Cheng's Conjecture in Geometric Analysis

• Cheng's Conjecture is a set of related rigidity and comparison principles that link curvature or metric bounds to extremal eigenvalue behavior and biholomorphic characterizations.
• It spans multiple formulations in Riemannian geometry, hyperbolic spectral theory, and several complex variables, establishing model comparisons with sharp rigidity results.
• Key extensions include applications to nonlinear operators like the Pucci operator and developments in synthetic metric measure spaces, affirming its significance in modern geometric analysis.

In current mathematical usage, “Cheng’s Conjecture” does not designate a single universally fixed statement. In the literature represented here, the expression is used for several distinct but structurally related rigidity problems: a Cheng-type extension of eigenvalue comparison for geodesic balls under curvature bounds; a sharp Yang-type universal inequality for Dirichlet eigenvalues on bounded domains in hyperbolic space; and the complex-analytic assertion that a bounded strictly pseudoconvex domain with Kähler–Einstein Bergman metric must be biholomorphic to the unit ball. Across these formulations, the common theme is that curvature or canonical-metric hypotheses force extremal spectral behavior or ball-type rigidity (Ariturk, 2016, Luo, 22 Apr 2026, Sha, 15 Oct 2025).

1. Principal formulations

The name is attached to different conjectural programs in geometric analysis, spectral theory, and several complex variables. The following synopsis organizes the usages that appear in recent arXiv work.

Setting	Standard formulation	Status in the cited literature
Geodesic-ball eigenvalue comparison	Cheng-type comparison of the first Dirichlet eigenvalue with the model space form under curvature bounds	Extended to the Riemannian Pucci operator and to synthetic $\mathsf{CD}^*(K,N)$ spaces (Ariturk, 2016, Luca et al., 31 Jul 2025)
Hyperbolic universal inequalities	For bounded $\Omega\subset \mathbb{H}^n(-1)$, replace the coefficient $4$ by $\frac{4}{n}$ in the Cheng–Yang hyperbolic inequality	Verified up to loss of $\varepsilon$ for two special kinds of bounded domains (Luo, 22 Apr 2026)
Bergman metric rigidity	If the Bergman metric of a bounded strictly pseudoconvex domain is Kähler–Einstein, then the domain is biholomorphic to the unit ball	Resolved by Huang–Xiao; extended to the cscK Bergman case and to an algebraic two-dimensional Stein-space setting (Sha, 15 Oct 2025, Ganguly et al., 2022)

This plurality is not accidental. In each case, the conjectural content is a sharp model comparison with rigidity in the equality case. In the spectral formulations, the model object is a constant-curvature ball or a one-dimensional curvature-dimension model; in the complex-analytic formulation, the model object is the unit ball with its Bergman metric.

2. Geometric spectral origin: Cheng’s comparison theorem and its Cheng-type extensions

Cheng’s classical eigenvalue comparison theorem concerns the first Dirichlet eigenvalue of the Laplace–Beltrami operator on geodesic balls. If $(M^n,g)$ is a Riemannian manifold, $x_0\in M$, and $B_M(x_0,R)$ is a geodesic ball, one denotes by $\lambda_1(B_M(x_0,R))$ the principal Dirichlet eigenvalue of $-\Delta_g$. Let $\Omega\subset \mathbb{H}^n(-1)$0 be the simply connected space form of constant sectional curvature $\Omega\subset \mathbb{H}^n(-1)$1, and $\Omega\subset \mathbb{H}^n(-1)$2 the corresponding model ball. Cheng’s theorem states that sectional curvature upper bounds yield lower bounds for $\Omega\subset \mathbb{H}^n(-1)$3, sectional curvature lower bounds yield upper bounds, and equality forces isometry with the model ball. Under a Ricci lower bound,

$\Omega\subset \mathbb{H}^n(-1)$4

Cheng proved an upper bound

$\Omega\subset \mathbb{H}^n(-1)$5

for suitable ranges of $\Omega\subset \mathbb{H}^n(-1)$6, again with rigidity (Ariturk, 2016).

In the Laplacian setting, one sometimes speaks of “Cheng’s conjecture” as the expectation that such eigenvalue comparison inequalities, with model space $\Omega\subset \mathbb{H}^n(-1)$7 and curvature bounds, should extend in various directions: to larger classes of domains, under weaker curvature assumptions, and analogously for other elliptic operators (Ariturk, 2016). This usage is programmatic rather than tied to one canonical sentence. It identifies Cheng’s theorem as the prototype of a broader geometric principle: curvature controls principal Dirichlet eigenvalues.

A direct nonlinear realization of this program is Ariturk’s study of the Riemannian Pucci operator $\Omega\subset \mathbb{H}^n(-1)$8. For $\Omega\subset \mathbb{H}^n(-1)$9, the scalar function

$4$0

induces the Riemannian Pucci operator $4$1. The Dirichlet problem

$4$2

has two principal half-eigenvalues $4$3, with corresponding positive and negative eigenfunctions (Ariturk, 2016). Ariturk proves that Cheng’s sectional-curvature comparison extends verbatim to these Pucci half-eigenvalues on geodesic balls of radius less than the injectivity radius, and that the Ricci-lower-bound upper estimate survives on admissible balls. This is explicitly framed as extending Cheng’s type of comparison to a fully nonlinear operator for which variational methods are unavailable (Ariturk, 2016).

A second major extension replaces smooth manifolds by synthetic curvature-dimension spaces. In essentially non-branching $4$4 spaces, the first Dirichlet eigenvalue of a metric ball satisfies the sharp bound

$4$5

where $4$6 is the one-dimensional model eigenvalue associated with the density

$4$7

When $4$8 is an integer, this coincides with Cheng’s classical model-ball eigenvalue. In the $4$9 setting, equality yields rigidity via exact model volume growth and cone-type structure theorems (Luca et al., 31 Jul 2025). This confirms the synthetic metric-measure version of the Cheng-type expectation.

3. Hyperbolic universal inequality

A different formulation of Cheng’s Conjecture concerns universal inequalities for the full Dirichlet spectrum of the Laplacian on bounded domains in hyperbolic space. For the Dirichlet problem

$\frac{4}{n}$0

on a bounded domain $\frac{4}{n}$1, let

$\frac{4}{n}$2

be the Dirichlet eigenvalues. In this context, a universal inequality is one depending only on the dimension $\frac{4}{n}$3, not on the geometry of $\frac{4}{n}$4 beyond its dimension (Luo, 22 Apr 2026).

The relevant background is Yang’s first inequality in Euclidean space and the Cheng–Yang analogues on the sphere and on hyperbolic space. In $\frac{4}{n}$5, Cheng and Yang proved

$\frac{4}{n}$6

The challenging problem is to improve the coefficient $\frac{4}{n}$7 to $\frac{4}{n}$8. The paper "New inequalities for eigenvalues of the Dirichlet Laplacian on the hyperbolic space" states Cheng’s conjecture, following Cheng’s 2017 article, as the assertion that for every bounded domain $\frac{4}{n}$9 and every positive integer $\varepsilon$0,

$\varepsilon$1

The shift $\varepsilon$2 reflects the bottom of the spectrum of $\varepsilon$3 on $\varepsilon$4, while the coefficient $\varepsilon$5 is known to be optimal by Cheng–Yang’s recursion method and Weyl’s asymptotics (Luo, 22 Apr 2026).

The 2026 paper verifies the conjecture up to loss of $\varepsilon$6 for two special kinds of bounded domains (Luo, 22 Apr 2026). In the upper half-space model,

$\varepsilon$7

the first result imposes the condition

$\varepsilon$8

which means that $\varepsilon$9 varies little across $(M^n,g)$0. The second uses

$(M^n,g)$1

which constrains the horizontal slope and places $(M^n,g)$2 near a vertical geodesic. Under either condition, the paper proves

$(M^n,g)$3

Thus the only deviation from the conjectured bound is the multiplicative factor $(M^n,g)$4 (Luo, 22 Apr 2026).

This formulation differs from the geodesic-ball comparison theorem in two respects. First, it concerns all Dirichlet eigenvalues, not only the first. Second, it is universal over bounded domains in hyperbolic space rather than local to geodesic balls. The conjectural sharpness, however, again takes the form of a dimension-only inequality with a curvature-dependent shift.

4. Bergman metric rigidity in several complex variables

In several complex variables, Cheng’s Conjecture refers to a biholomorphic rigidity statement for the Bergman metric. Let $(M^n,g)$5 be a bounded strictly pseudoconvex domain with smooth boundary. The Bergman kernel is

$(M^n,g)$6

for an orthonormal basis $(M^n,g)$7 of $(M^n,g)$8, and the Bergman metric is

$(M^n,g)$9

A Kähler metric $x_0\in M$0 is Kähler–Einstein if

$x_0\in M$1

for some constant $x_0\in M$2 (Sha, 15 Oct 2025).

In this setting, the conjecture is:

Let $x_0\in M$3 be a bounded strictly pseudoconvex domain with $x_0\in M$4 boundary. If the Bergman metric $x_0\in M$5 is Kähler–Einstein, then $x_0\in M$6 is biholomorphic to the unit ball (Sha, 15 Oct 2025).

This is the strictly pseudoconvex smooth-boundary case of a broader picture associated with Yau’s conjecture on domains admitting complete Kähler–Einstein Bergman metrics (Sha, 15 Oct 2025). The conjecture was resolved affirmatively by Huang–Xiao: the Bergman metric of a bounded strictly pseudoconvex domain with $x_0\in M$7-boundary is Kähler–Einstein if and only if the domain is biholomorphic to the unit ball (Sha, 15 Oct 2025).

Recent work extends the statement from the Einstein condition to the constant-scalar-curvature condition. On a bounded strictly pseudoconvex domain with $x_0\in M$8 boundary, if the Bergman metric $x_0\in M$9 has constant scalar curvature, then $B_M(x_0,R)$0 is Kähler–Einstein; if moreover the boundary is $B_M(x_0,R)$1, Huang–Xiao’s theorem implies that the domain is biholomorphic to the unit ball (Sha, 15 Oct 2025). The key identity is

$B_M(x_0,R)$2

and tracing yields

$B_M(x_0,R)$3

Under the cscK hypothesis, $B_M(x_0,R)$4 is harmonic; Diederich’s boundary theorem gives constant boundary value, hence $B_M(x_0,R)$5 is constant and the Bergman metric is Kähler–Einstein (Sha, 15 Oct 2025).

A singular analogue has also been established in complex dimension $B_M(x_0,R)$6. The paper "Bergman-Einstein metrics on two-dimensional Stein spaces" proves that the Bergman metric of the ball quotients $B_M(x_0,R)$7, where $B_M(x_0,R)$8 is a finite and fixed point free group, is Kähler–Einstein if and only if $B_M(x_0,R)$9 is trivial. As a consequence, among $\lambda_1(B_M(x_0,R))$0-dimensional Stein spaces with isolated normal singularities, smooth strongly pseudoconvex boundary, and algebraic CR boundary, the only one with Kähler–Einstein Bergman metric is $\lambda_1(B_M(x_0,R))$1 (Ganguly et al., 2022). This is described there as an algebraic version of Cheng’s conjecture for $\lambda_1(B_M(x_0,R))$2-dimensional Stein spaces.

5. Extensions beyond the classical smooth setting

One of the most active interpretations of Cheng’s Conjecture is that Cheng-type comparison should persist for broader operators and broader spaces.

For fully nonlinear elliptic operators, Ariturk studies the Riemannian Pucci operator and proves a Pucci analogue of Cheng’s theorem. If $\lambda_1(B_M(x_0,R))$3 on a geodesic ball $\lambda_1(B_M(x_0,R))$4, then

$\lambda_1(B_M(x_0,R))$5

while $\lambda_1(B_M(x_0,R))$6 yields the reversed inequalities

$\lambda_1(B_M(x_0,R))$7

with equality forcing isometry to the model ball. Under a Ricci lower bound and an admissibility hypothesis, one obtains the Cheng-type upper bounds

$\lambda_1(B_M(x_0,R))$8

The paper explicitly notes a limitation: with only a Ricci bound and without admissibility, one does not yet know if the full Cheng inequality holds for Pucci eigenvalues (Ariturk, 2016).

For metric measure spaces satisfying synthetic Ricci curvature bounds, the paper "Cheng's eigenvalue comparison on metric measure spaces and applications" proves the sharp upper bound

$\lambda_1(B_M(x_0,R))$9

in essentially non-branching $-\Delta_g$0 spaces, and establishes rigidity in $-\Delta_g$1 spaces (Luca et al., 31 Jul 2025). Equality forces exact model volume growth and leads to a classification into one-dimensional model cases or local isometry to a $-\Delta_g$2-cone over an $-\Delta_g$3 space (Luca et al., 31 Jul 2025).

The same paper derives further spectral consequences. For a compact essentially non-branching $-\Delta_g$4 space of diameter $-\Delta_g$5, the $-\Delta_g$6-th Neumann eigenvalue satisfies

$-\Delta_g$7

For non-compact $-\Delta_g$8 spaces with $-\Delta_g$9 and $\Omega\subset \mathbb{H}^n(-1)$00, the essential spectrum intersects

$\Omega\subset \mathbb{H}^n(-1)$01

The authors further interpret these bounds as upper bounds on the masses of spin-$\Omega\subset \mathbb{H}^n(-1)$02 Kaluza–Klein excitations in warped compactifications of higher-dimensional gravity (Luca et al., 31 Jul 2025). This suggests that Cheng-type comparison has become a template not only for geometric rigidity but also for synthetic spectral analysis and mathematical physics.

6. Methods, rigidity, and unresolved directions

The methods used across the various formulations are notably different, and the differences explain the present boundary between resolved and unresolved cases.

In the original eigenvalue-comparison setting, Cheng’s proof in one paper is based on Barta’s theorem and does not use the Rayleigh–Ritz characterization, whereas the later extension to larger radii under Ricci bounds crucially uses the variational formula (Ariturk, 2016). This distinction remains decisive in nonlinear extensions. For the Pucci operator, variational methods are unavailable; Ariturk replaces them with a Barta-type inequality for principal half-eigenvalues, formulas for the Pucci operator on radial functions,

$\Omega\subset \mathbb{H}^n(-1)$03

and comparison geometry for Jacobi fields via Rauch and Bishop–Gromov (Ariturk, 2016).

The hyperbolic universal-inequality problem uses a different analytic scheme. Its main tool is the Cheng–Yang general inequality for Dirichlet eigenvalues, combined with conformal formulas in the upper half-space model. The proofs use the test functions $\Omega\subset \mathbb{H}^n(-1)$04, $\Omega\subset \mathbb{H}^n(-1)$05, and in the second theorem

$\Omega\subset \mathbb{H}^n(-1)$06

together with integration by parts and an $\Omega\subset \mathbb{H}^n(-1)$07-parameter Cauchy–Schwarz argument. The appearance of the factor $\Omega\subset \mathbb{H}^n(-1)$08 is therefore built into the method, not merely a technical afterthought (Luo, 22 Apr 2026).

The Bergman-metric formulation is driven by complex-analytic invariants rather than spectral comparison. The decisive ingredients are the Cheng–Yau complete Kähler–Einstein metric, boundary asymptotics for the Bergman invariant, harmonicity of $\Omega\subset \mathbb{H}^n(-1)$09 under the cscK assumption, and rigidity statements for complete Kähler–Einstein metrics under cscK perturbations (Sha, 15 Oct 2025). In the two-dimensional singular Stein setting, the proof passes through explicit formulas for the Bergman kernel on $\Omega\subset \mathbb{H}^n(-1)$10, reduction of the Kähler–Einstein condition to a concrete functional equation, and a group-theoretic analysis of finite fixed point free subgroups of $\Omega\subset \mathbb{H}^n(-1)$11 (Ganguly et al., 2022).

The synthetic metric-measure extension rests on localization. An essentially non-branching $\Omega\subset \mathbb{H}^n(-1)$12 space is disintegrated into one-dimensional transport rays,

$\Omega\subset \mathbb{H}^n(-1)$13

and the spectral comparison is reduced to one-dimensional model inequalities on the fibers. Rigidity is then recovered by exact Bishop–Gromov ratios and volume-cone-implies-metric-cone theorems (Luca et al., 31 Jul 2025).

Several open directions remain explicit in the recent literature. In the Pucci setting, extending Ricci-based Cheng inequalities beyond admissible balls remains unresolved (Ariturk, 2016). In the hyperbolic universal-inequality setting, the central task is to remove the factor $\Omega\subset \mathbb{H}^n(-1)$14 and the geometric restrictions on the domain (Luo, 22 Apr 2026). In the Bergman-metric setting, natural problems include moving beyond strictly pseudoconvex domains, weakening regularity assumptions, and extending the singular Stein-space theorem beyond complex dimension $\Omega\subset \mathbb{H}^n(-1)$15 and beyond the algebraic boundary hypothesis (Sha, 15 Oct 2025, Ganguly et al., 2022). In the synthetic spectral setting, a stated open problem is quantitative stability: whether near-equality of $\Omega\subset \mathbb{H}^n(-1)$16 with the model value forces pointed measured Gromov–Hausdorff closeness to one of the rigid model spaces (Luca et al., 31 Jul 2025).

A recurrent misconception is terminological rather than mathematical. Cheng’s Conjecture is unrelated to Chang’s Conjecture in set theory, whose classical form is

$\Omega\subset \mathbb{H}^n(-1)$17

The similarity of names conceals a complete difference of subject matter (Cox, 2019).

Taken together, these developments show that “Cheng’s Conjecture” functions less as a single isolated problem than as a family of sharp comparison-rigidity principles. In spectral geometry it asserts that curvature controls extremal eigenvalues with model-space sharpness; in several complex variables it asserts that the Kähler–Einstein Bergman metric characterizes the ball; and in both smooth and singular settings it continues to organize current work on rigidity, synthetic curvature, and canonical metrics (Ariturk, 2016, Luo, 22 Apr 2026, Sha, 15 Oct 2025, Luca et al., 31 Jul 2025).