Growth, complexity, and dimensionality of the type IIB/F-theory flux-vacuum locus: the Grimm–Monnee counting conjectures

The number of flux vacua of type IIB/F-theory compactifications on a Calabi–Yau fourfold with bounded flux tadpole is known to be finite: finiteness for Hodge vacua (vanishing flux superpotential) follows from the theorem of Cattani–Deligne–Kaplan on the locus of Hodge classes, and finiteness for general self-dual vacua was proven by Bakker–Grimm–Schnell–Tsimerman using o-minimal geometry. The open problem is the finer quantitative structure of the vacuum locus that Grimm and Monnee (JHEP 08 (2024) 039) formulated as three mathematically precise conjectures: (1) for variations of Hodge structure of level at least 3, the number of Hodge classes with self-intersection L at points isolated in their Hodge-tensor locus grows sub-polynomially in the tadpole bound L; (2) the self-dual vacuum locus is definable in a sharply o-minimal structure with format O(h^{3,1}) and degree poly(L); and (3) a generalized tadpole conjecture asserting that, for sufficiently many moduli relative to L, every connected component of the bounded Hodge-class locus has positive dimension.

Background

In flux compactifications of type IIB string theory, and in their F-theory formulation on a Calabi–Yau fourfold $Y_4$, the four-dimensional effective theory depends on the complex-structure moduli of $Y_4$ and on a background four-form flux $G_4\in H^4(Y_4,\mathbb{Z})$ (an integral cohomology class; integrality implements flux quantization). The flux couples to the moduli through the superpotential $W_{\rm flux}=\int_{Y_4}G_4\wedge\Omega$, with $\Omega$ the holomorphic four-form, and supersymmetric Minkowski vacua are solutions of $D_i W_{\rm flux}=0$. Consistency with the dual M-theory description imposes the tadpole cancellation condition $N_{D3}+\frac{1}{2}\int G_4\wedge G_4=\chi(Y_4)/24$ ($\chi$ is the Euler characteristic), so the flux self-intersection $(v,v)=\int v\wedge v$ is bounded by an integer tadpole bound $L$ of order $\chi(Y_4)/24$.

Mathematically the setting is a polarized integral variation of Hodge structure (VHS) $E\to\mathcal{M}$: over each point $z$ of the complex-structure moduli space $\mathcal{M}$ (of dimension $h^{3,1}$) the fiber carries the primitive middle cohomology with its Hodge decomposition $H^{4,0}\oplus H^{3,1}\oplus H^{2,2}\oplus H^{1,3}\oplus H^{0,4}$ and the intersection pairing. Two classes of vacua arise. A Hodge vacuum ($D_iW=0$, $W=0$) has flux of pure type $(2,2)$ — a Hodge class. A self-dual vacuum ($D_iW=0$, $W\neq 0$) has flux that is an eigenvector of the Weil operator $C$ (the algebraic avatar of the Hodge star, acting as $Cv=i^{p-q}v$ on $H^{p,q}$); for weight 4 this means the flux has no $(3,1)$ or $(1,3)$ component. The locus of Hodge classes with self-intersection at most $L$ is denoted $E_{\rm Hodge}(L)$; its connected components count Hodge vacua, possibly with flat directions.

Bare finiteness of the vacuum count for bounded tadpole — the question made famous by early landscape studies — is now a theorem. Cattani, Deligne and Kaplan (1995) proved that $E_{\rm Hodge}(L)$ is an algebraic variety, finite over $\mathcal{M}$, which implies finitely many Hodge vacua. Bakker, Grimm, Schnell and Tsimerman (2021, published 2023) extended finiteness to the locus of self-dual classes with fixed self-intersection; their proof uses the definability of period maps in an o-minimal structure (an expansion of the real field in which every definable set has finitely many connected components — a model-theoretic tameness condition) established by Bakker, Klingler and Tsimerman. (The period map assigns to each point of moduli space the Hodge decomposition, i.e. the multivalued periods, of the cohomology of the fiber; its definability is the statement that this transcendental object is built from finitely many tame analytic operations.) Grimm and Monnee (2024) review these results, reprove finiteness of self-dual vacua within a controlled asymptotic approximation via new multi-variable asymptotics of Hodge inner products near boundaries of $\mathcal{M}$, and then pose the sharper questions that remain open.

The open questions replace heuristic censuses by rigorous ones. The classical counting heuristics of Ashok–Douglas and Denef–Douglas treat the flux as continuous and estimate the number of vacua by the volume of a high-dimensional flux ball, producing the famous $\sim 10^{500}$-type estimates; Douglas, Shiffman and Zelditch made related continuum statements rigorous, but still with the quantization condition relaxed. For Hodge vacua the quantization is expected to matter drastically: the vacuum equations are overdetermined ($h^{3,1}+1$ equations for $h^{3,1}$ moduli) and highly transcendental (they involve the periods of $Y_4$), so integer solutions should be rare. The Pila–Wilkie theorem — rational points of bounded height on the transcendental part of a definable set grow slower than any positive power of the height — is the model for this expectation (the height of an integral flux is $\max_i|v_i|$).

Two refinements are needed to turn this intuition into a defensible statement. First, Baldi, Klingler and Ullmo proved that for a VHS of level at least 3 — the level is a measure of the complexity of the variation, defined through the adjoint Mumford–Tate Lie algebra (the Mumford–Tate group being the smallest real algebraic group acting on the cohomology whose invariant tensors are exactly the Hodge tensors) and related to the length of the Hodge filtration; it is not the weight: a generic Calabi–Yau fourfold has level 4, while products such as $K3\times K3$ or $Y_3\times T^2$ have level 1 — the Hodge locus (the locus where additional Hodge tensors, i.e. rational classes of pure type on tensor powers of the fiber, appear) is a finite union of algebraic special subvarieties (irreducible algebraic subvarieties of the base that are maximal loci of additional Hodge tensors), at least after restricting to components of positive period dimension, meaning components whose image under the period map has positive dimension; the isolated, zero-period-dimension case is exactly the exception their methods leave open. Second, Grimm and van de Heisteeg (2024), studying the Hulek–Verrill fourfold, showed that on loci with enhanced Hodge tensors the restricted variation has reduced level, the vacuum conditions become algebraic, and vacua become dense when the tadpole bound is removed; this falsified the original form of the counting conjecture and forced its refinement to points isolated in their Hodge-tensor locus. In parallel, the tadpole conjecture of Bena, Blåbäck, Graña and Lüst postulates that stabilizing all $h^{3,1}$ complex-structure moduli requires a tadpole growing linearly in $h^{3,1}$; it has been supported in asymptotic regimes of moduli space (Graña et al.) and tested on Fermat-type and Landau–Ginzburg models, with the Landau–Ginzburg tests producing both confirmations and tensions for stronger variants. Finally, sharply o-minimal structures (Binyamini–Novikov–Zak) refine o-minimality by endowing definable sets with a complexity pair (format $F$, degree $D$) that controls the number of zeros of definable functions, in analogy with the degree of a polynomial, offering a candidate route to rigorous census bounds for self-dual vacua. What is not known is whether the three precise conjectures Grimm and Monnee formulated to consolidate these threads are true.

Problem Statement

Prove or disprove, in their stated generality, the three conjectures that Grimm and Monnee (arXiv:2311.09295, Section 4) formulated for the locus of flux vacua of type IIB/F-theory compactifications on Calabi–Yau fourfolds — equivalently, for the Hodge-theoretic loci of an even-weight polarized integral variation of Hodge structure $E\to\mathcal{M}$. Throughout, $\mathbb{H}_{\mathbb{Z}}$ denotes the integral local system of the variation (the lattice of integral classes carried by each fiber) and $(v,v)$ the polarization pairing. The three statements jointly quantify one objective: characterize how the vacuum census and the organization of the vacuum locus scale with the tadpole bound $L$ and the number of moduli, beyond the already-proven fact of finiteness.

Conjecture 1 (counting growth). Let $E\to\mathcal{M}$ be a variation of polarized Hodge structure of even weight $2k$ whose level is at least 3. Fix a positive integer $L$ and consider the locus of Hodge classes with self-intersection $L$,

$$\hat E_{\rm Hodge}(L)=\{(z,v)\in E:\ v\in H^{k,k}\cap\mathbb{H}_{\mathbb{Z}},\ (v,v)=L\},$$

and denote by $\hat E^{\rm iso}_{\rm Hodge}(L)$ the subset of points $(z,v)$ for which $z$ is an isolated point of its locus of Hodge tensors (the Hodge tensors at $z$ are the rational classes of pure Hodge type on the fiber over $z$ and its tensor powers; the locus of Hodge tensors of $z$ is the smallest special subvariety containing $z$, i.e. the common locus of all Hodge tensors present at $z$, and $z$ is isolated when this locus is zero-dimensional). The conjecture states that for every $\epsilon>0$ there exists $C>0$, independent of $L$, such that

$$\#\hat E^{\rm iso}_{\rm Hodge}(L)<C\,L^{\epsilon}.$$

Conjecture 2 (sharp complexity). The locus of self-dual flux vacua with tadpole bound $L$ is definable in a sharply o-minimal structure, and its sharp complexity, in the sense of the format $F$ and degree $D$ assigned to definable sets in such structures, satisfies $D=\mathrm{poly}(L)$ and $F=O(h^{3,1})$, where $h^{3,1}=\dim\mathcal{M}$ is the number of complex-structure moduli.

Conjecture 3 (generalized tadpole conjecture for the Hodge locus). Let $E\to\mathcal{M}$ be a variation of polarized Hodge structure of even weight $2k$ and

$$E_{\rm Hodge}(L)=\{(z,v)\in E:\ v\in H^{k,k}\cap\mathbb{H}_{\mathbb{Z}},\ (v,v)\le L\}.$$

There exist constants $C_1,C_2>0$, independent of $L$ and $\dim\mathcal{M}$ (but possibly depending on other data of the variation, such as the weight), such that if $\dim\mathcal{M}>C_1$ and $\dim\mathcal{M}>C_2\,L$, then every connected component of $E_{\rm Hodge}(L)$ has strictly positive dimension. When the variation arises from the middle cohomology of a family of Calabi–Yau fourfolds, $C_2$ is expected to be of order one.

The three conjectures are separate statements about the vacuum locus, and each can be resolved on its own, in whole or in part. Bare finiteness of the vacuum count for bounded $L$ is already a theorem (Cattani–Deligne–Kaplan for Hodge classes; Bakker–Grimm–Schnell–Tsimerman for self-dual classes).

References

1. T. W. Grimm and J. Monnee, "Finiteness Theorems and Counting Conjectures for the Flux Landscape", JHEP 08 (2024) 039, arXiv:2311.09295, DOI 10.1007/JHEP08(2024)039, https://arxiv.org/abs/2311.09295
2. B. Bakker, T. W. Grimm, C. Schnell and J. Tsimerman, "Finiteness for self-dual classes in integral variations of Hodge structure", Épijournal de Géométrie Algébrique, Special volume in honour of C. Voisin (2023), arXiv:2112.06995, DOI 10.46298/epiga.2023.specialvolumeinhonourofclairevoisin.9626, https://arxiv.org/abs/2112.06995
3. B. Bakker, B. Klingler and J. Tsimerman, "Tame topology of arithmetic quotients and algebraicity of Hodge loci", J. Amer. Math. Soc. 33 (2020) 917–939, arXiv:1810.04801, DOI 10.1090/jams/952, https://arxiv.org/abs/1810.04801
4. E. Cattani, P. Deligne and A. Kaplan, "On the Locus of Hodge Classes", J. Amer. Math. Soc. 8 (1995) 483–506, arXiv:alg-geom/9402009, DOI 10.1090/s0894-0347-1995-1273413-2, https://arxiv.org/abs/alg-geom/9402009
5. S. Ashok and M. R. Douglas, "Counting Flux Vacua", JHEP 01 (2004) 060, arXiv:hep-th/0307049, DOI 10.1088/1126-6708/2004/01/060, https://arxiv.org/abs/hep-th/0307049
6. F. Denef and M. R. Douglas, "Distributions of flux vacua", JHEP 05 (2004) 072, arXiv:hep-th/0404116, DOI 10.1088/1126-6708/2004/05/072, https://arxiv.org/abs/hep-th/0404116
7. M. R. Douglas, B. Shiffman and S. Zelditch, "Critical Points and Supersymmetric Vacua I", Commun. Math. Phys. 252 (2004) 325–358, arXiv:math/0402326, DOI 10.1007/s00220-004-1228-y, https://arxiv.org/abs/math/0402326
8. G. Baldi, B. Klingler and E. Ullmo, "On the distribution of the Hodge locus", Inventiones mathematicae 235 (2024) 441–487, arXiv:2107.08838, DOI 10.1007/s00222-023-01226-0, https://arxiv.org/abs/2107.08838
9. T. W. Grimm and D. van de Heisteeg, "Exact flux vacua, symmetries, and the structure of the landscape", JHEP 01 (2025) 005, arXiv:2404.12422, DOI 10.1007/JHEP01(2025)005, https://arxiv.org/abs/2404.12422
10. I. Bena, J. Blåbäck, M. Graña and S. Lüst, "The Tadpole Problem", JHEP 11 (2021) 223, arXiv:2010.10519, DOI 10.1007/JHEP11(2021)223, https://arxiv.org/abs/2010.10519
11. M. Graña, T. W. Grimm, D. van de Heisteeg, A. Herraez and E. Plauschinn, "The Tadpole Conjecture in Asymptotic Limits", JHEP 08 (2022) 237, arXiv:2204.05331, DOI 10.1007/JHEP08(2022)237, https://arxiv.org/abs/2204.05331
12. E. Plauschinn and L. Schlechter, "Flux vacua of the mirror octic", JHEP 01 (2024) 157, arXiv:2310.06040, DOI 10.1007/JHEP01(2024)157, https://arxiv.org/abs/2310.06040
13. M. Rajaguru, A. Sengupta and T. Wrase, "Fully stabilized Minkowski vacua in the $2^6$ Landau-Ginzburg model", JHEP 10 (2024) 095, arXiv:2407.16756, DOI 10.1007/JHEP10(2024)095, https://arxiv.org/abs/2407.16756
14. A. P. Braun, H. Fortin, D. Lopez Garcia and R. Villaflor Loyola, "More on $G$-flux and General Hodge Cycles on the Fermat Sextic", JHEP 06 (2024) 046, arXiv:2401.00470, DOI 10.1007/JHEP06(2024)046, https://arxiv.org/abs/2401.00470
15. H. Fortin and D. López Garcia, "Hodge theory and $G_4$ fluxes in weighted projective spaces: Galois action", arXiv:2606.05530, https://arxiv.org/abs/2606.05530
16. D. Urbanik, "Degrees of Hodge Loci", arXiv:2412.08924, https://arxiv.org/abs/2412.08924
17. N. Khelifa and D. Urbanik, "Existence and density of typical Hodge loci", arXiv:2303.16179, https://arxiv.org/abs/2303.16179
18. N. Khelifa, "On the distribution of mixed Hodge loci", arXiv:2603.20272, https://arxiv.org/abs/2603.20272
19. S. Lüst, "An index for flux vacua", JHEP 01 (2026) 144, arXiv:2405.04584, DOI 10.1007/JHEP01(2026)144, https://arxiv.org/abs/2405.04584
20. A. Chauhan, M. Cicoli, S. Krippendorf, A. Maharana, P. Piantadosi and A. Schachner, "Deep observations of the Type IIB flux landscape", JHEP 07 (2025) 271, arXiv:2501.03984, DOI 10.1007/JHEP07(2025)271, https://arxiv.org/abs/2501.03984
21. G. Binyamini, D. Novikov and B. Zak, "Sharply o-minimal structures and sharp cellular decomposition", arXiv:2209.10972, https://arxiv.org/abs/2209.10972
22. J. Pila and A. J. Wilkie, "The rational points of a definable set", Duke Mathematical Journal 133 (2006) 591–616, DOI 10.1215/s0012-7094-06-13336-7, https://doi.org/10.1215/s0012-7094-06-13336-7
23. A. Harrison-Migochi and R. McCulloch, "Sharp o-minimality and lattice point counting", arXiv:2503.01731, https://arxiv.org/abs/2503.01731
24. O. Carmon, "Analytically generated sharply o-minimal structures", arXiv:2604.06144, https://arxiv.org/abs/2604.06144
25. T. W. Grimm, D. Prieto and M. van Vliet, "Tame Embeddings, Volume Growth, and Complexity of Moduli Spaces", arXiv:2503.15601, https://arxiv.org/abs/2503.15601