← All problems

Ultrapowers (problem 1)

Does every separable C*-algebra embed into an ultrapower of O2\mathcal O_2 with respect to an ultrafilter on N\mathbb N? Here an ultrapower associated to an ultrafilter U\mathcal U is

O2U=ℓ∞(N,O2)/{(xn):lim⁡n→U∥xn∥=0},\mathcal O_2^{\mathcal U}=\ell^\infty(\mathbb N,\mathcal O_2)\big/\{(x_n):\lim_{n\to\mathcal U}\|x_n\|=0\},

where ℓ∞(N,O2)\ell^\infty(\mathbb N,\mathcal O_2) consists of bounded sequences. An embedding means an injective *-homomorphism; the question does not require it to be unital when the domain is nonunital. AimPL: Set theory and C∗ algebras

Background

This is the separable form of the Kirchberg Embedding Problem (KEP), which asks whether O2\mathcal O_2 is locally universal for C*-algebras: every algebra in the specified class embeds into an ultrapower of O2\mathcal O_2. The problem is a C*-algebraic analogue of the Connes Embedding Problem, with ultrapowers of a nuclear C*-algebra playing the role of the tracial ultrapower of the hyperfinite factor. The problem is attributed to Kirchberg; a survey reports that it was discussed with him in 2007 and that its first mention in the literature appeared in work of Goldbring and Sinclair in 2015. Goldbring, Model Theory and Ultrapower Embedding Problems

The locally universal viewpoint fits a general model-theoretic framework: for a theory with the joint embedding property, existentially closed models yield locally universal ultrapowers for countable models. This explains why the question is closely tied to whether O2\mathcal O_2 is existentially closed, rather than being only a question about a particular construction of embeddings. Goldbring, Model Theory and Ultrapower Embedding Problems

References

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 1

Claimed by OpenAI. Claims that the full group C*-algebra of Z[1/2]^3 semidirect (SL3(Z) times Z), with (M,k) acting by 2^k M, has no unital embedding into any norm ultrapower of any nonzero unital nuclear C*-algebra. The group is explicit and fixed; the target need not be separable. This is a negative norm-ultrapower embedding assertion, including the O2 target.See full solutionHide full solution

Scope relative to this problem: The actual theorem rules out unital embeddings of the fixed explicit full group Cstar algebra into norm ultrapowers of any nonzero unital nuclear algebra, including O2, for every free ultrafilter on N. The manuscript Introduction explicitly states, with a literature citation, that the unital and arbitrary-embedding formulations of the O2 norm-ultrapower problem are equivalent. That equivalence is source-reported; the theorem attribution remains no unital embedding. No tracial-ultrapower, QWEP or Connes embedding claim is made.

GitHub repository: https://github.com/openai/math

Manuscript: https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/An-explicit-obstruction-to-nuclear-norm-ultrapower-embeddings-September-23-2026/paper.pdf

  • OpenAI-292-01-An-explicit-obstruction-to-nuclear-norm-ultrapower-embeddings.pdf367,447 bytesOpen