The most important examples of dualizable categories which are not compactly generated arise in condensed mathematics, as categories of nuclear modules. These are extracted from large compactly generated categories of “complete” modules, which are not rigid, via the following general construction due to Clausen and Scholze [Scholze (2026)].
Throughout this section, let \(C \in \CAlg (\PrL _{\st })\) be compactly generated with compact unit. We write \(x^{\vee } = \iHom (x,\unit _C)\) for the predual of \(x\).
Definition 6.14 ([Efimov (2025), Definition 1.25]; cf. [Clausen and Scholze (2026), Definition 8.5] and [Scholze (2026), Definitions 13.10 and 13.12]). An object \(M \in C\) is called nuclear if every morphism \(P \to M\) from a compact object \(P \in C^{\omega }\) is trace-class. We denote by \(\Nuc (C) \subseteq C\) the full subcategory of nuclear objects. An object \(M\) is called basic nuclear if \(M \simeq \colim (P_0 \to P_1 \to P_2 \to \cdots )\) with each \(P_n\) compact and each map \(P_n \to P_{n+1}\) trace-class.
Note that a compact object \(P\) is nuclear if and only if \(\id _P\) is trace-class, i.e. if and only if \(P\) is dualizable (Remark 6.6(3)). So the nuclear objects should be thought of as the objects which are “locally dualizable”.
The definition only asks that every morphism \(P \to M\) from a compact object be trace-class, i.e. that a certain map is surjective on \(\pi _0\). We will see that this map is then automatically an equivalence. The key input is the following elementary observation, which says that trace-class morphisms can be “pulled back” through the canonical map \(y \otimes P^{\vee } \to \iHom (P,y)\). For \(x \in C\) and objects \(P, Q \in C\) we write \[ \lambda _x\colon \Map _C(Q, x \otimes P^{\vee }) \to \Map _C(Q \otimes P, x), \qquad s \mapsto (x \otimes \ev _P)(s \otimes P), \] where \(\ev _P\colon P^{\vee } \otimes P \to \unit _C\) is the evaluation. Under the adjunction \(\Map _C(Q \otimes P,x) \simeq \Map _C(Q,\iHom (P,x))\), this is the map induced by the canonical map \(x \otimes P^{\vee } \to \iHom (P,x)\). It is natural in \(x\). For \(Q = \unit _C\), a morphism \(P \to x\) is trace-class if and only if its homotopy class lies in the image of \(\pi _0\lambda _x\), by Remark 6.6(1).
Lemma 6.15. Let \(f\colon x \to y\) be a trace-class morphism in \(C\), and let \(P, Q \in C\). Then there is a map of sets \[ \psi _f\colon \pi _0\Map _C(Q \otimes P, x) \to \pi _0\Map _C(Q, y \otimes P^{\vee }) \] such that \(\lambda _y \circ \psi _f = f \circ (-)\) and \(\psi _f \circ \lambda _x = (f \otimes P^{\vee }) \circ (-)\) on homotopy classes.
Proof.
By Remark 6.6(1), there is a map \(t\colon \unit _C \to y \otimes x^{\vee }\) such that \(f \simeq (y \otimes \ev _x)(t \otimes x)\). For \(b\colon Q \otimes P \to x\), let \(\tilde {b}\colon x^{\vee } \otimes Q \to P^{\vee }\) be the map adjoint to \(\ev _x(x^{\vee } \otimes b)\colon x^{\vee } \otimes Q \otimes P \to \unit _C\), so that \(\ev _P(\tilde {b} \otimes P) \simeq \ev _x(x^{\vee } \otimes b)\), and put \[ \psi _f(b) := (y \otimes \tilde {b})(t \otimes Q)\colon Q \to y \otimes x^{\vee } \otimes Q \to y \otimes P^{\vee }. \] Here and below we suppress the unit and symmetry isomorphisms. Then \[ \lambda _y(\psi _f(b)) = (y \otimes \ev _P)(y \otimes \tilde {b} \otimes P)(t \otimes Q \otimes P) \simeq (y \otimes \ev _x)(y \otimes x^{\vee } \otimes b)(t \otimes Q \otimes P) \simeq (y \otimes \ev _x)(t \otimes x)\,b \simeq f b. \] Conversely, let \(s\colon Q \to x \otimes P^{\vee }\) and \(b = \lambda _x(s) = (x \otimes \ev _P)(s \otimes P)\). Then \(\ev _x(x^{\vee } \otimes b) \simeq (\ev _x \otimes \ev _P)(x^{\vee } \otimes s \otimes P)\), which is also adjoint to \((\ev _x \otimes P^{\vee })(x^{\vee } \otimes s)\colon x^{\vee } \otimes Q \to P^{\vee }\); hence \(\tilde {b} \simeq (\ev _x \otimes P^{\vee })(x^{\vee } \otimes s)\). Therefore \[ \psi _f(\lambda _x(s)) \simeq (y \otimes \ev _x \otimes P^{\vee })(y \otimes x^{\vee } \otimes s)(t \otimes Q) \simeq (y \otimes \ev _x \otimes P^{\vee })(t \otimes s) \simeq (y \otimes \ev _x \otimes P^{\vee })(t \otimes x \otimes P^{\vee })\,s \simeq (f \otimes P^{\vee })\,s. \] □
Lemma 6.16. Let \(M \simeq \colim _{i \in I} M_i\) for a directed poset \(I\) such that for every \(i\) there is \(j \geq i\) with \(M_i \to M_j\) trace-class. Let \(P, Q \in C\) be such that \(Q\) and \(Q \otimes P\) are compact. Then the map \(\hom _C(Q, M \otimes P^{\vee }) \to \hom _C(Q \otimes P, M)\) induced by \(M \otimes P^{\vee } \to \iHom (P,M)\) is an equivalence of spectra.
Proof.
Since \(\pi _n\hom _C(Q,-) = \pi _0\Map _C(\Sigma ^nQ,-)\) for \(n \in \Z \), and \(\Sigma ^nQ\) satisfies the same hypotheses as \(Q\), it suffices to show that \(\pi _0\lambda _M\colon \pi _0\Map _C(Q, M \otimes P^{\vee }) \to \pi _0\Map _C(Q \otimes P, M)\) is a bijection. As \(Q\) and \(Q \otimes P\) are compact and \(- \otimes P^{\vee }\) preserves colimits, and as \(\pi _0\) commutes with filtered colimits, \(\pi _0\lambda _M\) is the colimit over \(i \in I\) of the maps \(\pi _0\lambda _{M_i}\). Write \(t_{ij}\colon M_i \to M_j\) for the transition maps.
Surjectivity: an element of the target is represented by some \(b\colon Q \otimes P \to M_i\). Choose \(j \geq i\) such that \(t_{ij}\) is trace-class. By Lemma 6.15, \(\lambda _{M_j}(\psi _{t_{ij}}(b)) = t_{ij}b\), which represents the same element.
Injectivity: let \(s,s'\colon Q \to M_i \otimes P^{\vee }\) with the same image in the colimit. Then there is \(i' \geq i\) with \(t_{ii'}\lambda _{M_i}(s) = t_{ii'}\lambda _{M_i}(s')\), i.e. \(\lambda _{M_{i'}}(\sigma ) = \lambda _{M_{i'}}(\sigma ')\) for \(\sigma = (t_{ii'} \otimes P^{\vee })s\) and \(\sigma ' = (t_{ii'} \otimes P^{\vee })s'\). Choose \(j \geq i'\) such that \(f = t_{i'j}\) is trace-class. By Lemma 6.15, \((f \otimes P^{\vee })\sigma = \psi _f(\lambda _{M_{i'}}(\sigma )) = \psi _f(\lambda _{M_{i'}}(\sigma ')) = (f \otimes P^{\vee })\sigma '\), so \(s\) and \(s'\) have the same image in \(\pi _0\Map _C(Q, M_j \otimes P^{\vee })\). □
Proposition 6.17 (Clausen–Scholze; see [Clausen and Scholze (2026), Theorem 8.6] and [Efimov (2025), Propositions 1.27 and 1.28]).
- (1)
-
For \(M \in C\), the following are equivalent:
- (a)
-
\(M\) is nuclear;
- (b)
-
for every compact \(P\), the map \(\hom _C(\unit _C,P^{\vee } \otimes M) \to \hom _C(P,M)\) is an equivalence of spectra (equivalently, since we may replace \(P\) by its shifts, the corresponding map of mapping spaces is an equivalence for every compact \(P\));
- (d)
-
\(M \simeq \colim _{i \in I} M_i\) for a directed poset \(I\) and compact objects \(M_i\), such that for every \(i\) there is \(j \geq i\) with \(M_i \to M_j\) trace-class.
Moreover, these conditions are implied by
- (c)
-
for every compact \(P\), the map \(P^{\vee } \otimes M \to \iHom (P,M)\) is an equivalence,
and if the tensor product of any two compact objects of \(C\) is compact, then (c) is equivalent to (a), (b) and (d).
- (2)
-
The subcategory \(\Nuc (C) \subseteq C\) is closed under colimits. It is \(\omega _1\)-presentable, and its \(\omega _1\)-compact objects are precisely the basic nuclear objects.
The main step is to turn the pointwise trace-class condition into a presentation whose transition maps are eventually trace-class. The interleaving lemma then upgrades surjectivity on \(\pi _0\) to the isomorphism of mapping spectra in (b). For (2), we decompose these presentations into countable pieces, as in the proof of \(\omega _1\)-presentability for compactly assembled categories.
Proof.
We first prove (1). (b) \(\Rightarrow \) (a): Taking \(\pi _0\) of the mapping spaces, every morphism \(P \to M\) lies in the image of \(\pi _0\lambda _M\) (with \(Q = \unit _C\)), i.e. is trace-class.
(a) \(\Rightarrow \) (d): Since \(C\) is compactly generated, \(M\) is a filtered colimit of compact objects, and by Theorem A.8 we may write \(M \simeq \colim _{i \in I} M_i\) with \(I\) a directed poset and \(M_i\) compact; denote by \(t_{ij}\) the transition maps and by \(\iota _i\colon M_i \to M\) the canonical maps. Fix \(i\). Since \(\iota _i\) is trace-class, it is \(\lambda _M(w)\) for some \(w \in \pi _0\Map _C(\unit _C, M \otimes M_i^{\vee })\). As \(\unit _C\) is compact and \(M \otimes M_i^{\vee } \simeq \colim _j M_j \otimes M_i^{\vee }\), the class \(w\) is the image of some \(w' \in \pi _0\Map _C(\unit _C, M_j \otimes M_i^{\vee })\) with \(j \geq i\). Then \(g := \lambda _{M_j}(w')\colon M_i \to M_j\) is trace-class, and \(\iota _j g \simeq \lambda _M(w) \simeq \iota _i\) by naturality of \(\lambda \). As \(M_i\) is compact, \(\pi _0\Map _C(M_i,M) \simeq \colim _k \pi _0\Map _C(M_i,M_k)\), so there is \(k \geq j\) with \(t_{jk}g \simeq t_{ik}\). Since trace-class morphisms form an ideal (Remark 6.6(2)), \(t_{ik}\) is trace-class.
(d) \(\Rightarrow \) (b): This is Lemma 6.16 with \(Q = \unit _C\).
(c) \(\Rightarrow \) (b): Apply \(\hom _C(\unit _C,-)\) and use \(\hom _C(\unit _C,\iHom (P,M)) \simeq \hom _C(P,M)\).
(d) \(\Rightarrow \) (c), if compact objects are closed under tensor products: For compact \(Q\), the map \(\hom _C(Q,P^{\vee } \otimes M) \to \hom _C(Q,\iHom (P,M)) \simeq \hom _C(Q \otimes P,M)\) is an equivalence by Lemma 6.16. Since \(C\) is compactly generated, the compact objects detect equivalences, so \(P^{\vee } \otimes M \to \iHom (P,M)\) is an equivalence.
We now prove (2). By (1), \(M\) is nuclear if and only if the maps in (b) are equivalences. Since \(\unit _C\) and \(P\) are compact, the functors \(\hom _C(\unit _C,P^{\vee } \otimes -)\) and \(\hom _C(P,-)\) preserve filtered colimits, and being exact they preserve all small colimits. Hence \(\Nuc (C)\) is closed under colimits in \(C\); in particular, colimits in \(\Nuc (C)\) are computed in \(C\).
Basic nuclear objects are nuclear by (d) (with \(I = \N \)). They are countable colimits of compact objects, hence \(\omega _1\)-compact in \(C\) (Corollary A.4), and hence also in \(\Nuc (C)\).
Next, every nuclear \(M\) is an \(\omega _1\)-filtered colimit of basic nuclear objects. Write \(M \simeq \colim _{i \in I} M_i\) as in (d), and choose for every \(i\) an element \(\sigma (i) \geq i\) such that \(M_i \to M_{\sigma (i)}\) is trace-class. Let \(\Sigma _{\sigma }\) be the set of countable directed subsets \(J \subseteq I\) with \(\sigma (J) \subseteq J\), ordered by inclusion. The proof of Proposition A.9 applies verbatim to \(\Sigma _{\sigma }\) instead of \(\Sigma \) (adding in the inductive construction also \(\sigma (u)\) for \(u \in U_k\)), and shows that \(\Sigma _{\sigma }\) is \(\omega _1\)-filtered and that \(M \simeq \colim _{J \in \Sigma _{\sigma }} M_J\) with \(M_J = \colim _{i \in J} M_i\). Each \(M_J\) is basic nuclear: enumerate \(J = \{q_0,q_1,\dots \}\), and choose inductively \(j_0 = q_0\) and \(j_{n+1} \in J\) with \(j_{n+1} \geq \sigma (j_n), q_{n+1}\). Then \(n \mapsto j_n\) is cofinal (Lemma A.7), so \(M_J \simeq \colim _n M_{j_n}\), and each \(M_{j_n} \to M_{j_{n+1}}\) factors through the trace-class map \(M_{j_n} \to M_{\sigma (j_n)}\), hence is trace-class.
The basic nuclear objects form an essentially small collection, since \(C^{\omega }\) is essentially small. Thus \(\Nuc (C)\) admits small colimits and is generated under \(\omega _1\)-filtered colimits by an essentially small collection of \(\omega _1\)-compact objects, so it is \(\omega _1\)-accessible by [Lurie (2009), Proposition 5.4.2.2], and hence \(\omega _1\)-presentable.
Finally, let \(X \in \Nuc (C)\) be \(\omega _1\)-compact. Writing \(X \simeq \colim _{J \in \Sigma _{\sigma }} X_J\) as above, the identity of \(X\) factors through some \(X_J\), so \(X\) is a retract of a basic nuclear object \(Y = \colim _n a_n\), where the \(a_n\) are compact and the transition maps \(a_n \to a_{n+1}\) are trace-class. Now we argue as in the proof of Proposition A.13, applied in \(C \simeq \Ind (C^{\omega })\): by Lemma A.11 and Lemma A.12, we find a strictly increasing \(f\colon \N \to \N \) with \(f(n) \geq n\) and a natural transformation \(\eta \colon a_{\bullet } \to a_{f(\bullet )}\) such that \(X \simeq \colim _n \Phi (n,n)\), where \(\Phi (k,n) = a_{f^k(n)}\). The transition map \(\Phi (n,n) \to \Phi (n+1,n+1)\) is the composite of the transition map \(a_{f^n(n)} \to a_{f^n(n+1)}\) of \(a_{\bullet }\), which is trace-class as \(f^n(n+1) > f^n(n)\), with \(\eta _{f^n(n+1)}\). So it is trace-class, and \(X\) is basic nuclear. □
Remark 6.18. Proposition 6.17 shows that the condition in Definition 6.14, which only asks that the maps \(\pi _0\Map _C(\unit _C,P^{\vee } \otimes M) \to \pi _0\Map _C(P,M)\) be surjective, already implies that \(\hom _C(\unit _C,P^{\vee } \otimes M) \to \hom _C(P,M)\) is an equivalence. Clausen–Scholze [Clausen and Scholze (2026), Definition 8.5] instead define nuclearity by the latter condition (and [Scholze (2026), Definition 13.10] by an analogous condition for analytic rings), whereas Definition 6.14 follows [Efimov (2025), Definition 1.25]. The two definitions agree; the injectivity comes from the fact that, by (a) \(\Rightarrow \) (d), \(M\) is a filtered colimit of compact objects along trace-class maps, for which the two sides become interleaved (Lemma 6.16). The hypothesis on compact objects in (1)(c) holds for instance for the derived categories of solid and liquid modules, which are generated by compact objects closed under tensor products.
In general, \(\Nuc (C)\) need not be rigid, or even dualizable. The following criterion is the one that is used in practice.
Proposition 6.19 ([Efimov (2025), Propositions 1.23 and 1.29], [Ramzi (2024b), Construction 4.79 and Theorem 4.81]). Suppose that every trace-class morphism between compact objects of \(C\) is a composite of two trace-class morphisms between compact objects. Then \(\Nuc (C)\) is a rigid symmetric monoidal category. More precisely, \(\Nuc (C)\) is equivalent to the rigidification \(C^{\rig }\) of \(C\), the value at \(C\) of the right adjoint to the inclusion of rigid categories into \(\CAlg (\PrL _{\st ,\kappa })\), for an uncountable regular cardinal \(\kappa \) for which \(C\) belongs to this category. Here the morphisms preserve \(\kappa \)-compact objects, and the tensor product preserves them separately. This value is independent of the sufficiently large choice of \(\kappa \) [Efimov (2025), Proposition 1.23].
Sketch.
By [Efimov (2025), Proposition 1.23 and Remark 1.24(i)], based on [Ramzi (2024b), Theorem 4.81 and Corollary 4.75], the rigidification \(C^{\rig } \subseteq C\) is the full subcategory generated under colimits by colimits \(\colim _{a \in \Q } F(a)\) of diagrams \(F\colon (\Q ,\leq ) \to C^{\omega }\) in which every map \(F(a) \to F(b)\), \(a < b\), is trace-class. Such colimits are nuclear by Proposition 6.17(1)(d), so \(C^{\rig } \subseteq \Nuc (C)\). Conversely, let \(P_0 \to P_1 \to \cdots \) be a sequence of trace-class maps between compact objects. By the assumption and Lemma 4.7(2), applied to \(D = C^{\omega }\) and the class of trace-class morphisms (which is closed under composition by Remark 6.6(2)), it extends to a functor \(F\colon \Q _{\geq 0} \to C^{\omega }\) with trace-class transition maps and the same colimit. Restricting \(F\) to the cofinal subset \(\Q _{>0} \cong \Q \), we see that the basic nuclear objects lie in \(C^{\rig }\). By Proposition 6.17(2), these generate \(\Nuc (C)\). □
Remark 6.20. Compare this with the construction of \((C,S)^{\dual }\) in Clausen’s lectures: given a class \(S\) of morphisms forming an ideal with \(S = S^2\) (and some further closure properties), one obtains a dualizable category generated by colimits along morphisms in \(S\). For \(S\) the trace-class morphisms of \(C\), this gives \(C^{\rig }\). The condition \(S = S^2\) in Proposition 6.19 is what allows one to work with sequential colimits instead of \(\Q \)-indexed ones (Lemma 4.7; for compact morphisms in compactly assembled categories, see Proposition 4.8), cf. also the construction of \(C_S\) in Lecture 7 of the Bonn workshop.
Examples from condensed mathematics
We now describe the main examples, without proofs. We refer to [Scholze (2026)] for the solid and liquid theory, to [Clausen and Scholze (2026)] for the gaseous theory, and to [Andreychev (2023)] and [Efimov (2025), Section 7.2] for the comparison with Efimov’s theory.
Example 6.21 (Solid modules and nuclear modules over formal schemes). Ambient category. Let \(\D (\mathrm {Solid}_{\Z })\) be the derived category of light solid abelian groups, as in Efimov’s convention in [Efimov (2025), Section 7.2]. This is a compactly generated symmetric monoidal category, with compact projective generator \(\prod _{\N }\Z \) and compact unit \(\Z \). It is far from rigid: the compact object \(\prod _{\N }\Z \) is not dualizable (its predual is \(\bigoplus _{\N }\Z \)). In fact, there is a symmetric monoidal equivalence \[ \D (\mathrm {Solid}_{\Z }) \simeq \Ind \big (\D ^b(\mathrm {Proj}^{\omega _1}\text {-}\Z )\catop \big ), \] where \(\D ^b(\mathrm {Proj}^{\omega _1}\text {-}\Z ) \subseteq \D (\Z )\) is the stable subcategory generated by \(\bigoplus _{\N }\Z \), under which \(\prod _{\N }\Z \) corresponds to \((\bigoplus _{\N }\Z )\catop \) [Efimov (2025), Section 7.2].
Completion and nuclear objects. More generally, let \(R\) be a commutative ring and \(I = (a_1,\dots ,a_m) \subseteq R\) a finitely generated ideal, and let \(R^{\wedge }_I\) denote its derived \(I\)-adic completion. We use the induced solid analytic structure, with the notation of [Efimov (2025), Section 7.2]. All completions and tensor products in this example are derived; for Noetherian \(R\), the derived completion of \(R\) agrees with its ordinary completion. Its category of modules \(\D (\mathrm {Solid}_{(R^{\wedge }_I,\Z )}) = \Mod _{R^{\wedge }_I}(\D (\mathrm {Solid}_{\Z }))\) is compactly generated by \(R^{\wedge }_I \otimes ^{\square } \prod _{\N }\Z \), with compact unit. The category of nuclear solid modules is \[ \Nuc ^{CS}(R^{\wedge }_I) := \Nuc \big (\D (\mathrm {Solid}_{(R^{\wedge }_I,\Z )})\big ), \] with respect to this specified solid structure. (For \(R = \Z \) and \(I = 0\) one recovers \(\D (\Z )\): an object of \(\D (\mathrm {Solid}_{\Z })\) is nuclear if and only if it is discrete [Clausen and Scholze (2026), Lecture VIII, Exercise 1].) One has \(\Nuc ^{CS}(R^{\wedge }_I) \simeq \D (\mathrm {Solid}_{(R^{\wedge }_I,\Z )})^{\rig }\).
Rigidity and invariants. The key technical input is that the preduals of compact objects are nuclear, which goes back to Andreychev [Andreychev (2023), Lemma 3.8 and Korollar 3.14]; see [Efimov (2025), Propositions 7.4 and 7.5] (Efimov deduces the rigidity from this via [Efimov (2025), Propositions 1.33 and 1.34] rather than via Proposition 6.19). The tensor products here are those of the complete solid module category; the cited propositions supply the nuclearity and rigidity statements. Hence:
- \(\Nuc ^{CS}(R^{\wedge }_I)\) is rigid, hence dualizable, but in general not compactly generated.
- Its compact objects are the perfect complexes over \(R^{\wedge }_I\).
- Its continuous K-theory is \(K^{\cont }(\Nuc ^{CS}(R^{\wedge }_I)) \simeq \lim _n K(R \quot (a_1^n,\dots ,a_m^n))\), which is \(\lim _n K(R/I^n)\) if \(R\) is Noetherian (Efimov [Efimov (2025), Corollary 7.10], as conjectured by Clausen–Scholze), cf. Theorem 1.15.
Failure of compact generation. To see that compact generation fails already for \(\Nuc ^{CS}(\Z _p)\), suppose otherwise. Its compact objects are \(\Perf (\Z _p)\), so the canonical \(\Z \)-linear functor \(\D (\Z _p) \to \Nuc ^{CS}(\Z _p)\) would be an equivalence. But Efimov computes \[ \mathrm {HH}^{\cont }(\Nuc ^{CS}(\Z _p)/\Z ) \simeq \Z _p \] [Efimov (2025), Corollary 7.20(i)], whereas \[ \mathrm {HH}_1(\Z _p/\Z ) \cong \Omega ^1_{\Z _p/\Z } \cong \Omega ^1_{\Q _p/\Q } \neq 0. \] Here \(\mathrm {HH}^{\cont }\) denotes the continuous extension of relative Hochschild homology, which agrees with ordinary Hochschild homology on compactly generated categories. This gives the required contradiction; see also the discussion after [Efimov (2025), Remark 7.21].
Operator-algebra presentation. There is also an operator-algebra presentation, analogous to the compact-operator example of Example 1.14. Let \(R\) be a commutative ring, \(I = (f_1,\dots ,f_m) \subseteq R\) a finitely generated ideal. In this paragraph all completions and tensor products are derived, following [Efimov (2025), Section 7.2]. Let \(P := (\bigoplus _{\N } R)^{\wedge }_I\) be the \(I\)-completed countable direct sum, which plays the role of a separable Hilbert space. In place of \(B(\Hh )\) we take the endomorphism ring \(\End _R(P)\), and in place of the Calkin algebra \(B(\Hh )/\Kk \) the \(I\)-adically completed endomorphism ring \[ \Calk ^{\mathrm {top}}_R(P) := \End _{\Calk (R)}(P)^{\wedge }_I \] of \(P\) in the Calkin category of \(\Perf (R)\). The fiber of \(\End _R(P) \to \Calk ^{\mathrm {top}}_R(P)\) is \(\big (\prod _{\N }R \otimes _R \bigoplus _{\N }R\big )^{\wedge }_I\), which one may think of as the \(I\)-adic completion of the endomorphisms of finite rank, just as \(\Kk \) is the norm closure of the operators of finite rank. Efimov shows that \(\End _R(P) \to \Calk ^{\mathrm {top}}_R(P)\) is a homological epimorphism, whose kernel is the category of nuclear modules over \(R^{\wedge }_I\) in the sense of Clausen–Scholze [Efimov (2025), Corollary 7.7]: \[ \Nuc ^{CS}(R^{\wedge }_I) \simeq \D \big (\End _R(P) \to \Calk ^{\mathrm {top}}_R(P)\big ). \]
Efimov’s variant. Efimov also introduced a variant \(\Nuc (R^{\wedge }_I)\) [Efimov (2025), Definition 3.28], which can be described as \(\lim ^{\dual }_n \D (R \quot (a_1^n,\dots ,a_m^n))\), the limit computed in \(\Pr ^{\dual }\) [Efimov (2025), Proposition 5.23]. It receives a fully faithful strongly continuous functor from \(\Nuc ^{CS}(R^{\wedge }_I)\) [Efimov (2025), Corollary 7.6], which induces an equivalence on compact objects, both sides having compact objects \(\Perf (R^{\wedge }_I)\) [Efimov (2025), Proposition 3.29]. It can equivalently be described as the rigidification \(\D _{I\text {-}\mathrm {compl}}(R)^{\rig }\) [Efimov (2025), Proposition 4.3] of the category of derived \(I\)-complete \(R\)-modules; the latter need not itself be rigid. For example, for \(R = \Z \) and \(I = (p)\) its unit \(\Z _p\) is not compact.
Example 6.22 (Banach spaces as nuclear modules). Heuristically, nuclear modules are “Banach-type” objects, as opposed to the compact objects such as \(\prod _{\N }\Z _p\), which are of “profinite” or “compact” type. For instance, for every profinite set \(S\) the module \(C(S,\Z _p)\) of continuous functions is nuclear over \(\Z _p^{\square }\) [Andreychev (2023), Lemma 3.8]. Taking \(S = \N \cup \{\infty \}\), it follows that the \(p\)-adic Banach space \(c_0(\N ,\Z _p)\) of null sequences, a direct summand of \(C(S,\Z _p)\), is nuclear, and inverting \(p\) gives nuclear solid \(\Q _p\)-vector spaces such as \(c_0(\N ,\Q _p)\). In contrast, the compact object \(\prod _{\N }\Z _p\) of \(\D (\mathrm {Solid}_{\Z _p})\) is not nuclear: a compact nuclear object is dualizable, and the dualizable objects of \(\D (\mathrm {Solid}_{\Z _p})\) are just the perfect complexes of \(\Z _p\)-modules [Scholze (2026), Theorem 14.10].
Example 6.23 (Liquid and gaseous modules). Liquid modules. Over the real and complex numbers, the solid theory is replaced by the liquid theory: for \(0 < p \leq 1\) there is a compactly generated symmetric monoidal category \(\D (\mathrm {Liq}_p(\R ))\) of \(p\)-liquid \(\R \)-vector spaces [Scholze (2026)], and one can form \(\Nuc (\R ^{p\text {-}\mathrm {liq}})\). The nuclear objects of \(\D (\mathrm {Liq}_p)\) do not depend on \(p\) [Clausen and Scholze (2026), Corollary 8.18].
The lecture model. According to Clausen [Clausen (2023b)], there is an explicit model for the rigidification of the category of liquid \(\C \)-vector spaces in terms of operator algebras: \[ \D ^{\mathrm {liq}}(\C )^{\rig } \simeq \D \big (B(\Hh ) \to B(\Hh )/S_{0+}\big ), \] Here \(S_{0+} = \bigcap _{p > 0} S_p\) is the ideal of operators which are \(p\)-summable for all \(p > 0\). The kernel on the right and its Hilbert-space model are constructed in Proposition 5.13.
Gaseous modules. The gaseous theory of Clausen–Scholze [Clausen and Scholze (2026)] provides yet another analytic ring structure on \(\C \), with a category of nuclear modules \(\Nuc (\C ^{\mathrm {gas}})\). Rigidification produces rigid, hence dualizable, categories whose continuous K-theory is a candidate for analytic K-theory, cf. Section 1.5. Failure of compact generation is a separate assertion; we do not deduce it from rigidity.
Remark 6.24. The reason for working with \(\Nuc \) rather than with all complete (solid) modules is that the former has much better formal properties: for instance, it is rigid, hence dualizable with a good continuous K-theory, and \(\operatorname {Spa}(A,A^+) \mapsto \Nuc (A)\) satisfies descent for the analytic topology on affinoid adic spaces (Andreychev), see Section 1.5.