If quantum superposition creates repulsive gravity, would it only work on microscopic objects?

Your question conflates two distinct ideas, so let me untangle them.

The papers describe a measurement-conditional effect, not a new fundamental force. Here's what's actually happening:

When a source mass is prepared in a quantum superposition of two locations, it creates a superposition of gravitational fields pointing in opposite directions [AG-2026.02-534]. If you then *postselect* — that is, measure the source mass in one of those locations — you can retroactively "filter" the data to find that a probe mass moved *away* from the source. This looks like repulsion, but it's an artifact of quantum inference: you're conditioning on which outcome occurred, which mathematically produces a negative weak value (a counterintuitive quantum measurement quantity) [AG-2026.02-268].

The catch: This repulsion only appears *after postselection* — after you've observed where the source mass ended up. It's not a real force pushing the probe away; it's a statistical correlation that emerges from quantum measurement. You cannot use this to levitate a macroscopic object.

As for whether it scales: the proposed experiment uses spin-bearing nanocrystals, so yes, it's designed for microscopic objects [AG-2026.02-534]. But there's no physical law preventing it from working on larger masses — the quantum superposition itself could exist at any scale. The real barrier is decoherence: larger objects are harder to keep in quantum superposition because they constantly interact with their environment [AG-2025.09-021]. So the bottleneck isn't the effect itself, but maintaining the quantum state needed to create it.

None of these papers show how postselection-based repulsion could do useful macroscopic work.