The electron microscope improved our vision by a factor of 1 million, enabling us to see atoms. In this study, I aim for a similar leap by examining trades executed in nanoseconds, a million times more precise than the oft-used milliseconds. This approach allows us to observe asset reallocations among rapid-fire “tradebots,” including those used by high-frequency traders (HFTs). Some 20% of trades occur in submillisecond clusters, which seem to have no price instability. Although submillisecond trade bursts are costly to non-HFTs in terms of adverse selection, these costs can be avoided.