The employer might be liable for a discrimination claim, under the doctrine of disparate impact. See Texas Department of Housing & Community Affairs v. The Inclusive Communities Project, Inc and references cited therein. The idea is that an employer can be liable absent proof of intentional discrimination when a practice disproportionately affects protected classes of individuals, and the practice is not justified by reasonable business considerations. So it would depend on why this particular state of affairs in employment came about.
There is a test known as the 80% rule which attempts to quantify the notion of "under-representation" as evidence of discrimination. This test (not widely respected by the courts these days) might constitute evidence of discrimination, if a protected class is demonstrably under-represented. The current standard seems to be by comparison to random selection.
In EEOC v. Sambo's of Georgia, Inc., 530 F. Supp. 86, the court found that a grooming policy had a disparate impact on members of a religion (Sikhism), and was thus contrary to Title VII of the Civil Rights Act of 1964, although a requirement to shave does not obviously discriminate on the basis of religion. This points to an important element of a successful disparate impact claim, that there has to be a policy with a causal effect. In the above scenario, there is no proposed policy that has this effect. Note that the burden of proof is on the person suing for relief – they must have a theory of something the company does that causes this hiring pattern, a practice that is discriminatory.
The identified policy (whatever it might be) could be justified by a business necessity defense (Griggs v. Duke Power Co., 401 U.S. 424), by showing that the practice has a demonstrable relationship to the requirements of the job. That, b.t.w., would not excuse a racially or religiously discriminatory hiring policy for a factory manufacturing menorahs or kinaras.
If a company recruiting locally in Boise ID had a 75% black work force, it would be reasonable to suspect that something was up. But legally, without showing that this results from a unjustified policy of the company, mere statistically anomalous distribution is does not sustain a claim of discrimination.