The control here sits inside the same atom, which is the design's best feature. Both arms come from one cloud of rubidium on one chip, so the systematic that dogs any comparison between two separate samples never arises [6]. Free fall is what the design intends as the only difference between the two paths [5].
That intent is not quite achieved, by the release's own description. The arm that stayed put had to remain responsive to the chip's magnetic field in order to be held; the arm that fell had to be switched into a state almost unaffected by that field in order to be let go [15]. The two arms therefore differed in internal state as well as in trajectory, and any phase accumulated for internal reasons has to be accounted for before the remainder can be called gravitational. The release does not say how that was done [14]. The paper presumably does.
What is under test is narrower than "gravity in the quantum world" suggests. The prediction being checked is the one you get by applying the equivalence principle to a matter wave, and the measured phase agrees with it [3][11]. Gravity remains a classical background throughout. Ron Folman of Ben-Gurion University leaves the unification problem where it has always been, describing the experiment as a source of hints rather than a framework [13], and the release itself notes that quantum particles have been used to measure gravity before, with the novelty claimed being the direct measurement of the falling phase [12].
The thing this does not tell you is how much room was closed. A confirmation is worth what its error bar is worth, and the release carries no phase value and no uncertainty [14]. Until those numbers are read, the honest statement is that the predicted effect was seen, not that alternatives were excluded at some level. Note also that the release names the observable as a difference in quantum phase and never reaches for gravitational redshift [16]; a reader who meets that term in coverage is meeting an interpretation rather than a quoted result.
My view, with its condition attached: the durable output of this work is the instrument [5]. Holding one arm of a superposition against gravity while releasing the other is a geometry that can be pushed, and pushing it is how a consistency check turns into a test with teeth. If the published uncertainty is loose, the physics claim shrinks and the apparatus claim survives intact. That is a fair trade for one experiment, and it leaves the equivalence principle roughly where more than a century of precise tests on ordinary matter had already put it [4], now with a quantum wave added to the ledger.
