In Big Data, "scalable" usually means being able to parallelize the workload on a cluster, in ML it means "can deal with large data sets".