More flagella improve motility, particularly in spatially structured environments. But the benefits saturate. Our model predicts an optimum near 3% of proteome invested in flagella, close to the estimated WT investment. Beyond this point, the costs outweigh the benefits.