Modern cosmology operates like a high-stakes game of telephone across cosmic time. We take a precise snapshot of the infant universe—the Cosmic Microwave Background (CMB)—and use our standard model of physics, $\Lambda$CDM, to project how that universe should expand and clump together over 13.8 billion years. At the same time, we look directly at our local cosmic neighborhood to measure the current expansion rate ($H_0$) and the distribution of matter ($S_8$).
If our understanding of the universe is complete, the forward-projected prediction and the direct local measurement should match perfectly. They do not. This persistent mismatch is known as cosmological parameter tension, and it represents either the dawn of new physics or the discovery of deeply hidden systematic errors.
The Geometry of Expansion: The Hubble Tension
To understand why the discrepancy in the Hubble constant ($H_0$) is so stubborn, we must look at how different methods anchor their measurements.
Early-universe methods rely on an absolute physical scale stamped into the plasma of the early universe before recombination: the sound horizon. As analyzed by Hu et al. 2001, the positions and relative heights of the acoustic peaks in the CMB lock down the universe's initial conditions with sub-percent precision. Legacy results from the Planck satellite map these temperature anisotropies to imply a current expansion rate of roughly 67.4 km/s/Mpc (Planck Collaboration 2019; Aghanim et al. 2020). This early-universe anchor is independently verified when combining Baryon Acoustic Oscillations (BAO) with Big Bang Nucleosynthesis (BBN), yielding a similarly low expansion rate (Cuesta et al. 2016; Baryon Acoustic Oscillations and the Hubble constant: past, present and future (Journal of Cosmology and Astroparticle Physics 2019)).
Conversely, late-universe methods measure $H_0$ locally without needing to assume an initial state or an evolutionary model. These direct approaches include local distance ladders, gravitational wave standard sirens (Nicolaou et al. 2020), and time delays from strongly lensed quasars (Huber et al. 2021; Storfer et al. 2023). These local measurements consistently yield a higher value of around 73 km/s/Mpc.
Early-Universe Anchor (CMB / BAO + BBN) ---> [ 67.4 km/s/Mpc ]
▲
| Significant Tension!
▼
Late-Universe Direct (Lensing / Sirens) ---> [ 73.0 km/s/Mpc ]
It is tempting to think that an evolving dark energy component could accelerate the late universe just enough to reconcile the two numbers. However, Lee et al. 2018 used an "inverse distance ladder" approach to show why this intuition fails. Because the relative expansion history at intermediate redshifts is already tightly constrained by BAO and Type Ia supernovae, the shape of the late-time expansion curve is locked in. Simply modifying physics at low redshifts cannot bridge the gap. The discrepancy is structural, meaning any physical solution likely requires modifying the physics before recombination.
The Clumpiness of Matter: The $S_8$ Tension
The Hubble constant is not the only parameter under fire. A parallel discrepancy has emerged regarding $S_8$ (or $\sigma_8$), which measures how intensely matter clumps together across the cosmos.
When we take the Planck CMB data and project it forward, $\Lambda$CDM predicts a highly structured, clumpy late universe. But when we look at the actual late universe using weak gravitational lensing—the subtle distortion of distant galaxy shapes by intervening dark matter—we find a smoother distribution. Large-scale surveys like the Dark Energy Survey (DES) (Jain et al. 2022; Chang et al. 2023) and BOSS LOWZ (Wibking et al. 2019) consistently find that the matter clustering amplitude is lower than the CMB-predicted value.
Why does this tension persist? One possibility is that our models fail to capture how visible galaxies trace the underlying dark matter halos—a phenomenon known as the "lensing is low" problem. As explored by Yuan et al. 2021 and Yuan et al. 2022, environmental effects on galaxy formation and complex baryonic physics can suppress the small-scale lensing signal. If these localized, non-linear effects are not perfectly modeled, they can mimic a cosmological discrepancy. However, even after accounting for these adjustments, a moderate statistical gap remains, as noted by Efstathiou and Handley et al. 2019.
Evaluating the Path Forward
To resolve these tensions, cosmologists are actively testing extensions to the standard model. One intuitive route is replacing the static cosmological constant ($\Lambda$) with dynamical dark energy that evolves over time. Hazra et al. 2015 and Zhao et al. 2017 found that certain dynamical dark energy configurations can improve data fits and alleviate some of the strain between early and late epochs.
Yet, as comprehensive multi-probe analyses show (Camilleri et al. 2025), no single alternative model definitively dethrones $\Lambda$CDM across all observational datasets without introducing new complications elsewhere. Ultimately, whether these discrepancies are the first cracks in our standard cosmological model or the result of subtle, unaddressed systematic errors across different experimental setups remains the central debate of modern precision cosmology.