Dark mode is now everywhere (this blog has it). It’s both accessibility-driven and aesthetically pleasing. But it often isn’t the highest benefit feature (and rarely, if ever, a feature that leads to product-market-fit).
It’s not a technically difficult thing to build — across my entire blog, supporting dark mode takes up about 20 lines of code. For more complicated UIs, or retrofitting legacy applications, it’s a bit more involved (not to mention the technical debt of supporting it for future changes).
In their quest for product-market fit, most startups should prioritize features that bring them closer to that goal. Dark mode might marginally increase conversion and retention, but it won’t be a stepwise change. Good product managers must allocate resources effectively (even when personally partial to features).
After you nail the core features, you can build things that further delight users (marginal benefit > 0). Sometimes the details can be differentiating. Perfectly crafted products make a difference (but never forget the core value proposition).
Dark mode is one of those 'not excited that you have it, but i'll be very upset if you don't' features. On 2x products I worked on that lacked it, the demand was tremendous, placing it in the top #5 demand. By itself, you're right, it would not lead to PMF but lack of dark mode created friction that spurned users to look for other items they did not like to justify decisions on whether to stay or leave. An unusual experience.
Whenever I see "now with dark mode!" on a product I'm already using, I roll my eyes.