Based on my network, their reasoning is a mix of:
- communication breakdowns due to the difficulty of building an async communication model (remember that most people are not as text/writing and async driven as Internet commenters on a tech forum), and operational work like quarterly planning, QBRs, product line prioritizations, etc can be done much faster in person instead of over 10-15 Zoom calls over 3-5 weeks.
- situations where remote workers are domiciled in a different location from where they were hired AND they didn't tell HR (this leaves companies legally liable for various tax, business incorporation, and safety regulations)
- the occasional case where a major security incident occurs due to lack of in-person verification (eg. The recent North Korea remote worker scandal, the ongoing Snowflake breach)
- Justify US headcount. One of the portfolio companies I was an observer for went public a couple years ago. When they switched to remote during the pandemic, they realized they could cut down their entire US engineering presence and overhire in Eastern Europe, Israel, and India and drastically reduce margins. Most larger institutional investors are clamoring for this now, and they have a point to a certain extent. IBM, Google, Crowdstrike, and Amazon have unofficially started this, and I'm not surprised if other companies are doing the same
There are ways to change an organization to resolve these issues, but they are hard, expensive, and not scalable for large organizations.