It may well be that the cross-membrane voltage difference is largely homogeneous, but not because intracellular protons are very mobile. To the contrary, the cytosol mobility (rate of diffusion) of protons, and ions in general, is very different from the high mobility measured in water or dilute solutions of small molecular species. According to this interesting review on "Voltage-Gated Proton Channels and Other Proton Transfer Pathways", the mobility of protons in cytoplasm was measured at $8-12.5 \times 10^{-7} \text{cm}^2/\text{s}$, which is over 200-fold lower than that for $H^+$ in water. Surprisingly, intracellular proton mobility is actually on par with the mobility of much bigger ions: for example the cytoplasmic mobility of $\text{Ca}^{2+}$ is in the range $7-12 \times 10^{-7} \text{cm}^2/\text{s}$, whereas $\text{Ca}^{2+}$ mobility in vitro is $\sim 5.2-5.4 \times 10^{-6} \text{cm}^2/\text{s}$, see for instance here.

So, although proton transfer through proton pumps and channels is extremely fast, intracellular proton diffusion may be much, much slower. Just to get an idea of how slow: this paper measured pH gradients and proton diffusion dynamics in rabbit ventricular myocytes and found that "intracellular $H^+$ mobility was low, acid taking 20-30s to move 40μm down the cell". And again, they estimated the $H^+$ diffusion coefficient at "$3.78 \times 10^{−7} \text{cm}^2/\text{s}$, a value more than 300-fold lower than the $H^+$ diffusion coefficient in a dilute, unbuffered solution". On the other hand, they find that "intrinsic $H^+_i$ mobility is consistent with an average diffusion coefficient for the intracellular mobile buffers of $\sim 9 \times 10^{−7} \text{cm}^2/\text{s}$". Add to this that protons may also concentrate and diffuse toward channels and pumps from the inner surface of the membrane, as opposed to the intracellular bulk, and the picture gets even more complicated.

Anyways, as far as the voltage difference across the membrane is concerned, the picture that follows from all this is that in the exterior there is an aqueous solution with a relatively high concentration of highly mobile protons, whereas in the interior there is a low concentration of comparatively very slow protons. Even if proton pumps are very fast in expelling individual protons across the membrane, the limiting factor remains cytosol diffusion. If you now assume a sparse, but largely homogeneous distribution of pumps, channels, and sinks, it's probably also safe to assume a homogeneous voltage, simply by symmetry. But for the diametrically opposite pump and sink situation, I think the answer is much less clear cut and depends on the relative rates of the sink and the pump on the background of slow cytosol diffusion.

It may well be that the cross-membrane voltage difference is largely homogeneous, but not because intracellular protons are very mobile. To the contrary, the cytosol mobility (rate of diffusion) of protons, and ions in general, is very different from the high mobility measured in water or dilute solutions of small molecular species. According to this interesting review on "Voltage-Gated Proton Channels and Other Proton Transfer Pathways", the mobility of protons in cytoplasm was measured at $8-12.5 \times 10^{-7} \text{cm}^2/\text{s}$, which is over 200-fold lower than that for $H^+$ in water. Surprisingly, intracellular proton mobility is actually on par with the mobility of much bigger ions: for example the cytoplasmic mobility of $\text{Ca}^{2+}$ is in the range $7-12 \times 10^{-7} \text{cm}^2/\text{s}$, whereas $\text{Ca}^{2+}$ mobility in vitro is $\sim 5.2-5.4 \times 10^{-6} \text{cm}^2/\text{s}$, see for instance here.

So, although proton transfer through proton pumps and channels is extremely fast, intracellular proton diffusion may be much, much slower. Just to get an idea of how slow: this paper measured pH gradients and proton diffusion dynamics in rabbit ventricular myocytes and found that "intracellular $H^+$ mobility was low, acid taking 20-30s to move 40μm down the cell". And again, they estimated the $H^+$ diffusion coefficient at "$3.78 \times 10^{−7} \text{cm}^2/\text{s}$, a value more than 300-fold lower than the $H^+$ diffusion coefficient in a dilute, unbuffered solution". On the other hand, they find that "intrinsic $H^+_i$ mobility is consistent with an average diffusion coefficient for the intracellular mobile buffers of $\sim 9 \times 10^{−7} \text{cm}^2/\text{s}$". Add to this that protons may also concentrate and diffuse toward channels and pumps from the inner surface of the membrane, as opposed to the intracellular bulk, and the picture gets even more complicated.

Anyways, as far as the voltage difference across the membrane is concerned, the picture that follows from all this is that in the exterior there is an aqueous solution with a relatively high concentration of highly mobile protons, whereas in the interior there is a low concentration of comparatively very slow protons. Even if proton pumps are very fast in expelling individual protons across the membrane, the limiting factor remains cytosol diffusion. If you now assume a sparse, but largely homogeneous distribution of pumps, channels, and sinks, it's probably also safe to assume a homogeneous voltage, simply by symmetry. But for the diametrically opposite pump and sink situation, I think the answer is much less clear cut and depends on the relative rates of the sink and the pump on the background of slow cytosol diffusion.