysics). The Sun stands in for typical enriched local matter; the cosmic average metallicity is lower.
04 The hiding places
Most atoms never joined a galaxy.
Even the ordinary matter is mostly invisible. It sits in thin, million-degree plasma strung between galaxies — so diffuse that for two decades roughly a third of all atoms simply could not be found. In 2020 fast radio bursts finally weighed the gap.
Source: Shull, Smith & Danforth 2012 (The Astrophysical Journal 759, 23) — "30% of the baryons may still be missing." Whiskers show the quoted 1σ uncertainties. The shortfall was subsequently closed: Macquart et al. 2020 (Nature 581, 391) used the dispersion of localised fast radio bursts to measure Ωb = 0.051 (+0.021 / −0.025) h₇₀⁻¹ at 95% confidence — an entirely independent method that lands on the same 4.9% the CMB predicted.
05 The takeover
Dark energy has only been the majority owner for 3.6 billion years.
The recipe is not fixed. Matter thins out as space expands; dark energy, as far as anyone can measure, does not. So its share climbs. For most of cosmic history the universe was a matter universe — the crossover happened after Earth already existed.
Modelled curve, not a measurement. Dark-energy fraction ΩΛ(t) computed from the flat ΛCDM model with Planck 2018 values (ΩΛ = 0.685, Ωm = 0.315, H₀ = 67.4 km/s/Mpc); radiation neglected. The 50% crossover falls at redshift z ≈ 0.30. Age of the universe 13.797 ± 0.023 Gyr (Planck 2018).
06 The hunt
What it costs to keep finding nothing.
Dark matter is flowing through you right now at roughly 230 kilometres per second. Detectors buried under kilometres of rock have been waiting for one of those particles to nudge a single atomic nucleus. Nothing has shown up — and the silence itself is the result.
Null result2.2×10−48
cm² — the strongest limit yet on how strongly a 40 GeV dark-matter particle can touch an atomic nucleus, from 4.2 tonne-years of xenon exposure with zero candidate events. LUX-ZEPLIN, 2024
A crack in the model3.1σ
preference for dark energy that weakens over time, from 14 million galaxy and quasar redshifts. Not yet a discovery — but the constant in the standard model may not be constant. DESI DR2, 2025
In this room0.4
GeV per cm³ — the local dark-matter density measured from the motions of 90,000 nearby stars. About half a proton's worth of it in every sugar-cube of space. Guo et al. 2020 (LAMOST + Gaia)
Standard values for the present-day photon and neutrino relic number densities at Tγ = 2.725 K, as tabulated in the Particle Data Group's Review of Particle Physics (Astrophysical Constants). LZ limit: LUX-ZEPLIN Collaboration, Physical Review Letters, 2024 (4.2 ± 0.1 tonne-year exposure, 90% CL). DESI: DESI Collaboration DR2 BAO + CMB, March 2025.
Why it matters
The 4.9% is not a footnote — it is the entire subject matter of chemistry, geology, biology and every telescope image you have ever seen. Everything we know how to study is the trace ingredient. That is the honest position: 95% of the universe is currently named after our ignorance of it.
It is also the most productive kind of ignorance in science. Dark matter was inferred because galaxies rotate too fast; dark energy because distant supernovae are too faint. Neither was invented to be mysterious — both were forced on physics by measurements that refused to fit. The fact that fast radio bursts, the microwave background, and the abundance of primordial helium all independently agree on 4.9% is why the strange 95% is taken seriously at all.
If you want to watch this change: DESI, Euclid and the Vera C. Rubin Observatory are each mapping billions of galaxies through the 2020s. The first hint that dark energy varies with time is already on the table at 3.1σ. The recipe printed above may not survive the decade.
Shull, Smith & Danforth, "The Baryon Census in a Multiphase Intergalactic Medium: 30% of the Baryons May Still Be Missing," The Astrophysical Journal 759, 23 (2012). Lyα forest 28 ± 11%, WHIM 25 ± 8%, collapsed phases 18 ± 4%, unaccounted 29 ± 13%.
Macquart et al., "A census of baryons in the Universe from localized fast radio bursts," Nature 581, 391 (2020). Ωb = 0.051 (+0.021 / −0.025) h₇₀⁻¹, 95% confidence.
Asplund, Grevesse, Sauval & Scott, "The Chemical Composition of the Sun," Annual Review of Astronomy & Astrophysics 47, 481 (2009). X = 0.7381, Y = 0.2485, Z = 0.0134 by mass.
LUX-ZEPLIN (LZ) Collaboration, "Dark Matter Search Results from 4.2 Tonne-Years of Exposure," Physical Review Letters (2024). Strongest spin-independent exclusion 2.2 × 10⁻⁴⁸ cm² at 40 GeV/c², 90% CL; no WIMP signal above 9 GeV/c².
DESI Collaboration, Data Release 2 baryon-acoustic-oscillation results, March 2025. DESI + CMB give a 3.1σ preference for evolving dark energy (w₀ = −0.42 ± 0.21, wa = −1.75 ± 0.58); with Type Ia supernovae, w₀ = −0.838 ± 0.055. Contested in the literature — treat as a hint, not a detection.
Guo, Liu, Liu & Zhang, "Measuring the local dark matter density with LAMOST DR5 and Gaia DR2," MNRAS 495, 4828 (2020). ρDM = 0.0133 M☉ pc⁻³ ≈ 0.4 GeV cm⁻³.
Particle Data Group,Review of Particle Physics — Astrophysical Constants. Present-day CMB photon density nγ ≈ 411 cm⁻³ at T = 2.7255 K; relic neutrino density ≈ 336 cm⁻³ summed over three species.
Neutrino mass floor: Σmν ≥ 0.06 eV from atmospheric and solar oscillation measurements; Planck 2018 upper bound Σmν < 0.12 eV (95% CL).
Flagged as derived or modelled, not measured: the ~0.3% stellar share of the total energy budget (Planck baryon fraction × the collapsed-phase stellar budget); the photon share Ωγ ≈ 0.005% and minimum neutrino share ~0.14% (computed from relic physics and Planck parameters); the mean cosmic density of roughly one atom per four cubic metres (from the critical density at H₀ = 67.4 and Ωb = 0.049); and the entire ΩΛ(t) curve in §05, which is a ΛCDM model evaluation rather than an observation. Percentages are energy-density fractions today and are rounded; they sum to slightly under 100% because photons and neutrinos are listed separately.