UV Fixed Point in Quantum Gravity

• Ultraviolet fixed points in quantum gravity are scale-invariant coupling values where beta functions vanish, establishing a finite set of relevant directions.
• Functional renormalization group and lattice methods like CDT provide converging nonperturbative evidence for asymptotic safety across different truncations.
• Their existence underpins a predictive theory by dynamically limiting free parameters and explaining phenomena such as the small observed cosmological constant.

Ultraviolet (UV) fixed points in quantum gravity are central to the asymptotic safety scenario, which posits that all couplings of the quantum theory of gravity approach finite, scale-invariant values at high energies. This provides a route to non-perturbative renormalizability, circumventing the non-renormalizability of general relativity under standard perturbative quantization. The UV fixed point, if it exists and possesses a finite number of relevant directions, guarantees predictivity: the theory requires only a discrete set of low-energy measurements to fix all couplings. A variety of continuum and discrete approaches, from functional renormalization group (FRG) methods to lattice-based techniques such as Causal Dynamical Triangulations (CDT), support the existence of such nontrivial fixed points, albeit with distinct technical frameworks and approximations.

1. Core Definitions and Renormalization Group Structure

A UV fixed point in quantum gravity is defined by the simultaneous vanishing of beta functions for a set of appropriately rescaled, dimensionless couplings. In the Einstein–Hilbert truncation, the action reads

$$S = \frac{1}{16\pi G} \int d^4x\,\sqrt{g}\,(-R + 2\Lambda)$$

and the dimensionless couplings are

$$g(k) = G(k)\,k^2, \quad \lambda(k) = \Lambda(k)\,k^{-2}$$

where $k$ is the renormalization (RG) scale. The RG flow is governed by the Wetterich equation for the effective average action $\Gamma_k$, leading to flow equations of the form

$$\frac{dg}{dt} = \beta_g(g, \lambda), \qquad \frac{d\lambda}{dt} = \beta_\lambda(g, \lambda), \quad t = \ln k$$

A UV fixed point $(g_*, \lambda_*)$ satisfies $\beta_g(g_*, \lambda_*) = 0$, $\beta_\lambda(g_*, \lambda_*) = 0$ (Christiansen et al., 2012, Hamber, 2010).

The linearized flow near the fixed point yields the stability (or critical) exponents $\theta_i$, defined as eigenvalues of the stability matrix

$$M_{ij} = \left. \frac{\partial \beta_{g_i}}{\partial g_j} \right|_{\text{fixed point}}$$

The number of exponents with positive real part (UV-attractive/relevant) determines the dimensionality of the critical surface and thus the number of free parameters in the theory (Christiansen et al., 2012, Christiansen, 2016).

2. Functional Renormalization Group and Continuum Evidence

The FRG, centered on the scale-dependent effective average action $\Gamma_k$, enables non-perturbative investigation of metric quantum gravity. Various truncations have been explored:

• Einstein–Hilbert truncation: Typically finds a non-Gaussian fixed point in $g$ and $\lambda$ with two UV-attractive directions (e.g., $g_* \simeq 0.94,\, \lambda_* \simeq 0.20$ in one scheme), with complex-conjugate critical exponents $\theta \simeq 1.49\pm 2.68 i$, consistent with a spiral approach to the fixed point (Litim et al., 2012).
• Functional $f(R)$ truncation: Global in curvature, this setting reveals a unique globally regular fixed functional $f_*(R)$ that interpolates between an $R^2$-dominated UV regime and an Einstein–Hilbert-like IR (Demmel et al., 2015). The leading asymptotics are

$$f_*(R) \sim A R^2 + A' R^2 \ln R \quad \text{for} \ R \to \infty$$

The critical surface is found to be three-dimensional, corresponding to three UV-relevant directions (Falls et al., 2018). Padé resummation and global analyses confirm the weak-coupling nature of most operators.

• Higher-derivative extensions: When including $R^2$, $R_{\mu\nu}^2$, and $R_{\mu\nu\rho\sigma}^2$, fixed points with (typically) two to four relevant directions are identified. Notably, Riemann-invariant truncations reveal a four-dimensional UV critical surface due to quantum-induced relevance of operators such as $R\cdot(R_{\mu\nu\rho\sigma}R^{\mu\nu\rho\sigma})$ (Kluth et al., 2020).
• Gauge-invariant and background-independent flows: Approaches using the Vilkovisky–DeWitt formalism achieve fully background-independent results and find robust fixed points with similar number/dimension of relevant directions (Donkin et al., 2012).
• Inclusion of matter: The inclusion of scalars, spinors, and gauge fields modifies the position but not the existence of the fixed point, unless the matter content exceeds certain thresholds. The interplay between the irrelevance of certain matter couplings (e.g., the Higgs mass term) and the gravitational scaling solution provides a predictive mechanism for the Standard Model gauge hierarchy (Wetterich, 23 Jan 2026).

3. Lattice and Discrete Approaches: Causal Dynamical Triangulations

CDT investigates the RG flow nonperturbatively via Monte Carlo simulation of discretized spacetime manifolds built from simplicial building blocks with causality-preserving rules. The phase diagram in the $(K_0,\Delta)$ plane reveals three phases: branched-polymer (A), crumpled (B), and de Sitter-like (C). The C phase is of particular interest, as it possesses a semiclassical four-dimensional universe with quantum fluctuations.

A UV fixed point is associated with a continuous (second-order) phase transition, which allows for a continuum limit. The B–C line contains triple points that are candidates for such transitions. At these critical endpoints, correlation lengths diverge and the lattice spacing $a$ can be taken to zero, matching the continuum RG notion of a UV fixed point. Finite-size scaling yields nontrivial critical exponents, and ongoing work aims to locate genuine second-order endpoints to realize the asymptotic safety scenario in lattice quantum gravity (Ambjørn et al., 2019).

4. Truncation Dependence, Universality, and Predictivity

Numerical values for fixed point locations and critical exponents do depend on truncation schemes (e.g., inclusion of higher derivatives, specific gauge/ghost sectors, background independence, or the choice of regularization). However, the existence of the fixed point and the finiteness of the number of relevant directions are robust. For instance:

• Einstein–Hilbert truncations yield two relevant directions (Christiansen et al., 2012, Donkin et al., 2012).
• $f(R)$ and Ricci-tensor truncations yield three (Falls et al., 2018).
• Riemann-invariant expansions find four (Kluth et al., 2020).

Irrelevant operators increasingly exhibit near-Gaussian scaling, with exponents approaching their classical values at high order. The equal-weight criterion provides a stable identification of physical scaling directions across orders, revealing that relevant directions are strong mixtures of low-dimensional invariants, while irrelevant ones rapidly align with pure monomials (Kluth et al., 2020).

In unimodular gravity, the conformal mode is frozen, eliminating $\Lambda$-running. Still, an interacting UV-attractive fixed point is found for Newton's coupling, with a single relevant direction and robust approach to scaling (Eichhorn, 2013, Brito et al., 2020).

5. Phase Structure, Trajectories, and Physical Implications

The global RG flow in quantum gravity exhibits a rich phase diagram, with UV fixed points connected to the infrared (IR) domain—classical general relativity—via trajectories on the critical surface. Distinct IR fixed points correspond to asymptotically vanishing, positive, or negative cosmological constant. The existence of global trajectories connecting the UV fixed point to the IR classical regime is crucial for the viability of asymptotic safety (Christiansen et al., 2012, Christiansen et al., 2014).

A hallmark of the UV fixed point is the emergence of a nonperturbative RG-invariant scale $m$ (related to the vacuum energy), set dynamically as an integration constant of the RG flow

$$m \sim \mu \exp\left(-\int^g \frac{dg'}{\beta_g(g',\lambda(g'))}\right)$$

This provides a dynamical explanation for the observed smallness of the cosmological constant as a vacuum condensate effect (Hamber, 2010).

In the presence of matter, particularly in scale-invariant extensions or models with multiple scalar fields, the fixed point structure tightly constrains the running of couplings such as the Higgs mass parameter, leading to predictive relations among observed quantities (e.g., the Fermi/Planck mass ratio) (Wetterich, 23 Jan 2026).

6. Higher-Derivative Gravity, Dimensionality of Critical Surface, and Weak Coupling

Nontrivial fixed points persist when including four-derivative or higher curvature terms (e.g., $R^2$, $R_{\mu\nu}^2$, $C^2$). Notably, in four-dimensional truncations non-Gaussian fixed points have two to four relevant directions—counter to expectations from perturbative nonrenormalizability—indicating control over the proliferation of couplings (Christiansen, 2016, Sen et al., 2021). The majority of operators display near-Gaussian scaling, with quantum corrections decaying exponentially with operator dimension, thus most of "quantum gravity" is weakly coupled beyond a small core of relevant interactions (Falls et al., 2018, Kluth et al., 2020).

Comparison of different truncations demonstrates the importance of Riemann-tensor invariants in increasing the dimension of the UV critical surface (four in Riemann truncation versus two or three in $f(R)$ or Ricci curvature models) (Kluth et al., 2020).

7. Discrete Matter–Gravity Models and Alternative Scenarios

When classical gravity is minimally coupled to free quantum fields on a fixed background, the RG running of gravitational couplings can still admit nontrivial fixed points. For certain field content, corresponding closely to the Standard Model plus additional scalars, all gravitational beta functions vanish in the UV, leading to UV-finite stress tensor correlators and a realization of asymptotic safety in this sector (Boyle et al., 11 Sep 2025). This setting underscores the influence of matter content on the fate of gravity in the ultraviolet.

Table: Representative Fixed Points and Critical Exponents in Quantum Gravity

Truncation/Class	$(g_*, \lambda_*)$ or $\{\lambda_n^*\}$	$\{\theta_i\}$ (critical exponents)	$\dim(\mathcal{S}_{\text{UV}})$
Einstein–Hilbert (FRG, $d=4$)	$(0.94, 0.20)$	$1.49\pm2.68i$	2
$f(R)$ functional [upto $N\to70$]	$\lambda_0^*=0.25, \lambda_1^*=-1.08, \ldots$	$-2.53\pm2.20i$, $-1.66$, $+3.82$	3
$f(R, Ric^2)$ [upto $N=30$]	...	...	3
Riemann truncation [upto $N=144$]	$\lambda_0^*=0.055, \lambda_1^*=-0.352, \ldots$	$-4.96$, $-3.01\pm 1.46i$, $-1.33$, $+2.99$	4
Four-derivative (FRG)	$(0.43, -0.34, -0.41, 0.91)$	$-1.5\pm2.7i$, $2.4$, $8.3$	2
Unimodular gravity (FRG)	$G_* = 0.876$	$+3.88$	1
CDT (de Sitter–like phase C)	—	— [critical exponents determined via scaling]	1–2 (pending)

Dimensions correspond to the number of UV-attractive directions (relevant couplings) (Christiansen et al., 2012, Falls et al., 2018, Kluth et al., 2020, Christiansen, 2016, Eichhorn, 2013, Ambjørn et al., 2019).

• Functional RG: (Christiansen et al., 2012, Donkin et al., 2012, Demmel et al., 2015, Falls et al., 2018, Christiansen, 2016, Eichhorn, 2013, Hamber, 2010, Sen et al., 2021, Pawlowski et al., 2020, Christiansen et al., 2014, Christiansen et al., 2015, Kluth et al., 2020, Kluth, 2024)
• CDT: (Ambjørn et al., 2019)
• Matter-coupled fixed points: (Boyle et al., 11 Sep 2025, Wetterich, 23 Jan 2026)

Outlook

The nontrivial UV fixed point structure, first conceived in the context of scalar and gauge theories, is now established in a variety of gravitational settings, both continuum and discrete. The UV critical surface is finite-dimensional for a broad class of truncations, providing grounds for a predictive and UV-complete quantum theory of the metric. While precise numbers of relevant directions and critical exponents shift with truncation and approximation, the existence and qualitative features of the ultraviolet fixed point display a high degree of universality. Current challenges include extending the analysis to complete truncations (all diffeomorphism invariants, matter interactions), nonperturbative confirmation via lattice or CDT, and extraction of genuine physical observables and spectral quantities. These avenues are under active investigation and remain central to assessing the viability and empirical consequences of asymptotically safe quantum gravity.