They could point out the inaccuracies or misrepresentations. For example, the claim that 20% of the staff were let go, but 100% of the laid off folk were "of historically marginalized groups. Women, people of color, queer folks, and multiple people on disability and even maternity leave"
They could respond with information. Maybe this is a broad enough group and their workforce is unusually diverse, and 90% of their workforce fits that description. If their layoff cohort was remotely representative of their demographics in the first place, this would be fine, but it really doesn't sound like it.
For example, the claim that they laid off people on maternity leave. Did their 20% company-wide layoff comprise 20% of their staff which was on maternity leave (in other words "sure, we laid off 2 people on maternity leave, but another 8 of our employees are actually on maternity leave, and they didn't get laid off").
Or was this overall inaccurate? ("actually our layoff cohort contained groups of people who are underrepresented in tech, but also groups of people who are overrepresented in tech. Protected statuses were not a factor in our layoff decisions").
Perhaps they fucked up and realize they did discriminate based on protected class, and can only dig the hole deeper by responding now.